def compute_data(_tuples, _arguments, _padding_y_by, _padding_z_by, _space_y_by, _space_z_by): _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True, _padding_y_by=_padding_y_by, _padding_z_by=_padding_z_by, _space_y_by=_space_y_by, _space_z_by=_space_z_by) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _correlations = [] for _tuple in _tuples: _experiment, _series, _group = _tuple _left_cell_fiber_densities = _experiments_fiber_densities[(_experiment, _series, _group, 'left_cell')] _right_cell_fiber_densities = _experiments_fiber_densities[(_experiment, _series, _group, 'right_cell')] _properties = load.group_properties(_experiment, _series, _group) _left_cell_fiber_densities = compute.remove_blacklist( _experiment, _series, _properties['cells_ids']['left_cell'], _left_cell_fiber_densities ) _right_cell_fiber_densities = compute.remove_blacklist( _experiment, _series, _properties['cells_ids']['right_cell'], _right_cell_fiber_densities ) _left_cell_fiber_densities_filtered, _right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _left_cell_fiber_densities, _right_cell_fiber_densities ) # ignore small arrays if len(_left_cell_fiber_densities_filtered) < compute.minimum_time_frames_for_correlation(_experiment): continue _correlation = compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_right_cell_fiber_densities_filtered, _n=DERIVATIVE) ) _correlations.append(_correlation) return np.mean(_correlations)
def main(): _experiments = all_experiments() _experiments = filtering.by_categories(_experiments=_experiments, _is_single_cell=True, _is_high_temporal_resolution=False, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_time_frames_amount( _tuples, compute.density_time_frame(_experiments[0])) _tuples = filtering.by_main_cell(_tuples) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _time_frame = compute.density_time_frame(_experiment) for _direction in ['left', 'right']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': 'cell', 'direction': _direction, 'time_points': _time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'direction']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _tuples = organize.by_single_cell_id(_tuples) print('Total tuples:', len(_tuples)) _experiments_ids = list(_tuples.keys()) _y_arrays = [[] for _i in DERIVATIVES] for _index_1 in tqdm(range(len(_experiments_ids)), desc='Main loop'): _tuple_1 = _experiments_ids[_index_1] _experiment_1, _series_id_1, _cell_id_1 = _tuple_1 _fiber_densities_1 = compute_single_cell_mean( _experiment=_experiment_1, _series_id=_series_id_1, _cell_tuples=_tuples[_tuple_1], _windows_dictionary=_windows_dictionary, _fiber_densities=_fiber_densities) for _index_2 in range(_index_1 + 1, len(_experiments_ids)): _tuple_2 = _experiments_ids[_index_2] _experiment_2, _series_id_2, _cell_id_2 = _tuple_2 _fiber_densities_2 = compute_single_cell_mean( _experiment=_experiment_2, _series_id=_series_id_2, _cell_tuples=_tuples[_tuple_2], _windows_dictionary=_windows_dictionary, _fiber_densities=_fiber_densities) for _derivative_index, _derivative in enumerate(DERIVATIVES): _y_arrays[_derivative_index].append( compute_lib.correlation( compute_lib.derivative(_fiber_densities_1, _n=_derivative), compute_lib.derivative(_fiber_densities_2, _n=_derivative))) print('Total points:', len(_y_arrays[0])) print('Wilcoxon around the zero') for _y_array, _derivative in zip(_y_arrays, DERIVATIVES): print('Derivative:', _derivative, wilcoxon(_y_array)) # plot _colors_array = config.colors(3) _fig = go.Figure(data=[ go.Box(y=_y, name=_derivative, boxpoints='all', jitter=1, pointpos=0, line={'width': 1}, fillcolor='white', marker={ 'size': 10, 'color': _color }, opacity=0.7, showlegend=False) for _y, _derivative, _color in zip( _y_arrays, DERIVATIVES_TEXT, _colors_array) ], layout={ 'xaxis': { 'title': 'Fiber density derivative', 'zeroline': False }, 'yaxis': { 'title': 'Correlation', 'range': [-1, 1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot')
def compute_fiber_densities(_band=True, _high_temporal_resolution=False): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples, _band=_band) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows( _arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _tuples_by_experiment = organize.by_experiment(_tuples) _distances_from_y_equal_x = [] _z_positions_array = [] for _experiment in _tuples_by_experiment: print('Experiment:', _experiment) _experiment_tuples = _tuples_by_experiment[_experiment] for _same_index in tqdm(range(len(_experiment_tuples)), desc='Main loop'): _same_tuple = _experiment_tuples[_same_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'left_cell') ] _same_right_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'right_cell') ] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities) _same_left_cell_fiber_densities_filtered, _same_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_left_cell_fiber_densities, _same_right_cell_fiber_densities ) # ignore small arrays if len(_same_left_cell_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _same_experiment): continue _same_correlation = compute_lib.correlation( compute_lib.derivative( _same_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _same_right_cell_fiber_densities_filtered, _n=DERIVATIVE)) _same_group_mean_z_position = \ compute.group_mean_z_position_from_substrate(_same_experiment, _same_series, _same_group) for _different_index in range(len(_experiment_tuples)): if _same_index != _different_index: _different_tuple = _experiment_tuples[_different_index] _different_experiment, _different_series, _different_group = \ _different_tuple for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _same_cell_id)] _different_fiber_densities = _experiments_fiber_densities[ (_different_experiment, _different_series, _different_group, _different_cell_id)] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'] [_different_cell_id], _different_fiber_densities) _same_fiber_densities_filtered, _different_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_fiber_densities, _different_fiber_densities ) # ignore small arrays if len(_same_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _different_experiment): continue _different_correlation = compute_lib.correlation( compute_lib.derivative( _same_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _different_fiber_densities_filtered, _n=DERIVATIVE)) _point_distance = compute_lib.distance_from_a_point_to_a_line( _line=[-1, -1, 1, 1], _point=[_same_correlation, _different_correlation]) if _same_correlation > _different_correlation: _distances_from_y_equal_x.append(_point_distance) else: _distances_from_y_equal_x.append(-_point_distance) _z_positions_array.append(_same_group_mean_z_position) print('Total points:', len(_distances_from_y_equal_x)) print('Wilcoxon of distances from y = x around the zero:') print(wilcoxon(_distances_from_y_equal_x)) print( 'Pearson correlation of distances from y = x and z position distances:' ) print( compute_lib.correlation(_distances_from_y_equal_x, _z_positions_array, _with_p_value=True)) return _distances_from_y_equal_x, _z_positions_array
def main(_real_cells=True, _static=False, _dead_dead=False, _live_dead=False, _dead=False, _live=False, _bead=False, _metastasis=False, _bleb=False, _bleb_amount_um=None, _band=True, _high_temporal_resolution=False, _offset_y=0.5): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=_bleb, _is_dead_dead=_dead_dead, _is_live_dead=_live_dead, _is_bead=_bead, _is_metastasis=_metastasis) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples, _real_pairs=_real_cells) _tuples = filtering.by_fake_static_pairs(_tuples, _fake_static_pairs=_static) if _dead_dead is not False or _live_dead is not False: _tuples = filtering.by_dead_live(_tuples, _dead=_dead, _live=_live) _tuples = filtering.by_band(_tuples, _band=_band) if _bleb: _tuples = filtering.by_bleb_amount_um(_tuples, _amount_um=_bleb_amount_um) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': _offset_y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows( _arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _tuples_by_experiment = organize.by_experiment(_tuples) _n = 0 _cells_ranks = [] for _experiment in _tuples_by_experiment: print('Experiment:', _experiment) _experiment_tuples = _tuples_by_experiment[_experiment] for _pivot_tuple in tqdm(_experiment_tuples, desc='Main loop'): _pivot_experiment, _pivot_series_id, _pivot_group = _pivot_tuple _pivot_experiment_properties = load.group_properties( _pivot_experiment, _pivot_series_id, _pivot_group) for _pivot_cell_id, _pivot_cell_correct_match_cell_id in \ zip(['left_cell', 'right_cell'], ['right_cell', 'left_cell']): _pivot_cell = (_pivot_experiment, _pivot_series_id, _pivot_group, _pivot_cell_id) _pivot_cell_correct_match_cell = ( _pivot_experiment, _pivot_series_id, _pivot_group, _pivot_cell_correct_match_cell_id) _pivot_cell_fiber_densities = _experiments_fiber_densities[ _pivot_cell] _pivot_cell_fiber_densities = compute.remove_blacklist( _pivot_experiment, _pivot_series_id, _pivot_experiment_properties['cells_ids'][_pivot_cell_id], _pivot_cell_fiber_densities) _pivot_cell_correlations = [] # correct match _pivot_cell_correct_match_fiber_densities = _experiments_fiber_densities[ _pivot_cell_correct_match_cell] _pivot_cell_correct_match_fiber_densities = compute.remove_blacklist( _pivot_experiment, _pivot_series_id, _pivot_experiment_properties['cells_ids'] [_pivot_cell_correct_match_cell_id], _pivot_cell_correct_match_fiber_densities) _pivot_cell_fiber_densities_filtered, _pivot_cell_correct_match_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _pivot_cell_fiber_densities, _pivot_cell_correct_match_fiber_densities ) # ignore small arrays if len(_pivot_cell_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _pivot_experiment): continue _correlation = compute_lib.correlation( compute_lib.derivative( _pivot_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _pivot_cell_correct_match_fiber_densities_filtered, _n=DERIVATIVE)) _pivot_cell_correlations.append(_correlation) _pivot_cell_correct_match_correlation = _correlation # create list of potential cells _candidate_tuples = [] for _candidate_tuple in _experiment_tuples: _candidate_experiment, _candidate_series_id, _candidate_group = _candidate_tuple for _candidate_cell_id in ['left_cell', 'right_cell']: _candidate_cell = (_candidate_experiment, _candidate_series_id, _candidate_group, _candidate_cell_id) # skip if same cell or correct match if _candidate_cell == _pivot_cell or _candidate_cell == _pivot_cell_correct_match_cell: continue _candidate_tuples.append(_candidate_cell) # compare with each potential candidate, until reached the maximum or nothing to compare with while len(_pivot_cell_correlations ) < POTENTIAL_MATCHES and len(_candidate_tuples) > 0: # sample randomly _candidate_cell = random.choice(_candidate_tuples) _candidate_experiment, _candidate_series_id, _candidate_group, _candidate_cell_id = _candidate_cell _candidate_experiment_properties = load.group_properties( _candidate_experiment, _candidate_series_id, _candidate_group) _candidate_cell_fiber_densities = _experiments_fiber_densities[ _candidate_cell] _candidate_cell_fiber_densities = compute.remove_blacklist( _candidate_experiment, _candidate_series_id, _candidate_experiment_properties['cells_ids'] [_candidate_cell_id], _candidate_cell_fiber_densities) _pivot_cell_fiber_densities_filtered, _candidate_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _pivot_cell_fiber_densities, _candidate_cell_fiber_densities ) # ignore small arrays if len(_pivot_cell_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _pivot_experiment): _candidate_tuples.remove(_candidate_cell) continue _correlation = compute_lib.correlation( compute_lib.derivative( _pivot_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _candidate_cell_fiber_densities_filtered, _n=DERIVATIVE)) _pivot_cell_correlations.append(_correlation) # nothing to compare with if len(_pivot_cell_correlations) == 1: continue # check matchmaking _pivot_cell_correct_match_rank = 1 for _potential_match_correlation in sorted( _pivot_cell_correlations, reverse=True): if _pivot_cell_correct_match_correlation == _potential_match_correlation: break _pivot_cell_correct_match_rank += 1 _n += 1 _cells_ranks.append(_pivot_cell_correct_match_rank) # results _correct_match_probability = 1 / POTENTIAL_MATCHES _first_place_correct_matches = sum( [1 for _rank in _cells_ranks if _rank == 1]) _first_place_fraction = _first_place_correct_matches / _n print('Matchmaking results:') print('Total cells:', _n) print('Correct match probability:', round(_correct_match_probability, 2)) print('Fraction of first place correct matches:', round(_first_place_fraction, 2)) # plot _x = list(range(MAX_RANK)) _x_text = [str(_rank + 1) for _rank in _x[:-1]] + [str(MAX_RANK) + '+'] _ranks_sums = [0 for _rank in _x] for _rank in _cells_ranks: if _rank < MAX_RANK: _ranks_sums[_rank - 1] += 1 else: _ranks_sums[-1] += 1 _y = np.array(_ranks_sums) / _n _fig = go.Figure(data=go.Bar(x=_x, y=_y, marker={'color': '#ea8500'}), layout={ 'xaxis': { 'title': 'Correct match correlation rank', 'zeroline': False, 'tickmode': 'array', 'tickvals': _x, 'ticktext': _x_text }, 'yaxis': { 'title': 'Fraction', 'range': [0, 1.1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_real_' + str(_real_cells) + '_static_' + str(_static) + '_dead_dead_' + str(_dead_dead) + '_live_dead_' + str(_live_dead) + '_dead_' + str(_dead) + '_live_' + str(_live) + '_bead_' + str(_bead) + '_metastasis_' + str(_metastasis) + '_bleb_' + str(_bleb) + str(_bleb_amount_um) + '_band_' + str(_band) + '_high_time_' + str(_high_temporal_resolution) + '_y_' + str(_offset_y)) # correct match probability plot _y = [_correct_match_probability] * (MAX_RANK - 1) + [ _correct_match_probability * (POTENTIAL_MATCHES - MAX_RANK + 1) ] _fig = go.Figure(data=go.Bar(x=_x, y=_y, marker={'color': '#ea8500'}), layout={ 'xaxis': { 'title': 'Correct match correlation rank', 'zeroline': False, 'tickmode': 'array', 'tickvals': _x, 'ticktext': _x_text }, 'yaxis': { 'title': 'Fraction', 'range': [0, 1.1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_real_' + str(_real_cells) + '_static_' + str(_static) + '_dead_dead_' + str(_dead_dead) + '_live_dead_' + str(_live_dead) + '_dead_' + str(_dead) + '_live_' + str(_live) + '_bead_' + str(_bead) + '_metastasis_' + str(_metastasis) + '_bleb_' + str(_bleb) + str(_bleb_amount_um) + '_band_' + str(_band) + '_high_time_' + str(_high_temporal_resolution) + '_y_' + str(_offset_y) + '_correct_match_prob')
