def main():
parser = argparse.ArgumentParser(
description='create boxplots of interactions for timecourse' +
' at different distances')
parser.add_argument(
'-i',
help='timecourse hdf5 files in order Mock -> Overnight digestion',
nargs=7)
args = parser.parse_args()
# List of timecourse file objects in order
f_obj_list = []
for a in args.i:
f_obj_list.append(h5py.File(a, 'r'))
# List of chromosome 14 matrices for timecourse data
# in order
chr14_list = []
for o in f_obj_list:
chr14_list.append(mf.get_cis_matrix(o, 'chr14'))
if not check_shape(chr14_list):
print 'ERROR unequal dimensions for cis matrices'
sys.exit()
# Max
z = chr14_list[2]
max_idx = np.unravel_index(np.argmax(z), z.shape)
print max_idx
print z[max_idx]
quit()
# Boxplots for all interactions
data = []
for c in chr14_list:
iu1 = np.triu_indices(c.shape[0])
data.append(c[iu1])
my_boxplot(data)
# Mean interaction plots by distance
plot_mean_interaction_distance(chr14_list)
def main():
parser = argparse.ArgumentParser(
description='Return hdf5 with z-score cis matrices for input hdf5')
parser.add_argument('-i',
help='input hdf5 Hi-C file',
type=str,
required=True)
args = parser.parse_args()
# Generate z-score hdf5
if args.i[-5:] == '.hdf5':
filename_prefix = args.i[:-5]
zscore_file = filename_prefix + '_zScore.hdf5'
shutil.copy(args.i, zscore_file)
f = h5py.File(zscore_file, 'r+')
obs = f['interactions'][:]
blocksize = f['interactions'].chunks
# Loop through chromosomes:
chroms = f['chrs'][:]
for chrom in chroms:
obs_cis = mf.get_cis_matrix(f, chrom)
z = mf.z_score(obs_cis)
# Replace observed cis matrix with zscore matrix
chr_idx, = np.where(chroms == chrom)
chr_idx = chr_idx[0]
bins = f['chr_bin_range'][chr_idx]
obs[bins[0]:bins[1] + 1, bins[0]:bins[1] + 1] = z
del f['interactions']
f.create_dataset("interactions",
data=obs,
dtype='float64',
compression='gzip',
chunks=blocksize)
f.close()
else:
print 'ERROR: wrong file extension'
sys.exit()
def main():
parser = argparse.ArgumentParser(
description='create mean interactions vs timecourse and color' +
' by distance plots')
parser.add_argument(
'-i',
help='timecourse hdf5 files in order Mock -> Overnight digestion',
nargs=7)
parser.add_argument(
'-c',
help=
'compartment file (ex. HBHiC-K562-MN-Dp-1__hg19__genome__C-500000-raw_scaleBy_2.72.balanced_scaleBy_51.45__all.zScore.eigen1.sorted.bedGraph)',
type=str,
required=True)
args = parser.parse_args()
# List of timecourse file objects in order
f_obj_list = []
for a in args.i:
f_obj_list.append(h5py.File(a, 'r'))
# List of chromosome 14 matrices for timecourse data
# in order
chr14_list = []
for o in f_obj_list:
chr14_list.append(mf.get_cis_matrix(o, 'chr14'))
if not check_shape(chr14_list):
print 'ERROR unequal dimensions for cis matrices'
sys.exit()
# Make mean interaction plots by distane for
# A-A, B-B, A-B
CF = open(args.c, 'r')
eigen = get_chrom_eigen(CF, 'chr14')
plot_mean_interaction_distance_AB(chr14_list, eigen, 'AA')
plot_mean_interaction_distance_AB(chr14_list, eigen, 'BB')
plot_mean_interaction_distance_AB(chr14_list, eigen, 'AB')
def main():
parser = argparse.ArgumentParser(
description='Create half-life matrix from timecourse')
parser.add_argument(
'-i',
help='timecourse hdf5 files in order Mock -> Overnight digestion',
nargs=7)