def main(_real_cells=True, _static=False, _band=True, _high_temporal_resolution=False): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False ) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples, _real_pairs=_real_cells) _tuples = filtering.by_fake_static_pairs(_tuples, _fake_static_pairs=_static) _tuples = filtering.by_band(_tuples, _band=_band) _tuples = filtering.by_time_frames_amount(_tuples, compute.minimum_time_frames_for_correlation(_experiments[0])) print('Total tuples:', len(_tuples)) print('Density') _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _time_frame = compute.minimum_time_frames_for_correlation(_experiment) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y_DENSITY, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_point': _time_frame - 1 }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _densities_fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=False) _densities_experiments_fiber_densities = { _key: [_densities_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } print('Correlations') _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping(_experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y_CORRELATION, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _correlations_fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _correlations_experiments_fiber_densities = { _key: [_correlations_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _densities = [] _correlations = [] for _tuple in tqdm(_tuples, desc='Main loop'): _experiment, _series_id, _group = _tuple # density _left_cell_fiber_density = \ _densities_experiments_fiber_densities[(_experiment, _series_id, _group, 'left_cell')][0] _right_cell_fiber_density = \ _densities_experiments_fiber_densities[(_experiment, _series_id, _group, 'right_cell')][0] # not relevant if _left_cell_fiber_density[1] or _right_cell_fiber_density[1]: continue _normalization = load.normalization_series_file_data(_experiment, _series_id) _left_cell_fiber_density_normalized = compute_lib.z_score( _x=_left_cell_fiber_density[0], _average=_normalization['average'], _std=_normalization['std'] ) _right_cell_fiber_density_normalized = compute_lib.z_score( _x=_right_cell_fiber_density[0], _average=_normalization['average'], _std=_normalization['std'] ) _density = (_left_cell_fiber_density_normalized + _right_cell_fiber_density_normalized) / 2 # correlation _left_cell_fiber_densities = \ _correlations_experiments_fiber_densities[(_experiment, _series_id, _group, 'left_cell')] _right_cell_fiber_densities = \ _correlations_experiments_fiber_densities[(_experiment, _series_id, _group, 'right_cell')] _properties = load.group_properties(_experiment, _series_id, _group) _left_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_cell_fiber_densities) _right_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_cell_fiber_densities) _left_cell_fiber_densities_filtered, _right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _left_cell_fiber_densities, _right_cell_fiber_densities ) # ignore small arrays if len(_left_cell_fiber_densities_filtered) < compute.minimum_time_frames_for_correlation(_experiment): continue _correlation = compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_right_cell_fiber_densities_filtered, _n=DERIVATIVE) ) _densities.append(_density) _correlations.append(_correlation) print('Total tuples:', len(_densities)) print('Pearson correlation of densities and correlations:') print(compute_lib.correlation(_densities, _correlations, _with_p_value=True)) # plot _fig = go.Figure( data=go.Scatter( x=_densities, y=_correlations, mode='markers', marker={ 'size': 5, 'color': '#ea8500' }, showlegend=False ), layout={ 'xaxis': { 'title': 'Fiber density (z-score)', 'zeroline': False, 'range': [-1.1, 15.2], # 'tickmode': 'array', # 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'shapes': [ { 'type': 'line', 'x0': 0, 'y0': -1, 'x1': 0, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': 0, 'y0': -1, 'x1': 15, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } } ] } ) save.to_html( _fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_real_' + str(_real_cells) + '_static_' + str(_static) + '_band_' + str(_band) + '_high_time_' + str(_high_temporal_resolution) + '_y_density_' + str(OFFSET_Y_DENSITY) + '_y_correlation_' + str(OFFSET_Y_CORRELATION) )
def main(): _simulations = load.structured() _simulations = filtering.by_time_points_amount( _simulations, _time_points=SIMULATIONS_TIME_POINTS) _simulations = filtering.by_categories(_simulations, _is_single_cell=False, _is_heterogeneity=True, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False) _simulations = filtering.by_heterogeneity(_simulations, _std=STD) _simulations = filtering.by_pair_distances(_simulations, _distances=PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _fiber_densities = compute_fiber_densities(_simulations) _x_arrays = [[] for _i in PAIR_DISTANCE] _y_arrays = [[] for _i in PAIR_DISTANCE] for _distance_index, _distance in enumerate(PAIR_DISTANCE): _distance_simulations = filtering.by_pair_distance(_simulations, _distance=_distance) print('Distance:', _distance, 'Total simulations:', len(_distance_simulations)) for _simulation in tqdm(_distance_simulations, desc='Simulations loop'): for _direction in ['inside', 'outside']: _left_cell_fiber_densities = _fiber_densities[(_simulation, 'left_cell', _direction)] _right_cell_fiber_densities = _fiber_densities[(_simulation, 'right_cell', _direction)] _correlation = compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_right_cell_fiber_densities, _n=DERIVATIVE)) if _direction == 'inside': _x_arrays[_distance_index].append(_correlation) else: _y_arrays[_distance_index].append(_correlation) print('Wilcoxon of insides minus outsides around the zero:') print( wilcoxon( np.array(_x_arrays[_distance_index]) - np.array(_y_arrays[_distance_index]))) # 2d plots _colors_array = config.colors(4) _legendgroup_array = ['group_1', 'group_1', 'group_2', 'group_2'] _fig = go.Figure(data=[ go.Scatter(x=_x, y=_y, name='Dist. ' + str(_distance), mode='markers', marker={ 'size': 10, 'color': _color }, legendgroup=_legendgroup) for _x, _y, _distance, _color, _legendgroup in zip( _x_arrays, _y_arrays, PAIR_DISTANCE, _colors_array, _legendgroup_array) ], layout={ 'xaxis': { 'title': 'Inner correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Outer correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'legend': { 'x': 0.1, 'y': 1, 'xanchor': 'left', 'yanchor': 'top', 'bordercolor': 'black', 'borderwidth': 2, 'bgcolor': 'white', 'orientation': 'h' }, 'shapes': [{ 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot') for _x, _y, _distance, _color in zip(_x_arrays, _y_arrays, PAIR_DISTANCE, _colors_array): _fig = go.Figure(data=go.Scatter(x=_x, y=_y, name='Dist. ' + str(_distance), mode='markers', marker={ 'size': 10, 'color': _color }, showlegend=True), layout={ 'xaxis': { 'title': 'Inner correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Outer correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'legend': { 'xanchor': 'left', 'x': 0.1, 'yanchor': 'top', 'bordercolor': 'black', 'borderwidth': 2, 'bgcolor': 'white' }, 'shapes': [{ 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_distance_' + str(_distance)) # box plot _box_y_arrays = [[] for _i in PAIR_DISTANCE] for _x_array, _y_array, _distance_index in zip(_x_arrays, _y_arrays, range(len(PAIR_DISTANCE))): for _x, _y in zip(_x_array, _y_array): _point_distance = compute_lib.distance_from_a_point_to_a_line( _line=[-1, -1, 1, 1], _point=[_x, _y]) if _x > _y: _box_y_arrays[_distance_index].append(_point_distance) else: _box_y_arrays[_distance_index].append(-_point_distance) _fig = go.Figure(data=[ go.Box(y=_y, name=str(_distance), boxpoints='all', jitter=1, pointpos=0, line={'width': 1}, fillcolor='white', marker={ 'size': 10, 'color': _color }, showlegend=False) for _y, _distance, _color in zip( _box_y_arrays, PAIR_DISTANCE, _colors_array) ], layout={ 'xaxis': { 'title': 'Pair distance (cell diameter)', 'zeroline': False, 'tickmode': 'array', 'tickvals': PAIR_DISTANCE, 'type': 'category' }, 'yaxis': { 'title': 'Inner minus outer correlation', 'zeroline': False, 'tickmode': 'array', 'tickvals': [-0.2, 0, 0.2, 0.4] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_box')
def main(): _simulations = load.structured() _simulations = filtering.by_time_points_amount(_simulations, TIME_POINT) _simulations = filtering.by_categories(_simulations, _is_single_cell=False, _is_heterogeneity=True, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False) _simulations = filtering.by_heterogeneity(_simulations, _std=STD) _simulations = filtering.by_pair_distance(_simulations, _distance=PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _fiber_densities = compute_fiber_densities(_simulations) _n = 0 _cells_potential_matches = [] _cells_ranks = [] for _simulation_1 in tqdm(_simulations, desc='Main loop'): for _cell_1_id, _cell_1_correct_match_cell_id in zip( ['left_cell', 'right_cell'], ['right_cell', 'left_cell']): _cell_1 = (_simulation_1, _cell_1_id) _cell_1_correct_match = (_simulation_1, _cell_1_correct_match_cell_id) _cell_1_fiber_densities = _fiber_densities[_cell_1] _cell_1_correlations = [] _cell_1_correct_match_correlation = None for _simulation_2 in _simulations: for _cell_2_id in ['left_cell', 'right_cell']: _cell_2 = (_simulation_2, _cell_2_id) # same cell if _cell_1 == _cell_2: continue _cell_2_fiber_densities = _fiber_densities[_cell_2] _correlation = compute_lib.correlation( compute_lib.derivative(_cell_1_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_cell_2_fiber_densities, _n=DERIVATIVE)) _cell_1_correlations.append(_correlation) # correct match if _cell_2 == _cell_1_correct_match: _cell_1_correct_match_correlation = _correlation # correct match does not exist if _cell_1_correct_match_correlation is None: continue # check matchmaking _cell_1_total_potential_matches = len(_cell_1_correlations) if _cell_1_total_potential_matches > 1: _cell_1_correct_match_rank = 1 for _potential_match_correlation in sorted( _cell_1_correlations, reverse=True): if _cell_1_correct_match_correlation == _potential_match_correlation: break _cell_1_correct_match_rank += 1 _n += 1 _cells_potential_matches.append( _cell_1_total_potential_matches) _cells_ranks.append(_cell_1_correct_match_rank) # results _mean_cells_potential_matches = float(np.mean(_cells_potential_matches)) _mean_correct_match_probability = 1 / _mean_cells_potential_matches _first_place_correct_matches = sum( [1 for _rank in _cells_ranks if _rank == 1]) _first_place_fraction = _first_place_correct_matches / _n print('Matchmaking results:') print('Total cells:', _n) print('Average potential matches per cell:', round(_mean_cells_potential_matches, 2)) print('Average correct match probability:', round(_mean_correct_match_probability, 2)) print('Fraction of first place correct matches:', round(_first_place_fraction, 2)) # plot _x = list(range(MAX_RANK)) _x_text = [str(_rank + 1) for _rank in _x[:-1]] + [str(MAX_RANK) + '+'] _ranks_sums = [0 for _rank in _x] for _rank in _cells_ranks: if _rank < MAX_RANK: _ranks_sums[_rank - 1] += 1 else: _ranks_sums[-1] += 1 _y = np.array(_ranks_sums) / _n _fig = go.Figure(data=go.Bar(x=_x, y=_y, marker={'color': '#2e82bf'}), layout={ 'xaxis': { 'title': 'Correct match correlation rank', 'zeroline': False, 'tickmode': 'array', 'tickvals': _x, 'ticktext': _x_text }, 'yaxis': { 'title': 'Fraction', 'range': [0, 1.1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot') # correct match probability plot _y = [_mean_correct_match_probability] * (MAX_RANK - 1) + [ _mean_correct_match_probability * MAX_RANK ] _fig = go.Figure(data=go.Bar(x=_x, y=_y, marker={'color': '#2e82bf'}), layout={ 'xaxis': { 'title': 'Correct match correlation rank', 'zeroline': False, 'tickmode': 'array', 'tickvals': _x, 'ticktext': _x_text }, 'yaxis': { 'title': 'Fraction', 'range': [0, 1.1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_correct_match_prob')
def main(_band=True): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=None, _is_bleb=True, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False ) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, _distance_range=PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples, _real_pairs=REAL_CELLS) _tuples = filtering.by_fake_static_pairs(_tuples, _fake_static_pairs=STATIC) _tuples = filtering.by_band(_tuples, _band=_band) _tuples = filtering.by_bleb_from_start(_tuples, _from_start=False) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping(_experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _n_pairs = 0 _before_correlations = [] _after_correlations = [] for _tuple in tqdm(_tuples, desc='Experiments loop'): _experiment, _series_id, _group = _tuple _left_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'left_cell')] _right_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'right_cell')] _properties = load.group_properties(_experiment, _series_id, _group) _left_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_cell_fiber_densities) _right_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_cell_fiber_densities) _before_left_cell_fiber_densities = \ _left_cell_fiber_densities[:AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_experiment]] _before_right_cell_fiber_densities = \ _right_cell_fiber_densities[:AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_experiment]] _after_left_cell_fiber_densities = \ _left_cell_fiber_densities[AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_experiment]:] _after_right_cell_fiber_densities = \ _right_cell_fiber_densities[AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_experiment]:] _before_left_cell_fiber_densities_filtered, _before_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _before_left_cell_fiber_densities, _before_right_cell_fiber_densities) _after_left_cell_fiber_densities_filtered, _after_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _after_left_cell_fiber_densities, _after_right_cell_fiber_densities) # ignore small arrays _minimum_time_frame_for_correlation = compute.minimum_time_frames_for_correlation(_experiment) if len(_before_left_cell_fiber_densities_filtered) < _minimum_time_frame_for_correlation or \ len(_after_left_cell_fiber_densities_filtered) < _minimum_time_frame_for_correlation: continue _n_pairs += 1 _before_correlations.append(compute_lib.correlation( compute_lib.derivative(_before_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_before_right_cell_fiber_densities_filtered, _n=DERIVATIVE) )) _after_correlations.append(compute_lib.correlation( compute_lib.derivative(_after_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_after_right_cell_fiber_densities_filtered, _n=DERIVATIVE) )) print('Total pairs:', _n_pairs) _before_minus_after = np.array(_before_correlations) - np.array(_after_correlations) print('Wilcoxon of before minus after around the zero:') print(wilcoxon(_before_minus_after)) print('Higher before amount:', (_before_minus_after > 0).sum() / len(_before_minus_after)) # plot _fig = go.Figure( data=go.Scatter( x=_before_correlations, y=_after_correlations, mode='markers', marker={ 'size': 5, 'color': '#ea8500' }, showlegend=False ), layout={ 'xaxis': { 'title': 'Correlation before bleb', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Correlation after bleb', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'shapes': [ { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } } ] } ) save.to_html( _fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot' )
def main(_high_temporal_resolution=True): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple for _cell_id in ['left_cell', 'right_cell']: _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) if ALIGNMENT_OFFSET_Y != OFFSET_Y: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': ALIGNMENT_OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id', 'offset_y']) _fiber_densities = compute.fiber_densities(_windows_to_compute) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _experiments_fiber_densities_aligned = align_by_z_score( _tuples, _experiments_fiber_densities) _tuples_by_experiment = organize.by_experiment(_tuples) _same_correlations_array = [] _different_correlations_array = [] _valid_tuples = [] for _experiment in _tuples_by_experiment: print('Experiment:', _experiment) _experiment_tuples = _tuples_by_experiment[_experiment] for _same_index in tqdm(range(len(_experiment_tuples)), desc='Main loop'): _same_tuple = _experiment_tuples[_same_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = \ _experiments_fiber_densities_aligned[ (_same_experiment, _same_series, _same_group, 'left_cell') ] _same_right_cell_fiber_densities = \ _experiments_fiber_densities_aligned[ (_same_experiment, _same_series, _same_group, 'right_cell') ] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities) _same_left_cell_fiber_densities_filtered, _same_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_left_cell_fiber_densities, _same_right_cell_fiber_densities ) # ignore small arrays if len(_same_left_cell_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _same_experiment): continue _same_correlation = compute_lib.correlation( compute_lib.derivative( _same_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _same_right_cell_fiber_densities_filtered, _n=DERIVATIVE)) for _different_index in range(len(_experiment_tuples)): if _same_index != _different_index: _different_tuple = _experiment_tuples[_different_index] _different_experiment, _different_series, _different_group = \ _different_tuple for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities_aligned[ (_same_experiment, _same_series, _same_group, _same_cell_id)] _different_fiber_densities = _experiments_fiber_densities_aligned[ (_different_experiment, _different_series, _different_group, _different_cell_id)] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'] [_different_cell_id], _different_fiber_densities) _same_fiber_densities_filtered, _different_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_fiber_densities, _different_fiber_densities ) # ignore small arrays if len(_same_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _different_experiment): continue _different_correlation = compute_lib.correlation( compute_lib.derivative( _same_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _different_fiber_densities_filtered, _n=DERIVATIVE)) _same_correlations_array.append(_same_correlation) _different_correlations_array.append( _different_correlation) if _same_tuple not in _valid_tuples: _valid_tuples.append(_same_tuple) print('Total tuples:', len(_valid_tuples)) print('Total points:', len(_same_correlations_array)) _same_minus_different = \ np.array(_same_correlations_array) - np.array(_different_correlations_array) print('Wilcoxon of same minus different around the zero:') print(wilcoxon(_same_minus_different)) print('Higher same amount:', (_same_minus_different > 0).sum() / len(_same_minus_different)) # plot _fig = go.Figure(data=go.Scatter(x=_same_correlations_array, y=_different_correlations_array, mode='markers', marker={ 'size': 5, 'color': '#ea8500' }, showlegend=False), layout={ 'xaxis': { 'title': 'Same network correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Different network correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'shapes': [{ 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_high_time_' + str(_high_temporal_resolution))
def main(): _simulations = load.structured() _simulations = filtering.by_categories(_simulations, _is_single_cell=False, _is_heterogeneity=HETEROGENEITY, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False) _simulations = filtering.by_pair_distances(_simulations, PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _simulations_by_distances = organize.by_pair_distance(_simulations) for _distance in _simulations_by_distances: print('Distance ', _distance, ', total simulations:', len(_simulations_by_distances[_distance])) _fiber_densities = compute_fiber_densities(_simulations) _heatmap_fiber = [] _heatmap_fiber_change = [] for _simulation in _simulations: _simulation_normalization = load.normalization(_simulation) for _cell_id in ['left_cell', 'right_cell']: _cell_fiber_densities = _fiber_densities[(_simulation, _cell_id)] # not enough data if len(_cell_fiber_densities) < DERIVATIVE + 1: continue _z_score_fiber_density = compute_lib.z_score_array( _array=_cell_fiber_densities, _average=_simulation_normalization['average'], _std=_simulation_normalization['std']) _heatmap_fiber += _z_score_fiber_density[DERIVATIVE:] _heatmap_fiber_change += compute_lib.derivative( _z_score_fiber_density, _n=DERIVATIVE) print( compute_lib.correlation(_heatmap_fiber, _heatmap_fiber_change, _with_p_value=True)) if PLOT: _y_shape = int(round((Y_LABELS_END - Y_LABELS_START) * Y_BINS)) _x_shape = int(round((X_LABELS_END - X_LABELS_START) * X_BINS)) _total_points = 0 _z_array = np.zeros(shape=(_y_shape, _x_shape)) for _y, _x in zip(_heatmap_fiber_change, _heatmap_fiber): _y_rounded, _x_rounded = int(round(_y * Y_BINS)), int( round(_x * X_BINS)) _y_index, _x_index = int(_y_rounded - Y_LABELS_START * Y_BINS), int(_x_rounded - X_LABELS_START * X_BINS) if 0 <= _y_index < _z_array.shape[ 0] and 0 <= _x_index < _z_array.shape[1]: _z_array[_y_index][_x_index] += 1 _total_points += 1 _z_array = _z_array / _total_points if not CONDITIONAL_NORMALIZATION: _z_array[_z_array == 0] = None else: _z_array_plot = np.zeros(shape=np.array(_z_array).shape) for _fiber_index, _fiber_density_z_score in enumerate(_z_array): _sum = np.sum(_fiber_density_z_score) for _change_index, _change_z_score in enumerate( _fiber_density_z_score): _z_array_plot[_fiber_index][_change_index] = ( _change_z_score / _sum) if _sum != 0 else 0 _z_array_plot[_z_array_plot == 0] = None _fig = go.Figure( data=go.Heatmap(x=np.arange(start=X_LABELS_START, stop=X_LABELS_END, step=1 / X_BINS), y=np.arange(start=Y_LABELS_START, stop=Y_LABELS_END, step=1 / Y_BINS), z=_z_array, colorscale='Viridis', colorbar={ 'tickmode': 'array', 'tickvals': [0, 0.025, 0.05], 'ticktext': ['0', 'Fraction', '0.05'], 'tickangle': -90 }, zmin=Z_MIN, zmax=Z_MAX[CONDITIONAL_NORMALIZATION]), layout={ 'xaxis': { 'title': 'fiber densities z-score', 'zeroline': False }, 'yaxis': { 'title': 'Change in fiber<br>density (z-score)', 'zeroline': False }, 'shapes': [{ 'type': 'line', 'x0': X_LABELS_START, 'y0': Y_LABELS_START, 'x1': X_LABELS_END, 'y1': Y_LABELS_START, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': X_LABELS_START, 'y0': Y_LABELS_START, 'x1': X_LABELS_START, 'y1': Y_LABELS_END, 'line': { 'color': 'black', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_direction_' + DIRECTION)
def main(_band=True): _experiments = all_experiments() _experiments = filtering.by_categories(_experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=None, _is_bleb=True, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range( _tuples, _distance_range=PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples, _real_pairs=REAL_CELLS) _tuples = filtering.by_fake_static_pairs(_tuples, _fake_static_pairs=STATIC) _tuples = filtering.by_band(_tuples, _band=_band) _tuples = filtering.by_bleb_from_start(_tuples, _from_start=False) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows( _arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _tuples_by_experiment = organize.by_experiment(_tuples) # same (before, after), different (before, after) _correlations = [[[], []], [[], []]] _valid_real_tuples = [] for _experiment in _tuples_by_experiment: print('Experiment:', _experiment) _experiment_tuples = _tuples_by_experiment[_experiment] for _same_index in tqdm(range(len(_experiment_tuples)), desc='Main loop'): _same_tuple = _experiment_tuples[_same_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'left_cell') ] _same_right_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'right_cell') ] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities) _same_before_left_cell_fiber_densities = \ _same_left_cell_fiber_densities[:AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_same_experiment]] _same_before_right_cell_fiber_densities = \ _same_right_cell_fiber_densities[:AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_same_experiment]] _same_after_left_cell_fiber_densities = \ _same_left_cell_fiber_densities[AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_same_experiment]:] _same_after_right_cell_fiber_densities = \ _same_right_cell_fiber_densities[AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_same_experiment]:] _same_before_left_cell_fiber_densities_filtered, _same_before_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_before_left_cell_fiber_densities, _same_before_right_cell_fiber_densities ) _same_after_left_cell_fiber_densities_filtered, _same_after_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_after_left_cell_fiber_densities, _same_after_right_cell_fiber_densities ) # ignore small arrays _minimum_time_frame_for_correlation = compute.minimum_time_frames_for_correlation( _same_experiment) if len(_same_before_left_cell_fiber_densities_filtered) < _minimum_time_frame_for_correlation or \ len(_same_after_left_cell_fiber_densities_filtered) < _minimum_time_frame_for_correlation: continue _same_before_correlation = compute_lib.correlation( compute_lib.derivative( _same_before_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _same_before_right_cell_fiber_densities_filtered, _n=DERIVATIVE)) _same_after_correlation = compute_lib.correlation( compute_lib.derivative( _same_after_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _same_after_right_cell_fiber_densities_filtered, _n=DERIVATIVE)) for _different_index in range(len(_experiment_tuples)): if _same_index != _different_index: _different_tuple = _experiment_tuples[_different_index] _different_experiment, _different_series, _different_group = _different_tuple for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _same_cell_id)] _different_fiber_densities = _experiments_fiber_densities[ (_different_experiment, _different_series, _different_group, _different_cell_id)] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'] [_different_cell_id], _different_fiber_densities) _same_before_fiber_densities = \ _same_fiber_densities[:AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_same_experiment]] _same_after_fiber_densities = \ _same_fiber_densities[AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_same_experiment]:] _different_before_fiber_densities = \ _different_fiber_densities[:AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_different_experiment]] _different_after_fiber_densities = \ _different_fiber_densities[AFTER_BLEB_INJECTION_FIRST_TIME_FRAME[_different_experiment]:] _same_before_fiber_densities_filtered, _different_before_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_before_fiber_densities, _different_before_fiber_densities ) _same_after_fiber_densities_filtered, _different_after_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_after_fiber_densities, _different_after_fiber_densities ) # ignore small arrays if len(_same_before_fiber_densities_filtered) < _minimum_time_frame_for_correlation or \ len(_same_after_fiber_densities_filtered) < _minimum_time_frame_for_correlation: continue _different_before_correlation = compute_lib.correlation( compute_lib.derivative( _same_before_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _different_before_fiber_densities_filtered, _n=DERIVATIVE)) _different_after_correlation = compute_lib.correlation( compute_lib.derivative( _same_after_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _different_after_fiber_densities_filtered, _n=DERIVATIVE)) _correlations[0][0].append(_same_before_correlation) _correlations[0][1].append(_same_after_correlation) _correlations[1][0].append( _different_before_correlation) _correlations[1][1].append( _different_after_correlation) if _same_tuple not in _valid_real_tuples: _valid_real_tuples.append(_same_tuple) print('Total tuples:', len(_valid_real_tuples)) _distances_from_y_equal_x = [[], []] _same_correlations, _different_correlations = _correlations _same_before_correlations, _same_after_correlations = _same_correlations _different_before_correlations, _different_after_correlations = _different_correlations for _same_before, _same_after, _different_before, _different_after in \ zip(_same_before_correlations, _same_after_correlations, _different_before_correlations, _different_after_correlations): for _group_type_index, _same, _different in \ zip([0, 1], [_same_before, _same_after], [_different_before, _different_after]): _point_distance = compute_lib.distance_from_a_point_to_a_line( _line=[-1, -1, 1, 1], _point=[_same, _different]) if _same > _different: _distances_from_y_equal_x[_group_type_index].append( _point_distance) else: _distances_from_y_equal_x[_group_type_index].append( -_point_distance) print('Total points:', len(_distances_from_y_equal_x[0])) print('Higher before same amount:', (np.array(_distances_from_y_equal_x[0]) > 0).sum() / len(_distances_from_y_equal_x[0])) print('Wilcoxon of before points:', wilcoxon(_distances_from_y_equal_x[0])) print('Higher after same amount:', (np.array(_distances_from_y_equal_x[1]) > 0).sum() / len(_distances_from_y_equal_x[1])) print('Wilcoxon of after points:', wilcoxon(_distances_from_y_equal_x[1])) _before_minus_after = np.array(_distances_from_y_equal_x[0]) - np.array( _distances_from_y_equal_x[1]) print('Before > after amount:', (_before_minus_after > 0).sum() / len(_before_minus_after)) print('Wilcoxon before & after:', wilcoxon(_distances_from_y_equal_x[0], _distances_from_y_equal_x[1])) # box plot _colors_array = config.colors(2) _names_array = ['Before', 'After'] _fig = go.Figure(data=[ go.Box(y=_y_array, name=_name, boxpoints=False, line={'width': 1}, marker={'color': _color}, showlegend=False) for _y_array, _name, _color in zip( _distances_from_y_equal_x, _names_array, _colors_array) ], layout={ 'xaxis': { 'zeroline': False }, 'yaxis': { 'title': 'Same minus different correlation', 'zeroline': False, 'range': [-1, 1.1], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_box') # scatter plot _fig = go.Figure(data=go.Scatter(x=_distances_from_y_equal_x[0], y=_distances_from_y_equal_x[1], mode='markers', marker={ 'size': 5, 'color': '#ea8500' }, showlegend=False), layout={ 'xaxis': { 'title': 'Before bleb', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'After bleb', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'shapes': [{ 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot')
def main(): _simulations = load.structured() _simulations = filtering.by_time_points_amount(_simulations, _time_points=TIME_POINTS) _simulations = filtering.by_categories(_simulations, _is_single_cell=False, _is_heterogeneity=False, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False) _simulations = filtering.by_pair_distance(_simulations, _distance=PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _fiber_densities = compute_fiber_densities(_simulations) _y_arrays = [[] for _i in DERIVATIVES] for _simulation in tqdm(_simulations, desc='Simulations loop'): _normalization = load.normalization(_simulation) for _cell_id in ['left_cell', 'right_cell']: _cell_fiber_densities = _fiber_densities[(_simulation, _cell_id)] _cell_fiber_densities_normalized = compute_lib.z_score_array( _array=_cell_fiber_densities, _average=_normalization['average'], _std=_normalization['std']) for _derivative_index, _derivative in enumerate(DERIVATIVES): _y_arrays[_derivative_index].append( compute_lib.derivative(_cell_fiber_densities_normalized, _n=_derivative)) print('Total cells:', len(_y_arrays[0])) # plot for _derivative, _y_array in zip(DERIVATIVES, _y_arrays): _fig = go.Figure( data=go.Scatter(x=list(range(TIME_POINTS))[::TIME_POINTS_STEP], y=np.mean(_y_array, axis=0)[::TIME_POINTS_STEP], name='Fiber density (z-score)', error_y={ 'type': 'data', 'array': np.std(_y_array, axis=0)[::TIME_POINTS_STEP], 'thickness': 1, 'color': DERIVATIVES_COLORS[_derivative] }, mode='markers', marker={ 'size': 15, 'color': DERIVATIVES_COLORS[_derivative] }, opacity=0.7), layout={ 'xaxis': { 'title': 'Cell contraction (%)', 'zeroline': False }, 'yaxis': { 'title': 'Fiber density (z-score)', 'zeroline': False, 'tickmode': 'array', 'tickvals': DERIVATIVES_Y_TICKVALS[_derivative] }, 'shapes': [{ 'type': 'line', 'x0': -2, 'y0': (np.mean(_y_array, axis=0)[0] - np.std(_y_array, axis=0)[0]) * 1.5, 'x1': 53, 'y1': (np.mean(_y_array, axis=0)[0] - np.std(_y_array, axis=0)[0]) * 1.5, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -2, 'y0': (np.mean(_y_array, axis=0)[0] - np.std(_y_array, axis=0)[0]) * 1.5, 'x1': -2, 'y1': (np.mean(_y_array, axis=0)[-1] + np.std(_y_array, axis=0)[-1]), 'line': { 'color': 'black', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_derivative_' + str(_derivative))
def main(): _simulations = load.structured() _simulations = filtering.by_time_points_amount(_simulations, TIME_POINT) _simulations = filtering.by_categories( _simulations, _is_single_cell=False, _is_heterogeneity=None, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False ) _simulations = filtering.by_heterogeneities(_simulations, _stds=STDS) _simulations = filtering.by_pair_distance(_simulations, _distance=PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _fiber_densities = compute_fiber_densities(_simulations) # stds loop _std_communicating = [[] for _i in STDS] _std_non_communicating = [[] for _i in STDS] for _std_index, _std in enumerate(STDS): print('Std:', _std) _simulations_std = filtering.by_heterogeneity(_simulations, _std=_std) print('Total simulations:', len(_simulations_std)) # communicating loop for _simulation in tqdm(_simulations_std, desc='Communicating pairs loop'): _left_cell_fiber_densities = _fiber_densities[(_simulation, 'left_cell')] _right_cell_fiber_densities = _fiber_densities[(_simulation, 'right_cell')] _correlation = compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_right_cell_fiber_densities, _n=DERIVATIVE) ) _std_communicating[_std_index].append(_correlation) # non-communicating loop _simulations_indices = range(len(_simulations_std)) for _simulation_1_index in tqdm(_simulations_indices, desc='Non-communicating pairs loop'): _simulation_1 = _simulations_std[_simulation_1_index] for _simulation_2_index in _simulations_indices[_simulation_1_index + 1:]: _simulation_2 = _simulations_std[_simulation_2_index] for _simulation_1_cell_id, _simulation_2_cell_id in product(['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _simulation_1_fiber_densities = _fiber_densities[(_simulation_1, _simulation_1_cell_id)] _simulation_2_fiber_densities = _fiber_densities[(_simulation_2, _simulation_2_cell_id)] _correlation = compute_lib.correlation( compute_lib.derivative(_simulation_1_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_simulation_2_fiber_densities, _n=DERIVATIVE) ) _std_non_communicating[_std_index].append(_correlation) print('Wilcoxon rank-sum test:', ranksums(_std_communicating[_std_index], _std_non_communicating[_std_index])) # histogram plot, once with and once without the legend _colors_array = config.colors(len(STDS)) for _show_legend in [True, False]: _fig = go.Figure( data=[ *[ go.Histogram( x=_non_communicating, name='STD ' + str(_std), marker={ 'color': _color }, nbinsx=10, histnorm='probability', showlegend=_show_legend ) for _std, _non_communicating, _color in zip(STDS, _std_non_communicating, _colors_array) ], *[ go.Scatter( x=_communicating, y=[0.5 - 0.02 * _std_index for _i in range(len(_communicating))], mode='text', text='●', textfont={ 'color': _color, 'size': 15 }, showlegend=False ) for _std_index, _communicating, _color in zip(range(len(STDS)), _std_communicating, _colors_array) ] ], layout={ 'xaxis': { 'title': 'Correlation', 'zeroline': False }, 'yaxis': { 'title': 'Fraction', 'zeroline': False }, 'legend': { 'xanchor': 'right', 'yanchor': 'top', 'bordercolor': 'black', 'borderwidth': 2 }, 'bargap': 0.1 } ) save.to_html( _fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_legend_' + str(_show_legend) )
def compute_tuples(_tuples): _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple # stop when windows are overlapping _properties = load.group_properties(_experiment, _series_id, _group) _latest_time_frame = len(_properties['time_points']) _middle_offsets_x = [] for _time_frame in range(len(_properties['time_points'])): _pair_distance = \ compute.pair_distance_in_cell_size_time_frame(_experiment, _series_id, _group, _time_frame) if _time_frame == 0: print(_tuple, 'Pair distance:', _pair_distance) _middle_offsets_x.append( (_time_frame, (_pair_distance - 1) / 2 - QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER / 2)) if _pair_distance - 1 - OFFSET_X * 2 < QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER * 2: _latest_time_frame = _time_frame - 1 break for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame, 'middle_time_frame': -1 }) # middle one for (_time_frame, _offset_x) in _middle_offsets_x: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': _offset_x, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': 'left_cell', 'direction': 'inside', 'time_point': _time_frame, 'middle_time_frame': _time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id', 'middle_time_frame']) _fiber_densities = compute.fiber_densities(_windows_to_compute) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _tuples_data = [] for _tuple in tqdm(_tuples, desc='Main loop'): _experiment, _series_id, _group = _tuple _left_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'left_cell', -1)] _right_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'right_cell', -1)] _middle_fiber_densities = \ [_experiments_fiber_densities[(_experiment, _series_id, _group, 'left_cell', _time_frame)][0] for _time_frame in range(0, len(_left_cell_fiber_densities))] _properties = load.group_properties(_experiment, _series_id, _group) _left_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_cell_fiber_densities) _right_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_cell_fiber_densities) _left_cell_fiber_densities_filtered, _right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _left_cell_fiber_densities, _right_cell_fiber_densities) if len(_left_cell_fiber_densities_filtered) == 0: continue _start_time_frame = 0 for _left in _left_cell_fiber_densities: if _left[0] == _left_cell_fiber_densities_filtered[0]: break _start_time_frame += 1 _middle_fiber_densities_filtered = \ [_fiber_density[0] for _fiber_density in _middle_fiber_densities] _middle_fiber_densities_filtered = \ _middle_fiber_densities_filtered[_start_time_frame: _start_time_frame + len(_left_cell_fiber_densities_filtered)] _normalization = load.normalization_series_file_data( _experiment, _series_id) _left_cell_fiber_densities_normalized = compute_lib.z_score_array( _array=_left_cell_fiber_densities_filtered, _average=_normalization['average'], _std=_normalization['std']) _right_cell_fiber_densities_normalized = compute_lib.z_score_array( _array=_right_cell_fiber_densities_filtered, _average=_normalization['average'], _std=_normalization['std']) _middle_fiber_densities_normalized = compute_lib.z_score_array( _array=_middle_fiber_densities_filtered, _average=_normalization['average'], _std=_normalization['std']) _left_cell_fiber_densities_normalized = \ compute_lib.derivative(_left_cell_fiber_densities_normalized, _n=DERIVATIVE) _right_cell_fiber_densities_normalized = \ compute_lib.derivative(_right_cell_fiber_densities_normalized, _n=DERIVATIVE) _middle_fiber_densities_normalized = \ compute_lib.derivative(_middle_fiber_densities_normalized, _n=DERIVATIVE) _correlation = \ compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities_normalized, _n=1), compute_lib.derivative(_right_cell_fiber_densities_normalized, _n=1) ) _tuples_data.append([ _tuple, _start_time_frame, (_left_cell_fiber_densities_normalized, _right_cell_fiber_densities_normalized, _middle_fiber_densities_normalized), _correlation ]) # plots _names_array = ['Left cell', 'Right cell', 'Middle'] _colors_array = config.colors(3) for _tuple_data in _tuples_data: _tuple, _start_time_frame, _y_arrays, _correlation = _tuple_data _fig = go.Figure(data=[ go.Scatter(x=np.arange( start=_start_time_frame, stop=_start_time_frame + len(_y_arrays[0]) - DERIVATIVE, step=1) * TEMPORAL_RESOLUTION, y=_y, name=_name, mode='lines', line={'color': _color}) for _y, _name, _color in zip( _y_arrays, _names_array, _colors_array) ], layout={ 'xaxis': { 'title': 'Time (minutes)', 'zeroline': False }, 'yaxis': { 'title': 'Fiber density (z-score)', 'zeroline': False, 'range': [-3, 8], 'tickmode': 'array', 'tickvals': [-3, 0, 3, 6] }, 'legend': { 'xanchor': 'left', 'x': 0.1, 'yanchor': 'top', 'bordercolor': 'black', 'borderwidth': 2, 'bgcolor': 'white' }, }) _experiment, _series_id, _group = _tuple save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_' + _experiment + '_' + str(_series_id) + '_' + _group) print('Tuple:', _tuple, 'Correlation:', _correlation)
def compute_fiber_densities(_band=True, _high_temporal_resolution=True): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples, _band=_band) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows( _arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _same_correlation_vs_time_lag = {} _same_time_lags_arrays = [[] for _i in TIME_LAGS[_high_temporal_resolution]] _different_time_lags_arrays = [ [] for _i in TIME_LAGS[_high_temporal_resolution] ] _same_time_lags_highest = [ 0 for _i in TIME_LAGS[_high_temporal_resolution] ] _different_time_lags_highest = [ 0 for _i in TIME_LAGS[_high_temporal_resolution] ] _valid_tuples = [] for _same_index in tqdm(range(len(_tuples)), desc='Main loop'): _same_tuple = _tuples[_same_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'left_cell') ] _same_right_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'right_cell') ] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities) # time lag _same_highest_correlation = -1.1 _same_highest_correlation_time_lag_index = 0 _same_correlation_vs_time_lag[_same_tuple] = [] for _time_lag_index, _time_lag in enumerate( TIME_LAGS[_high_temporal_resolution]): # choose either negative or positive lag for _symbol in [-1, 1]: # if no time lag consider it only once if _time_lag == 0 and _symbol == -1: continue _time_lag_symbol = _time_lag * _symbol if _time_lag_symbol > 0: _same_left_cell_fiber_densities_time_lag = _same_left_cell_fiber_densities[: -_time_lag_symbol] _same_right_cell_fiber_densities_time_lag = _same_right_cell_fiber_densities[ _time_lag_symbol:] elif _time_lag_symbol < 0: _same_left_cell_fiber_densities_time_lag = _same_left_cell_fiber_densities[ -_time_lag_symbol:] _same_right_cell_fiber_densities_time_lag = _same_right_cell_fiber_densities[: _time_lag_symbol] else: _same_left_cell_fiber_densities_time_lag = _same_left_cell_fiber_densities _same_right_cell_fiber_densities_time_lag = _same_right_cell_fiber_densities _same_left_cell_fiber_densities_filtered, _same_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_left_cell_fiber_densities_time_lag, _same_right_cell_fiber_densities_time_lag ) # ignore small arrays if len(_same_left_cell_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _same_experiment): _same_correlation_vs_time_lag[_same_tuple].append(None) continue _same_correlation = compute_lib.correlation( compute_lib.derivative( _same_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _same_right_cell_fiber_densities_filtered, _n=DERIVATIVE)) _same_time_lags_arrays[_time_lag_index].append( _same_correlation) _same_correlation_vs_time_lag[_same_tuple].append( _same_correlation) if _same_correlation > _same_highest_correlation: _same_highest_correlation = _same_correlation _same_highest_correlation_time_lag_index = _time_lag_index _same_time_lags_highest[_same_highest_correlation_time_lag_index] += 1 for _different_index in range(len(_tuples)): if _same_index != _different_index: _different_tuple = _tuples[_different_index] _different_experiment, _different_series, _different_group = \ _different_tuple for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _same_cell_id)] _different_fiber_densities = _experiments_fiber_densities[( _different_experiment, _different_series, _different_group, _different_cell_id)] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'][_different_cell_id], _different_fiber_densities) # time lag _different_highest_correlation = -1.1 _different_highest_correlation_time_lag_index = 0 for _time_lag_index, _time_lag in enumerate( TIME_LAGS[_high_temporal_resolution]): # choose either negative or positive lag for _symbol in [-1, 1]: # if no time lag consider it only once if _time_lag == 0 and _symbol == -1: continue _time_lag_symbol = _time_lag * _symbol if _time_lag_symbol > 0: _same_fiber_densities_time_lag = _same_fiber_densities[: -_time_lag_symbol] _different_fiber_densities_time_lag = _different_fiber_densities[ _time_lag_symbol:] elif _time_lag_symbol < 0: _same_fiber_densities_time_lag = _same_fiber_densities[ -_time_lag_symbol:] _different_fiber_densities_time_lag = _different_fiber_densities[: _time_lag_symbol] else: _same_fiber_densities_time_lag = _same_fiber_densities _different_fiber_densities_time_lag = _different_fiber_densities _same_fiber_densities_filtered, _different_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_fiber_densities_time_lag, _different_fiber_densities_time_lag ) # ignore small arrays if len( _same_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _different_experiment): continue _different_correlation = compute_lib.correlation( compute_lib.derivative( _same_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _different_fiber_densities_filtered, _n=DERIVATIVE)) _different_time_lags_arrays[ _time_lag_index].append(_different_correlation) if _different_correlation > _different_highest_correlation: _different_highest_correlation = _different_correlation _different_highest_correlation_time_lag_index = _time_lag_index if _same_tuple not in _valid_tuples: _valid_tuples.append(_same_tuple) _different_time_lags_highest[ _different_highest_correlation_time_lag_index] += 1 print('Total tuples:', len(_valid_tuples)) return _same_correlation_vs_time_lag, _same_time_lags_arrays, _different_time_lags_arrays, \ _same_time_lags_highest, _different_time_lags_highest
def main(): _simulations = load.structured() _simulations = filtering.by_time_points_amount(_simulations, _time_points=TIME_POINTS) _simulations = filtering.by_categories(_simulations, _is_single_cell=True, _is_heterogeneity=False, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False) print('Total simulations:', len(_simulations)) _fiber_densities = compute_simulations_fiber_densities(_simulations) _y_arrays = [[] for _i in DERIVATIVES] for _index_1 in tqdm(range(len(_simulations)), desc='Simulations loop'): _simulation_1 = _simulations[_index_1] _cell_1_fiber_densities = \ [_fiber_densities[(_simulation_1, _direction)] for _direction in ['left', 'right', 'up', 'down']] _cell_1_fiber_densities = np.mean(_cell_1_fiber_densities, axis=0) for _index_2 in range(_index_1 + 1, len(_simulations)): _simulation_2 = _simulations[_index_2] _cell_2_fiber_densities = \ [_fiber_densities[(_simulation_2, _direction)] for _direction in ['left', 'right', 'up', 'down']] _cell_2_fiber_densities = np.mean(_cell_2_fiber_densities, axis=0) for _derivative_index, _derivative in enumerate(DERIVATIVES): _y_arrays[_derivative_index].append( compute_lib.correlation( compute_lib.derivative(_cell_1_fiber_densities, _n=_derivative), compute_lib.derivative(_cell_2_fiber_densities, _n=_derivative))) print('Total points:', len(_y_arrays[0])) print('Wilcoxon around the zero') for _y_array, _derivative in zip(_y_arrays, DERIVATIVES): print('Derivative:', _derivative, wilcoxon(_y_array)) # plot _colors_array = config.colors(3) _fig = go.Figure(data=[ go.Box(y=_y, name=_derivative, boxpoints='all', jitter=1, pointpos=0, line={'width': 1}, fillcolor='white', marker={ 'size': 10, 'color': _color }, opacity=0.7, showlegend=False) for _y, _derivative, _color in zip( _y_arrays, DERIVATIVES_TEXT, _colors_array) ], layout={ 'xaxis': { 'title': 'Fiber density derivative', 'zeroline': False }, 'yaxis': { 'title': 'Correlation', 'range': [-1, 1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot')
def compute_data(_arguments): _offset_y_index, _offset_y, _offset_z_index, _offset_z = _arguments _same_correlations_array = [] _different_correlations_array = [] for _experiment in _tuples_by_experiment: _experiment_tuples = _tuples_by_experiment[_experiment] for _same_index in range(len(_experiment_tuples)): _same_tuple = _experiment_tuples[_same_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _offset_y, _offset_z, 'left_cell')] _same_right_cell_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _offset_y, _offset_z, 'right_cell')] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities) _same_left_cell_fiber_densities_filtered, _same_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_left_cell_fiber_densities, _same_right_cell_fiber_densities ) # ignore small arrays if len(_same_left_cell_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _same_experiment): continue _same_correlation = compute_lib.correlation( compute_lib.derivative( _same_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _same_right_cell_fiber_densities_filtered, _n=DERIVATIVE)) for _different_index in range(len(_experiment_tuples)): if _same_index != _different_index: _different_tuple = _experiment_tuples[_different_index] _different_experiment, _different_series, _different_group = _different_tuple for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _offset_y, _offset_z, _same_cell_id)] _different_fiber_densities = _experiments_fiber_densities[ (_different_experiment, _different_series, _different_group, _offset_y, _offset_z, _different_cell_id)] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'] [_different_cell_id], _different_fiber_densities) _same_fiber_densities_filtered, _different_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_fiber_densities, _different_fiber_densities ) # ignore small arrays if len(_same_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _different_experiment): continue _different_correlation = compute_lib.correlation( compute_lib.derivative( _same_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _different_fiber_densities_filtered, _n=DERIVATIVE)) _same_correlations_array.append(_same_correlation) _different_correlations_array.append( _different_correlation) # compute fraction _annotation = None _same_minus_different = np.array(_same_correlations_array) - np.array( _different_correlations_array) _same_count = len(_same_minus_different[_same_minus_different > 0]) if len(_same_minus_different) > 0: _same_fraction = round(_same_count / len(_same_minus_different), 10) _wilcoxon = wilcoxon(_same_minus_different) _p_value = _wilcoxon[1] if _p_value > 0.05: _annotation = { 'text': 'x', 'showarrow': False, 'x': _offset_z, 'y': _offset_y, 'font': { 'size': 6, 'color': 'red' } } else: _same_fraction = None return _offset_y_index, _offset_z_index, _same_fraction, _annotation
def main(): _experiments = all_experiments() _experiments = filtering.by_categories(_experiments=_experiments, _is_single_cell=True, _is_high_temporal_resolution=False, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_time_frames_amount( _tuples, compute.density_time_frame(_experiments[0])) _tuples = filtering.by_main_cell(_tuples) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _time_frame = compute.density_time_frame(_experiment) for _direction in ['left', 'right']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': 'cell', 'direction': _direction, 'time_points': _time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'direction']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _tuples = organize.by_single_cell_id(_tuples) print('Total experiments:', len(_tuples)) _kpss_y_arrays = [[] for _i in DERIVATIVES] _adf_y_arrays = [[] for _i in DERIVATIVES] for _tuple in tqdm(_tuples, desc='Experiments loop'): _experiment, _series_id, _cell_id = _tuple _cell_fiber_densities = compute_single_cell_mean( _experiment=_experiment, _series_id=_series_id, _cell_tuples=_tuples[_tuple], _windows_dictionary=_windows_dictionary, _fiber_densities=_fiber_densities) for _derivative_index, _derivative in enumerate(DERIVATIVES): _cell_fiber_densities_derivative = compute_lib.derivative( _cell_fiber_densities, _n=_derivative) _, _kpss_p_value, _, _ = kpss(_cell_fiber_densities_derivative, nlags='legacy') _kpss_y_arrays[_derivative_index].append(_kpss_p_value) _, _adf_p_value, _, _, _, _ = adfuller( _cell_fiber_densities_derivative) _adf_y_arrays[_derivative_index].append(_adf_p_value) print('Total cells:', len(_kpss_y_arrays[0])) # print results print('KPSS:') for _derivative_index, _derivative in enumerate(DERIVATIVES): _stationary_count = len([ _value for _value in _kpss_y_arrays[_derivative_index] if _value > 0.05 ]) print( 'Derivative:', _derivative, 'Stationary:', str(_stationary_count / len(_kpss_y_arrays[_derivative_index]) * 100) + '%') print('ADF:') for _derivative_index, _derivative in enumerate(DERIVATIVES): _stationary_count = len([ _value for _value in _adf_y_arrays[_derivative_index] if _value < 0.05 ]) print( 'Derivative:', _derivative, 'Stationary:', str(_stationary_count / len(_adf_y_arrays[_derivative_index]) * 100) + '%') # plot _colors_array = config.colors(3) for _test_name, _y_title, _y_tickvals, _p_value_line, _y_arrays in \ zip( ['kpss', 'adf'], ['KPSS test p-value', 'ADF test p-value'], [[0.05, 0.1], [0.05, 1]], [0.05, 0.05], [_kpss_y_arrays, _adf_y_arrays] ): _fig = go.Figure(data=[ go.Box(y=_y, name=_derivative, boxpoints='all', jitter=1, pointpos=0, line={'width': 1}, fillcolor='white', marker={ 'size': 10, 'color': _color }, opacity=0.7, showlegend=False) for _y, _derivative, _color in zip( _y_arrays, DERIVATIVES_TEXT, _colors_array) ], layout={ 'xaxis': { 'title': 'Fiber density derivative', 'zeroline': False }, 'yaxis': { 'title': _y_title, 'zeroline': False, 'tickmode': 'array', 'tickvals': _y_tickvals }, 'shapes': [{ 'type': 'line', 'x0': DERIVATIVES[0] - 0.75, 'y0': _p_value_line, 'x1': DERIVATIVES[-1] + 0.75, 'y1': _p_value_line, 'line': { 'color': 'red', 'width': 2, 'dash': 'dash' } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_' + _test_name)
def compute_fiber_densities(_alpha=1, _beta=1, _low_connectivity=False): _simulations = load.structured() _simulations = filtering.by_time_points_amount( _simulations, TIME_POINT[_low_connectivity]) _simulations = filtering.by_categories( _simulations, _is_single_cell=False, _is_heterogeneity=True, _is_low_connectivity=_low_connectivity, _is_causality=True, _is_dominant_passive=False, _is_fibrin=False) _simulations = filtering.by_causality(_simulations, _alpha=_alpha, _beta=_beta) _simulations = filtering.by_pair_distance(_simulations, _distance=PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _arguments = [] for _simulation in _simulations: for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'simulation': _simulation, 'length_x': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': TIME_POINT[_low_connectivity] }) _fiber_densities = {} with Pool(CPUS_TO_USE) as _p: for _keys, _value in tqdm(_p.imap_unordered( compute.window_fiber_density_by_time, _arguments), total=len(_arguments), desc='Computing windows & fiber densities'): _fiber_densities[(_keys['simulation'], _keys['cell_id'])] = _value _p.close() _p.join() _same_correlation_vs_time_lag = {} _same_time_lags_arrays = [[] for _i in TIME_LAGS] _different_time_lags_arrays = [[] for _i in TIME_LAGS] _same_time_lags_highest = [0 for _i in TIME_LAGS] _different_time_lags_highest = [0 for _i in TIME_LAGS] for _same_index in tqdm(range(len(_simulations)), desc='Main loop'): _same_simulation = _simulations[_same_index] _same_left_cell_fiber_densities = _fiber_densities[(_same_simulation, 'left_cell')] _same_right_cell_fiber_densities = _fiber_densities[(_same_simulation, 'right_cell')] # time lag _same_highest_correlation = -1.1 _same_highest_correlation_time_lag_index = 0 _same_correlation_vs_time_lag[_same_simulation] = [] for _time_lag_index, _time_lag in enumerate(TIME_LAGS): if _time_lag > 0: _same_left_cell_fiber_densities_time_lag = _same_left_cell_fiber_densities[: -_time_lag] _same_right_cell_fiber_densities_time_lag = _same_right_cell_fiber_densities[ _time_lag:] elif _time_lag < 0: _same_left_cell_fiber_densities_time_lag = _same_left_cell_fiber_densities[ -_time_lag:] _same_right_cell_fiber_densities_time_lag = _same_right_cell_fiber_densities[: _time_lag] else: _same_left_cell_fiber_densities_time_lag = _same_left_cell_fiber_densities _same_right_cell_fiber_densities_time_lag = _same_right_cell_fiber_densities _same_correlation = compute_lib.correlation( compute_lib.derivative( _same_left_cell_fiber_densities_time_lag, _n=DERIVATIVE), compute_lib.derivative( _same_right_cell_fiber_densities_time_lag, _n=DERIVATIVE)) _same_time_lags_arrays[_time_lag_index].append(_same_correlation) _same_correlation_vs_time_lag[_same_simulation].append( _same_correlation) if _same_correlation > _same_highest_correlation: _same_highest_correlation = _same_correlation _same_highest_correlation_time_lag_index = _time_lag_index _same_time_lags_highest[_same_highest_correlation_time_lag_index] += 1 for _different_index in range(len(_simulations)): if _same_index != _different_index: _different_simulation = _simulations[_different_index] for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = \ _fiber_densities[(_same_simulation, _same_cell_id)] _different_fiber_densities = \ _fiber_densities[(_different_simulation, _different_cell_id)] # time lag _different_highest_correlation = -1.1 _different_highest_correlation_time_lag_index = 0 for _time_lag_index, _time_lag in enumerate(TIME_LAGS): if _time_lag > 0: _same_fiber_densities_time_lag = _same_fiber_densities[: -_time_lag] _different_fiber_densities_time_lag = _different_fiber_densities[ _time_lag:] elif _time_lag < 0: _same_fiber_densities_time_lag = _same_fiber_densities[ -_time_lag:] _different_fiber_densities_time_lag = _different_fiber_densities[: _time_lag] else: _same_fiber_densities_time_lag = _same_fiber_densities _different_fiber_densities_time_lag = _different_fiber_densities _different_correlation = compute_lib.correlation( compute_lib.derivative( _same_fiber_densities_time_lag, _n=DERIVATIVE), compute_lib.derivative( _different_fiber_densities_time_lag, _n=DERIVATIVE)) _different_time_lags_arrays[_time_lag_index].append( _different_correlation) if _different_correlation > _different_highest_correlation: _different_highest_correlation = _different_correlation _different_highest_correlation_time_lag_index = _time_lag_index _different_time_lags_highest[ _different_highest_correlation_time_lag_index] += 1 return _same_correlation_vs_time_lag, _same_time_lags_arrays, _different_time_lags_arrays, \ _same_time_lags_highest, _different_time_lags_highest
def main(_high_temporal_resolution=True, _offset_y=0.5): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_real_pairs(_tuples) # _tuples = filtering.by_band(_tuples) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id, _direction in product(['left_cell', 'right_cell'], ['inside', 'outside']): _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': _offset_y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': _direction, 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id', 'direction']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _inner_correlations = [] _outer_correlations = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _left_inner_tuple = (_experiment, _series_id, _group, 'left_cell', 'inside') _left_outer_tuple = (_experiment, _series_id, _group, 'left_cell', 'outside') _right_inner_tuple = (_experiment, _series_id, _group, 'right_cell', 'inside') _right_outer_tuple = (_experiment, _series_id, _group, 'right_cell', 'outside') if _left_inner_tuple not in _windows_dictionary or _left_outer_tuple not in _windows_dictionary or \ _right_inner_tuple not in _windows_dictionary or _right_outer_tuple not in _windows_dictionary: continue _properties = load.group_properties(_experiment, _series_id, _group) _left_inner_fiber_densities = _experiments_fiber_densities[ _left_inner_tuple] _left_outer_fiber_densities = _experiments_fiber_densities[ _left_outer_tuple] _right_inner_fiber_densities = _experiments_fiber_densities[ _right_inner_tuple] _right_outer_fiber_densities = _experiments_fiber_densities[ _right_outer_tuple] _left_inner_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_inner_fiber_densities) _left_outer_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_outer_fiber_densities) _right_inner_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_inner_fiber_densities) _right_outer_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_outer_fiber_densities) _left_inner_fiber_densities, _right_inner_fiber_densities = \ compute.longest_same_indices_shared_in_borders_sub_array( _left_inner_fiber_densities, _right_inner_fiber_densities) _left_outer_fiber_densities, _right_outer_fiber_densities = \ compute.longest_same_indices_shared_in_borders_sub_array( _left_outer_fiber_densities, _right_outer_fiber_densities) # ignore small arrays _minimum_time_frame_for_correlation = compute.minimum_time_frames_for_correlation( _experiment) if len(_left_inner_fiber_densities) < _minimum_time_frame_for_correlation or \ len(_left_outer_fiber_densities) < _minimum_time_frame_for_correlation: continue _inner_correlations.append( compute_lib.correlation( compute_lib.derivative(_left_inner_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_right_inner_fiber_densities, _n=DERIVATIVE))) _outer_correlations.append( compute_lib.correlation( compute_lib.derivative(_left_outer_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_right_outer_fiber_densities, _n=DERIVATIVE))) print('Total pairs:', len(_inner_correlations)) print( 'Pearson correlation:', compute_lib.correlation(_inner_correlations, _outer_correlations, _with_p_value=True)) print('Wilcoxon of inner around the zero:', wilcoxon(_inner_correlations)) print('Wilcoxon of outer around the zero:', wilcoxon(_outer_correlations)) _inner_minus_outer = np.array(_inner_correlations) - np.array( _outer_correlations) print('Wilcoxon of inner minus outer:', wilcoxon(_inner_minus_outer)) print('Higher inner amount:', (_inner_minus_outer > 0).sum() / len(_inner_minus_outer)) # plot _fig = go.Figure(data=go.Scatter(x=_inner_correlations, y=_outer_correlations, mode='markers', marker={ 'size': 5, 'color': '#ea8500' }, showlegend=False), layout={ 'xaxis': { 'title': 'Inner correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Outer correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'shapes': [{ 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_high_' + str(_high_temporal_resolution) + '_offset_y_' + str(_offset_y))
def main(): _experiments = all_experiments() _experiments = filtering.by_categories(_experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=False, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_time_frames_amount( _tuples, compute.density_time_frame(_experiments[0])) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _kpss_y_arrays = [[] for _i in DERIVATIVES] _adf_y_arrays = [[] for _i in DERIVATIVES] for _tuple in tqdm(_tuples, desc='Experiments loop'): _experiment, _series_id, _group = _tuple _properties = load.group_properties(_experiment, _series_id, _group) _left_cell_fiber_densities = _experiments_fiber_densities[( _experiment, _series_id, _group, 'left_cell')] _left_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_cell_fiber_densities) _right_cell_fiber_densities = _experiments_fiber_densities[( _experiment, _series_id, _group, 'right_cell')] _right_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_cell_fiber_densities) if not OUT_OF_BOUNDARIES: _left_cell_fiber_densities = \ compute.longest_fiber_densities_ascending_sequence(_left_cell_fiber_densities) _right_cell_fiber_densities = \ compute.longest_fiber_densities_ascending_sequence(_right_cell_fiber_densities) else: _left_cell_fiber_densities = [ _fiber_density[0] for _fiber_density in _left_cell_fiber_densities ] _right_cell_fiber_densities = [ _fiber_density[0] for _fiber_density in _right_cell_fiber_densities ] # ignore small arrays _minimum_time_frames_for_correlation = compute.minimum_time_frames_for_correlation( _experiment) if len(_left_cell_fiber_densities) < _minimum_time_frames_for_correlation or \ len(_right_cell_fiber_densities) < _minimum_time_frames_for_correlation: continue for _derivative_index, _derivative in enumerate(DERIVATIVES): for _cell_fiber_densities in [ _left_cell_fiber_densities, _right_cell_fiber_densities ]: _cell_fiber_densities_derivative = compute_lib.derivative( _cell_fiber_densities, _n=_derivative) _, _kpss_p_value, _, _ = kpss(_cell_fiber_densities_derivative, nlags='legacy') _kpss_y_arrays[_derivative_index].append(_kpss_p_value) _, _adf_p_value, _, _, _, _ = adfuller( _cell_fiber_densities_derivative) _adf_y_arrays[_derivative_index].append(_adf_p_value) print('Total pairs:', len(_kpss_y_arrays[0]) / 2) # print results print('KPSS:') for _derivative_index, _derivative in enumerate(DERIVATIVES): _stationary_count = len([ _value for _value in _kpss_y_arrays[_derivative_index] if _value > 0.05 ]) print( 'Derivative:', _derivative, 'Stationary:', str(_stationary_count / len(_kpss_y_arrays[_derivative_index]) * 100) + '%') print('ADF:') for _derivative_index, _derivative in enumerate(DERIVATIVES): _stationary_count = len([ _value for _value in _adf_y_arrays[_derivative_index] if _value < 0.05 ]) print( 'Derivative:', _derivative, 'Stationary:', str(_stationary_count / len(_adf_y_arrays[_derivative_index]) * 100) + '%') # plot _colors_array = config.colors(3) for _test_name, _y_title, _y_tickvals, _p_value_line, _y_arrays in \ zip( ['kpss', 'adf'], ['KPSS test p-value', 'ADF test p-value'], [[0.05, 0.1], [0.05, 1]], [0.05, 0.05], [_kpss_y_arrays, _adf_y_arrays] ): _fig = go.Figure(data=[ go.Box(y=_y, name=_derivative, boxpoints='all', jitter=1, pointpos=0, line={'width': 1}, fillcolor='white', marker={ 'size': 10, 'color': _color }, opacity=0.7, showlegend=False) for _y, _derivative, _color in zip( _y_arrays, DERIVATIVES_TEXT, _colors_array) ], layout={ 'xaxis': { 'title': 'Fiber density derivative', 'zeroline': False }, 'yaxis': { 'title': _y_title, 'zeroline': False, 'tickmode': 'array', 'tickvals': _y_tickvals }, 'shapes': [{ 'type': 'line', 'x0': DERIVATIVES[0] - 0.75, 'y0': _p_value_line, 'x1': DERIVATIVES[-1] + 0.75, 'y1': _p_value_line, 'line': { 'color': 'red', 'width': 2, 'dash': 'dash' } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_' + _test_name)
def main(): _simulations = load.structured() _simulations = filtering.by_time_points_amount( _simulations, _time_points=SIMULATIONS_TIME_POINTS) _simulations = filtering.by_categories(_simulations, _is_single_cell=False, _is_heterogeneity=False, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False) _simulations = filtering.by_pair_distance(_simulations, _distance=PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _fiber_densities = compute_fiber_densities(_simulations) _kpss_y_arrays = [[] for _i in DERIVATIVES] _adf_y_arrays = [[] for _i in DERIVATIVES] for _simulation in tqdm(_simulations, desc='Simulations loop'): for _cell_id in ['left_cell', 'right_cell']: _cell_fiber_densities = _fiber_densities[(_simulation, _cell_id)] for _derivative_index, _derivative in enumerate(DERIVATIVES): _cell_fiber_densities_derivative = compute_lib.derivative( _cell_fiber_densities, _n=_derivative) with warnings.catch_warnings(): warnings.simplefilter('ignore', category=InterpolationWarning) _, _kpss_p_value, _, _ = kpss( _cell_fiber_densities_derivative, nlags='legacy') _kpss_y_arrays[_derivative_index].append(_kpss_p_value) _, _adf_p_value, _, _, _, _ = adfuller( _cell_fiber_densities_derivative) _adf_y_arrays[_derivative_index].append(_adf_p_value) # print results print('KPSS:') for _derivative_index, _derivative in enumerate(DERIVATIVES): _stationary_count = len([ _value for _value in _kpss_y_arrays[_derivative_index] if _value > 0.05 ]) print( 'Derivative:', _derivative, 'Stationary:', str(_stationary_count / len(_kpss_y_arrays[_derivative_index]) * 100) + '%') print('ADF:') for _derivative_index, _derivative in enumerate(DERIVATIVES): _stationary_count = len([ _value for _value in _adf_y_arrays[_derivative_index] if _value < 0.05 ]) print( 'Derivative:', _derivative, 'Stationary:', str(_stationary_count / len(_adf_y_arrays[_derivative_index]) * 100) + '%') # plot _colors_array = config.colors(3) for _test_name, _y_title, _y_tickvals, _p_value_line, _y_arrays in \ zip( ['kpss', 'adf'], ['KPSS test p-value', 'ADF test p-value'], [[0.05, 0.1], [0.05, 1]], [0.05, 0.05], [_kpss_y_arrays, _adf_y_arrays] ): _fig = go.Figure(data=[ go.Box(y=_y, name=_derivative, boxpoints='all', jitter=1, pointpos=0, line={'width': 1}, fillcolor='white', marker={ 'size': 10, 'color': _color }, opacity=0.7, showlegend=False) for _y, _derivative, _color in zip( _y_arrays, DERIVATIVES_TEXT, _colors_array) ], layout={ 'xaxis': { 'title': 'Fiber density derivative', 'zeroline': False }, 'yaxis': { 'title': _y_title, 'zeroline': False, 'tickmode': 'array', 'tickvals': _y_tickvals }, 'shapes': [{ 'type': 'line', 'x0': DERIVATIVES[0] - 0.75, 'y0': _p_value_line, 'x1': DERIVATIVES[-1] + 0.75, 'y1': _p_value_line, 'line': { 'color': 'red', 'width': 2, 'dash': 'dash' } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_' + _test_name)
def main(_early_time_frames=True): _experiments = all_experiments() _experiments = filtering.by_categories(_experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=False, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside' }) _windows_dictionary, _windows_to_compute = compute.windows( _arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute) _heatmap_fiber = [] _heatmap_fiber_change = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _series_normalization = load.normalization_series_file_data( _experiment, _series_id) for _cell_id in ['left_cell', 'right_cell']: _fiber_densities_by_time = [ _fiber_densities[_tuple] for _tuple in _windows_dictionary[(_experiment, _series_id, _group, _cell_id)] ] _cell_fiber_densities = \ _fiber_densities_by_time[TIME_FRAMES[OFFSET_Y]['early'][0] if _early_time_frames else TIME_FRAMES[OFFSET_Y]['late'][0]: TIME_FRAMES[OFFSET_Y]['early'][1] if _early_time_frames else TIME_FRAMES[OFFSET_Y]['late'][1]] _properties = load.group_properties(_experiment, _series_id, _group) _cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids'][_cell_id], _cell_fiber_densities) _cell_fiber_densities = compute.longest_fiber_densities_ascending_sequence( _cell_fiber_densities) # fix if found nan if True in np.isnan(_cell_fiber_densities): _cell_fiber_densities = _cell_fiber_densities[:np.where( np.isnan(_cell_fiber_densities))[0][0]] # not enough data if len(_cell_fiber_densities) < DERIVATIVE + 1: continue _z_score_fiber_density = libs.compute_lib.z_score_array( _array=_cell_fiber_densities, _average=_series_normalization['average'], _std=_series_normalization['std']) if _experiment in ['SN41', 'SN44']: for _start_index in [0, 1, 2]: _heatmap_fiber += _z_score_fiber_density[_start_index::3][ DERIVATIVE:] _heatmap_fiber_change += compute_lib.derivative( _z_score_fiber_density[_start_index::3], _n=DERIVATIVE) else: _heatmap_fiber += _z_score_fiber_density[DERIVATIVE:] _heatmap_fiber_change += compute_lib.derivative( _z_score_fiber_density, _n=DERIVATIVE) print( compute_lib.correlation(_heatmap_fiber, _heatmap_fiber_change, _with_p_value=True)) if PLOT: _y_shape = int(round((Y_LABELS_END - Y_LABELS_START) * Y_BINS)) _x_shape = int(round((X_LABELS_END - X_LABELS_START) * X_BINS)) _total_points = 0 _z_array = np.zeros(shape=(_y_shape, _x_shape)) for _y, _x in zip(_heatmap_fiber_change, _heatmap_fiber): _y_rounded, _x_rounded = int(round(_y * Y_BINS)), int( round(_x * X_BINS)) _y_index, _x_index = int(_y_rounded - Y_LABELS_START * Y_BINS), int(_x_rounded - X_LABELS_START * X_BINS) if 0 <= _y_index < _z_array.shape[ 0] and 0 <= _x_index < _z_array.shape[1]: _z_array[_y_index][_x_index] += 1 _total_points += 1 _z_array = _z_array / _total_points if not CONDITIONAL_NORMALIZATION: _z_array[_z_array == 0] = None else: _z_array_plot = np.zeros(shape=np.array(_z_array).shape) for _fiber_index, _fiber_density_z_score in enumerate(_z_array): _sum = np.sum(_fiber_density_z_score) for _change_index, _change_z_score in enumerate( _fiber_density_z_score): _z_array_plot[_fiber_index][_change_index] = ( _change_z_score / _sum) if _sum != 0 else 0 _z_array_plot[_z_array_plot == 0] = None _fig = go.Figure( data=go.Heatmap(x=np.arange(start=X_LABELS_START, stop=X_LABELS_END, step=1 / X_BINS), y=np.arange(start=Y_LABELS_START, stop=Y_LABELS_END, step=1 / Y_BINS), z=_z_array, colorscale='Viridis', colorbar={ 'tickmode': 'array', 'tickvals': [0, 0.025, 0.05], 'ticktext': ['0', 'Fraction', '0.05'], 'tickangle': -90 }, zmin=Z_MIN, zmax=Z_MAX[CONDITIONAL_NORMALIZATION]), layout={ 'xaxis': { 'title': 'Fiber densities z-score', 'zeroline': False }, 'yaxis': { 'title': 'Change in fiber<br>density (z-score)', 'zeroline': False }, 'shapes': [{ 'type': 'line', 'x0': X_LABELS_START, 'y0': Y_LABELS_START, 'x1': X_LABELS_END, 'y1': Y_LABELS_START, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': X_LABELS_START, 'y0': Y_LABELS_START, 'x1': X_LABELS_START, 'y1': Y_LABELS_END, 'line': { 'color': 'black', 'width': 2 } }] }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_early_' + str(_early_time_frames))
def compute_fiber_densities(_offset_y=0.5, _high_temporal_resolution=False): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) _tuples = filtering.by_band(_tuples) _tuples = filtering.by_real_fake_pairs(_tuples, _real_fake_pairs=False) _experiments_matched = organize.by_matched_real_and_fake(_tuples) print('Total matched pairs:', len(_experiments_matched)) _arguments = [] for _matched_tuple in _experiments_matched: for _tuple in _matched_tuple: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': _offset_y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows( _arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _tuples_by_experiment = organize.by_experiment(_tuples) # same (real, fake), different (real, fake) _correlations = [[[], []], [[], []]] _valid_real_tuples = [] for _experiment in _tuples_by_experiment: print('Experiment:', _experiment) _experiment_tuples = _tuples_by_experiment[_experiment] _experiments_matched = organize.by_matched_real_and_fake( _experiment_tuples) print('Matched pairs:', len(_experiments_matched)) for _same_index in tqdm(range(len(_experiments_matched)), desc='Main loop'): for _group_type_index in [0, 1]: _same_tuple = _experiments_matched[_same_index][ _group_type_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'left_cell') ] _same_right_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'right_cell') ] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities) _same_left_cell_fiber_densities_filtered, _same_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_left_cell_fiber_densities, _same_right_cell_fiber_densities ) # ignore small arrays if len(_same_left_cell_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _same_experiment): for _different_index in range(len(_experiments_matched)): if _same_index != _different_index: # for all combinations for _i in range(4): _correlations[0][_group_type_index].append( None) _correlations[1][_group_type_index].append( None) continue _same_correlation = compute_lib.correlation( compute_lib.derivative( _same_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _same_right_cell_fiber_densities_filtered, _n=DERIVATIVE)) for _different_index in range(len(_experiments_matched)): if _same_index != _different_index: _different_tuple = _experiments_matched[ _different_index][_group_type_index] _different_experiment, _different_series, _different_group = _different_tuple for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, _same_cell_id)] _different_fiber_densities = _experiments_fiber_densities[ (_different_experiment, _different_series, _different_group, _different_cell_id)] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'] [_different_cell_id], _different_fiber_densities) _same_fiber_densities_filtered, _different_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_fiber_densities, _different_fiber_densities ) # ignore small arrays if len( _same_fiber_densities_filtered ) < compute.minimum_time_frames_for_correlation( _different_experiment): _correlations[0][_group_type_index].append( None) _correlations[1][_group_type_index].append( None) continue _different_correlation = compute_lib.correlation( compute_lib.derivative( _same_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _different_fiber_densities_filtered, _n=DERIVATIVE)) _correlations[0][_group_type_index].append( _same_correlation) _correlations[1][_group_type_index].append( _different_correlation) if _group_type_index == 0 and _same_tuple not in _valid_real_tuples: _valid_real_tuples.append(_same_tuple) print('Total real tuples:', len(_valid_real_tuples)) _distances_from_y_equal_x = [[], []] _same_correlations, _different_correlations = _correlations _same_real_correlations, _same_fake_correlations = _same_correlations _different_real_correlations, _different_fake_correlations = _different_correlations for _same_real, _same_fake, _different_real, _different_fake in \ zip(_same_real_correlations, _same_fake_correlations, _different_real_correlations, _different_fake_correlations): # one of the correlations is none - not valid if None in [_same_real, _same_fake, _different_real, _different_fake]: continue for _group_type_index, _same, _different in \ zip([0, 1], [_same_real, _same_fake], [_different_real, _different_fake]): _point_distance = compute_lib.distance_from_a_point_to_a_line( _line=[-1, -1, 1, 1], _point=[_same, _different]) if _same > _different: _distances_from_y_equal_x[_group_type_index].append( _point_distance) else: _distances_from_y_equal_x[_group_type_index].append( -_point_distance) return _distances_from_y_equal_x
def main(): _arguments = [] for _tuple in TRIPLET: _experiment, _series_id, _group = _tuple _pair_distance = compute.pair_distance_in_cell_size_time_frame(_experiment, _series_id, _group, _time_frame=0) print(_tuple, 'pairs distance:', round(_pair_distance, 2)) _latest_time_frame = compute.latest_time_frame_before_overlapping(_experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _same_correlations_arrays = [[], [], []] _different_correlations_arrays = [[], [], []] _names_array = [] for _same_index in tqdm(range(3), desc='Main loop'): _same_tuple = TRIPLET[_same_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'left_cell') ] _same_right_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'right_cell') ] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities ) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities ) _same_left_cell_fiber_densities_filtered, _same_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_left_cell_fiber_densities, _same_right_cell_fiber_densities ) _same_correlation = compute_lib.correlation( compute_lib.derivative(_same_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_same_right_cell_fiber_densities_filtered, _n=DERIVATIVE) ) for _different_index in range(3): if _same_index != _different_index: _different_tuple = TRIPLET[_different_index] _different_experiment, _different_series, _different_group = \ _different_tuple for _same_cell_id, _different_cell_id in product(['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _same_cell_id )] _different_fiber_densities = _experiments_fiber_densities[( _different_experiment, _different_series, _different_group, _different_cell_id )] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group ) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities ) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'][_different_cell_id], _different_fiber_densities ) _same_fiber_densities_filtered, _different_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_fiber_densities, _different_fiber_densities ) _different_correlation = compute_lib.correlation( compute_lib.derivative(_same_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_different_fiber_densities_filtered, _n=DERIVATIVE) ) _same_correlations_arrays[_same_index].append(_same_correlation) _different_correlations_arrays[_same_index].append(_different_correlation) _names_array.append('Cells ' + _same_group.split('_')[1] + ' & ' + _same_group.split('_')[2]) print('Group:', TRIPLET[_same_index]) print('Points:', len(_same_correlations_arrays[_same_index])) _same_minus_different = \ np.array(_same_correlations_arrays[_same_index]) - np.array(_different_correlations_arrays[_same_index]) print('Wilcoxon of same minus different around the zero:') print(wilcoxon(_same_minus_different)) print('Higher same amount:', (_same_minus_different > 0).sum() / len(_same_minus_different)) print('Total points:', len(np.array(_same_correlations_arrays).flatten())) _same_minus_different = \ np.array(_same_correlations_arrays).flatten() - np.array(_different_correlations_arrays).flatten() print('Wilcoxon of same minus different around the zero:') print(wilcoxon(_same_minus_different)) print('Higher same amount:', (_same_minus_different > 0).sum() / len(_same_minus_different)) # plot _colors_array = config.colors(3) _fig = go.Figure( data=[ go.Scatter( x=_same_correlations_array, y=_different_correlations_array, name=_name, mode='markers', marker={ 'size': 15, 'color': _color }, opacity=0.7 ) for _same_correlations_array, _different_correlations_array, _name, _color in zip(_same_correlations_arrays, _different_correlations_arrays, _names_array, _colors_array) ], layout={ 'xaxis': { 'title': 'Same network correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Different network correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'legend': { 'xanchor': 'left', 'x': 0.1, 'yanchor': 'top', 'bordercolor': 'black', 'borderwidth': 2, 'bgcolor': 'white' }, 'shapes': [ { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } } ] } ) save.to_html( _fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot' )
def main(_low_connectivity=False): _simulations = load.structured() _simulations = filtering.by_time_points_amount( _simulations, TIME_POINT[_low_connectivity]) _simulations = filtering.by_categories( _simulations, _is_single_cell=False, _is_heterogeneity=True, _is_low_connectivity=_low_connectivity, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False) _simulations = filtering.by_heterogeneity(_simulations, _std=STD) _simulations = filtering.by_pair_distances(_simulations, _distances=PAIR_DISTANCE) print('Total simulations:', len(_simulations)) _fiber_densities = compute_fiber_densities(_simulations, _low_connectivity) _y_arrays = [[] for _i in PAIR_DISTANCE] for _distance_index, _distance in enumerate(PAIR_DISTANCE): _distance_simulations = filtering.by_pair_distance(_simulations, _distance=_distance) print('Distance:', _distance, 'Total simulations:', len(_distance_simulations)) _higher_same_counter = 0 for _same_index in tqdm(range(len(_distance_simulations)), desc='Main loop'): _same_simulation = _distance_simulations[_same_index] _same_left_cell_fiber_densities = _fiber_densities[( _same_simulation, 'left_cell')] _same_right_cell_fiber_densities = _fiber_densities[( _same_simulation, 'right_cell')] _same_correlation = compute_lib.correlation( compute_lib.derivative(_same_left_cell_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_same_right_cell_fiber_densities, _n=DERIVATIVE)) for _different_index in range(len(_distance_simulations)): if _same_index != _different_index: _different_simulation = _distance_simulations[ _different_index] for _same_cell_id, _different_cell_id in product( ['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = \ _fiber_densities[(_same_simulation, _same_cell_id)] _different_fiber_densities = \ _fiber_densities[(_different_simulation, _different_cell_id)] _different_correlation = compute_lib.correlation( compute_lib.derivative(_same_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_different_fiber_densities, _n=DERIVATIVE)) _point_distance = compute_lib.distance_from_a_point_to_a_line( _line=[-1, -1, 1, 1], _point=[_same_correlation, _different_correlation]) if _same_correlation > _different_correlation: _y_arrays[_distance_index].append(_point_distance) _higher_same_counter += 1 else: _y_arrays[_distance_index].append(-_point_distance) print('Total points:', len(_y_arrays[_distance_index])) print('Wilcoxon around the zero:') print(wilcoxon(_y_arrays[_distance_index])) print('Higher same amount:', _higher_same_counter / len(_y_arrays[_distance_index])) # plot _colors_array = config.colors(4) _fig = go.Figure(data=[ go.Box(y=_y, name=str(_distance), boxpoints=False, line={'width': 1}, marker={ 'size': 10, 'color': _color }, showlegend=False) for _y, _distance, _color in zip( _y_arrays, PAIR_DISTANCE, _colors_array) ], layout={ 'xaxis': { 'title': 'Pair distance (cell diameter)', 'zeroline': False, 'tickmode': 'array', 'tickvals': PAIR_DISTANCE, 'type': 'category' }, 'yaxis': { 'title': 'Same minus different correlation', 'range': [-1, 1.1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot')
def main(): _simulations = load.structured() _simulations = filtering.by_time_points_amount(_simulations, TIME_POINTS) _simulations = filtering.by_categories( _simulations, _is_single_cell=False, _is_heterogeneity=True, _is_low_connectivity=False, _is_causality=False, _is_dominant_passive=False, _is_fibrin=False ) _simulations = filtering.by_pair_distance(_simulations, _distance=PAIR_DISTANCE) _simulations = filtering.by_heterogeneity(_simulations, _std=STD) print('Total simulations:', len(_simulations)) _fiber_densities = compute_fiber_densities(_simulations) _window_distances_communicating = [[] for _i in OFFSETS_X] _window_distances_non_communicating = [[] for _i in OFFSETS_X] # window distances loop for _window_distance_index, _window_distance in enumerate(OFFSETS_X): print('Window distance:', _window_distance) # communicating loop for _simulation in tqdm(_simulations, desc='Communicating loop'): _left_cell_fiber_densities = _fiber_densities[(_simulation, _window_distance, 'left_cell')] _right_cell_fiber_densities = _fiber_densities[(_simulation, _window_distance, 'right_cell')] _correlation = compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_right_cell_fiber_densities, _n=DERIVATIVE) ) _window_distances_communicating[_window_distance_index].append(_correlation) # non-communicating loop _simulations_indices = range(len(_simulations)) for _simulation_1_index in tqdm(_simulations_indices, desc='Non-communicating pairs loop'): _simulation_1 = _simulations[_simulation_1_index] for _simulation_2_index in _simulations_indices[_simulation_1_index + 1:]: _simulation_2 = _simulations[_simulation_2_index] for _simulation_1_cell_id, _simulation_2_cell_id in product(['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _simulation_1_fiber_densities = \ _fiber_densities[(_simulation_1, _window_distance, _simulation_1_cell_id)] _simulation_2_fiber_densities = \ _fiber_densities[(_simulation_2, _window_distance, _simulation_2_cell_id)] _correlation = compute_lib.correlation( compute_lib.derivative(_simulation_1_fiber_densities, _n=DERIVATIVE), compute_lib.derivative(_simulation_2_fiber_densities, _n=DERIVATIVE) ) _window_distances_non_communicating[_window_distance_index].append(_correlation) # rank sums print('Wilcoxon rank-sum tests between communicating and non-communicating:', ranksums(_window_distances_communicating[_window_distance_index], _window_distances_non_communicating[_window_distance_index])) # plot _data = [] _colors_array = config.colors(2) for _communicating, _communicating_text, _pair_distances, _color in \ zip([True, False], ['Communicating', 'Non-communicating'], [_window_distances_communicating, _window_distances_non_communicating], _colors_array): _y = [] _x = [] for _window_distance_index, _window_distance in enumerate(OFFSETS_X): _y += _pair_distances[_window_distance_index] _x += [_window_distance for _i in _pair_distances[_window_distance_index]] _data.append( go.Box( y=_y, x=_x, name=_communicating_text, boxpoints='all' if _communicating else False, jitter=1, pointpos=0, line={ 'width': 1 }, fillcolor='white', marker={ 'size': 10, 'color': _color }, opacity=0.7 ) ) _fig = go.Figure( data=_data, layout={ 'xaxis': { 'title': 'Window distance (cell diameter)', 'zeroline': False, 'tickmode': 'array', 'tickvals': OFFSETS_X, 'type': 'category' }, 'yaxis': { 'title': 'Correlation', 'range': [-1, 1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'boxmode': 'group', 'legend': { 'xanchor': 'right', 'yanchor': 'top', 'bordercolor': 'black', 'borderwidth': 2 } } ) save.to_html( _fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot' )
def main(): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=False, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False ) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples) _tuples = filtering.by_time_frames_amount(_tuples, MINIMUM_TIME_FRAMES) _tuples = filtering.by_pair_distance_range(_tuples, [MINIMUM_PAIR_DISTANCE, sys.maxsize]) _triplets = filtering.by_triplets(_tuples) print('Total triplets:', len(_triplets)) _arguments = [] for _triplet in _triplets: for _tuple in _triplet: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping(_experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } _same_correlations_arrays = [[] for _i in _triplets] _different_correlations_arrays = [[] for _i in _triplets] _names_array = [] for _triplet_index, _triplet in enumerate(_triplets): for _same_index in tqdm(range(len(_triplet)), desc='Main loop'): _same_tuple = _triplet[_same_index] _same_experiment, _same_series, _same_group = _same_tuple _same_left_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'left_cell') ] _same_right_cell_fiber_densities = \ _experiments_fiber_densities[ (_same_experiment, _same_series, _same_group, 'right_cell') ] _same_properties = \ load.group_properties(_same_experiment, _same_series, _same_group) _same_left_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['left_cell'], _same_left_cell_fiber_densities ) _same_right_cell_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids']['right_cell'], _same_right_cell_fiber_densities ) _same_left_cell_fiber_densities_filtered, _same_right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_left_cell_fiber_densities, _same_right_cell_fiber_densities ) _same_correlation = compute_lib.correlation( compute_lib.derivative(_same_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_same_right_cell_fiber_densities_filtered, _n=DERIVATIVE) ) for _different_index in range(len(_triplet)): if _same_index != _different_index: _different_tuple = _triplet[_different_index] _different_experiment, _different_series, _different_group = \ _different_tuple for _same_cell_id, _different_cell_id in product(['left_cell', 'right_cell'], ['left_cell', 'right_cell']): _same_fiber_densities = _experiments_fiber_densities[( _same_experiment, _same_series, _same_group, _same_cell_id )] _different_fiber_densities = _experiments_fiber_densities[( _different_experiment, _different_series, _different_group, _different_cell_id )] _different_properties = load.group_properties( _different_experiment, _different_series, _different_group ) _same_fiber_densities = compute.remove_blacklist( _same_experiment, _same_series, _same_properties['cells_ids'][_same_cell_id], _same_fiber_densities ) _different_fiber_densities = compute.remove_blacklist( _different_experiment, _different_series, _different_properties['cells_ids'][_different_cell_id], _different_fiber_densities ) _same_fiber_densities_filtered, _different_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _same_fiber_densities, _different_fiber_densities ) _different_correlation = compute_lib.correlation( compute_lib.derivative(_same_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative(_different_fiber_densities_filtered, _n=DERIVATIVE) ) _same_correlations_arrays[_triplet_index].append(_same_correlation) _different_correlations_arrays[_triplet_index].append(_different_correlation) _names_array.append('Triplet #' + str(_triplet_index + 1)) print('Total points:', len(np.array(_same_correlations_arrays).flatten())) _same_minus_different = \ np.array(_same_correlations_arrays).flatten() - np.array(_different_correlations_arrays).flatten() print('Wilcoxon of same minus different around the zero:') print(wilcoxon(_same_minus_different)) print('Higher same amount:', (_same_minus_different > 0).sum() / len(_same_minus_different)) # plot _colors_array = ['green', 'blue', config.colors(1)] _fig = go.Figure( data=[ go.Scatter( x=_same_correlations_array, y=_different_correlations_array, name=_name, mode='markers', marker={ 'size': 15, 'color': _color }, opacity=0.7 ) for _same_correlations_array, _different_correlations_array, _name, _color in zip(_same_correlations_arrays, _different_correlations_arrays, _names_array, _colors_array) ], layout={ 'xaxis': { 'title': 'Same network correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'yaxis': { 'title': 'Different network correlation', 'zeroline': False, 'range': [-1.1, 1.2], 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] }, 'legend': { 'xanchor': 'left', 'x': 0.1, 'yanchor': 'top', 'bordercolor': 'black', 'borderwidth': 2, 'bgcolor': 'white' }, 'shapes': [ { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': -1, 'y1': 1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': -1, 'line': { 'color': 'black', 'width': 2 } }, { 'type': 'line', 'x0': -1, 'y0': -1, 'x1': 1, 'y1': 1, 'line': { 'color': 'red', 'width': 2 } } ] } ) save.to_html( _fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot' )
def main(_high_temporal_resolution=True): _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=_high_temporal_resolution, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_time_frames_amount( _tuples, _time_frames=MOVING_WINDOW_LENGTH[_high_temporal_resolution]) _tuples = filtering.by_pair_distance_range( _tuples, _distance_range=PAIR_DISTANCE_RANGE) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping( _experiment, _series_id, _group, OFFSET_X) for _cell_id in ['left_cell', 'right_cell']: _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': 'inside', 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = compute.windows( _arguments, _keys=['experiment', 'series_id', 'group', 'cell_id']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } for _tuple in _tuples: _correlations = [] _experiment, _series_id, _group = _tuple _left_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'left_cell')] _right_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'right_cell')] _properties = load.group_properties(_experiment, _series_id, _group) _left_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_cell_fiber_densities) _right_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_cell_fiber_densities) for _start_time_frame in \ range(0, END_TIME_FRAME[_high_temporal_resolution], TIME_FRAME_STEP[_high_temporal_resolution]): _left_cell_fiber_densities_window = _left_cell_fiber_densities[ _start_time_frame:_start_time_frame + MOVING_WINDOW_LENGTH[_high_temporal_resolution]] _right_cell_fiber_densities_window = _right_cell_fiber_densities[ _start_time_frame:_start_time_frame + MOVING_WINDOW_LENGTH[_high_temporal_resolution]] _left_cell_fiber_densities_filtered, _right_cell_fiber_densities_filtered = \ compute.longest_same_indices_shared_in_borders_sub_array( _left_cell_fiber_densities_window, _right_cell_fiber_densities_window) # ignore small arrays if len(_left_cell_fiber_densities_filtered ) < MOVING_WINDOW_LENGTH[_high_temporal_resolution]: _correlations.append(None) continue _correlations.append( compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities_filtered, _n=DERIVATIVE), compute_lib.derivative( _right_cell_fiber_densities_filtered, _n=DERIVATIVE))) # plot _temporal_resolution = compute.temporal_resolution_in_minutes( _experiment) _fig = go.Figure(data=go.Scatter( x=np.arange(start=0, stop=len(_correlations), step=1) * _temporal_resolution * TIME_FRAME_STEP[_high_temporal_resolution], y=_correlations, mode='lines+markers', line={'dash': 'solid'}), layout={ 'xaxis': { 'title': 'Window start time (minutes)', 'zeroline': False }, 'yaxis': { 'title': 'Inner correlation', 'zeroline': False } }) save.to_html(_fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_' + str(_experiment) + '_' + str(_series_id) + '_' + str(_group))
def main(_directions=None): if _directions is None: _directions = ['inside', 'outside'] _experiments = all_experiments() _experiments = filtering.by_categories( _experiments=_experiments, _is_single_cell=False, _is_high_temporal_resolution=False, _is_bleb=False, _is_dead_dead=False, _is_live_dead=False, _is_bead=False, _is_metastasis=False ) _tuples = load.experiments_groups_as_tuples(_experiments) _tuples = filtering.by_real_pairs(_tuples) _tuples = filtering.by_band(_tuples) _tuples = filtering.by_pair_distance_range(_tuples, PAIR_DISTANCE_RANGE) print('Total tuples:', len(_tuples)) _arguments = [] for _tuple in _tuples: _experiment, _series_id, _group = _tuple _latest_time_frame = compute.latest_time_frame_before_overlapping(_experiment, _series_id, _group, OFFSET_X) for _cell_id, _direction in product(['left_cell', 'right_cell'], _directions): _arguments.append({ 'experiment': _experiment, 'series_id': _series_id, 'group': _group, 'length_x': QUANTIFICATION_WINDOW_LENGTH_IN_CELL_DIAMETER, 'length_y': QUANTIFICATION_WINDOW_HEIGHT_IN_CELL_DIAMETER, 'length_z': QUANTIFICATION_WINDOW_WIDTH_IN_CELL_DIAMETER, 'offset_x': OFFSET_X, 'offset_y': OFFSET_Y, 'offset_z': OFFSET_Z, 'cell_id': _cell_id, 'direction': _direction, 'time_points': _latest_time_frame }) _windows_dictionary, _windows_to_compute = \ compute.windows(_arguments, _keys=['experiment', 'series_id', 'group', 'cell_id', 'direction']) _fiber_densities = compute.fiber_densities(_windows_to_compute, _subtract_border=True) _experiments_fiber_densities = { _key: [_fiber_densities[_tuple] for _tuple in _windows_dictionary[_key]] for _key in _windows_dictionary } for _direction in _directions: _y_arrays = [[] for _i in DERIVATIVES] for _tuple in tqdm(_tuples, desc='Experiments loop'): _experiment, _series_id, _group = _tuple if (_experiment, _series_id, _group, 'left_cell', _direction) not in _windows_dictionary or \ (_experiment, _series_id, _group, 'right_cell', _direction) not in _windows_dictionary: continue _properties = load.group_properties(_experiment, _series_id, _group) _left_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'left_cell', _direction)] _right_cell_fiber_densities = \ _experiments_fiber_densities[(_experiment, _series_id, _group, 'right_cell', _direction)] _left_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['left_cell'], _left_cell_fiber_densities) _right_cell_fiber_densities = compute.remove_blacklist( _experiment, _series_id, _properties['cells_ids']['right_cell'], _right_cell_fiber_densities) if not OUT_OF_BOUNDARIES: _left_cell_fiber_densities, _right_cell_fiber_densities = \ compute.longest_same_indices_shared_in_borders_sub_array( _left_cell_fiber_densities, _right_cell_fiber_densities ) else: _left_cell_fiber_densities = [_fiber_density[0] for _fiber_density in _left_cell_fiber_densities] _right_cell_fiber_densities = [_fiber_density[0] for _fiber_density in _right_cell_fiber_densities] # ignore small arrays _minimum_time_frame_for_correlation = compute.minimum_time_frames_for_correlation(_experiment) if len(_left_cell_fiber_densities) < _minimum_time_frame_for_correlation or \ len(_right_cell_fiber_densities) < _minimum_time_frame_for_correlation: continue for _derivative_index, _derivative in enumerate(DERIVATIVES): _y_arrays[_derivative_index].append(compute_lib.correlation( compute_lib.derivative(_left_cell_fiber_densities, _n=_derivative), compute_lib.derivative(_right_cell_fiber_densities, _n=_derivative) )) print('Direction:', _direction) print('Total pairs:', len(_y_arrays[0])) print('Wilcoxon around the zero') for _y_array, _derivative in zip(_y_arrays, DERIVATIVES): print('Derivative:', _derivative, wilcoxon(_y_array)) # plot _y_title = 'Inner correlation' if _direction == 'inside' else 'Outer correlation' _colors_array = config.colors(3) _fig = go.Figure( data=[ go.Box( y=_y, name=_derivative, boxpoints='all', jitter=1, pointpos=0, line={ 'width': 1 }, fillcolor='white', marker={ 'size': 10, 'color': _color }, opacity=0.7, showlegend=False ) for _y, _derivative, _color in zip(_y_arrays, DERIVATIVES_TEXT, _colors_array) ], layout={ 'xaxis': { 'title': 'Fiber density derivative', 'zeroline': False }, 'yaxis': { 'title': _y_title, 'range': [-1, 1], 'zeroline': False, 'tickmode': 'array', 'tickvals': [-1, -0.5, 0, 0.5, 1] } } ) save.to_html( _fig=_fig, _path=os.path.join(paths.PLOTS, save.get_module_name()), _filename='plot_direction_' + _direction )