args = parser.parse_args()
# Create copy of mock and use that to write over
# with half life data
shutil.copy(args.i[0], 'half_life_chr14_6Mb.hdf5')
f = h5py.File('half_life_chr14_6Mb.hdf5', 'r+')
# List of timecourse file objects in order
f_obj_list = []
for a in args.i:
f_obj_list.append(h5py.File(a, 'r'))
# List of chromosome 14 matrices for timecourse data
# in order
chr14_list = []
for o in f_obj_list:
chr14_list.append(mf.get_cis_matrix(o, 'chr14'))
if not check_shape(chr14_list):
print 'ERROR unequal dimensions for cis matrices'
sys.exit()
time = np.array([5, 60, 120, 180, 240, 960], dtype=float)
# # 1 MB distance
# interactions = []
# for h in chr14_list:
# interactions.append(h[500, 510])
# # delta interactions
# delta_intxns = get_delta_intxns(interactions)
# mid = get_intxns_half(delta_intxns)
# hl, params = get_half_life(delta_intxns, time, mid)
# x = np.linspace(-10, 1000, 100, dtype=float)
# y = exp_decay(x, *params)
# scatter_plot(delta_intxns, time, x, y, 'test_1MB_distance_delta.png', hl, mid)
# # 400kb distance
# interactions = []
# for h in chr14_list:
# interactions.append(h[500, 504])
# # delta interactions
# delta_intxns = get_delta_intxns(interactions)
# mid = get_intxns_half(delta_intxns)
# hl, params = get_half_life(delta_intxns, time, mid)
# x = np.linspace(-10, 1000, 100, dtype=float)
# y = exp_decay(x, *params)
# scatter_plot(delta_intxns, time, x, y, 'test_400kb_distance_delta.png', hl, mid)
chrom = 'chr14'
binSize = f['bin_positions'][0, 2]
dist = (6000000 / binSize) / 2
#obs = f['interactions'][:]
bin_positions = f['bin_positions'][:]
num_bins = len(bin_positions)
chr_idx, = np.where(f['chrs'][:] == chrom)
chr_idx = chr_idx[0]
bins = f['chr_bin_range'][chr_idx]
for i in range(bins[0], bins[1] + 1):
#for i in range(bins[0] + 1250, bins[1] + 1):
print 'on row: ' + str(i)
# 6Mb window
# Check if row is all nan
if np.all(np.isnan(f['interactions'][i])) or np.nansum(
f['interactions'][i]) == 0:
f['interactions'][i] = np.nan
# Check if at start of chromosome (upstream range reaches trans)
elif bin_positions[i, 0] != bin_positions[i - dist, 0]:
f['interactions'][i] = np.nan
# Check if at end of chromosome (downstream range reaches trans)
elif bin_positions[i, 0] != bin_positions[i + (dist - 1), 0]:
f['interactions'][i] = np.nan
else:
for j in range(bins[0], bins[1] + 1):
# Check if in 6Mb window
if j >= i - dist and j < i + dist:
interactions = []
for h in chr14_list:
interactions.append(h[i - bins[0], j - bins[0]])
if np.any(np.isnan(interactions)) or (interactions[0]
== 0):
f['interactions'][i, j] = np.nan
else:
delta_intxns = get_delta_intxns(interactions)
mid = get_intxns_half(delta_intxns)
hl, params = get_half_life(delta_intxns, time, mid)
f['interactions'][i, j] = hl
f['interactions'][j, i] = hl
else:
f['interactions'][i, j] = np.nan
f.close()
def main():
parser = argparse.ArgumentParser(description='Create half-life matrix from timecourse')
parser.add_argument('-i', help= 'timecourse hdf5 files in order Mock -> Overnight digestion', nargs=7)
parser.add_argument('-t', help= 'median trans interaction file (eg. median_trans.txt)', type=str, required=True)
args = parser.parse_args()
# Create copy of mock and use that to write over
# with half life data
shutil.copy(args.i[0], 'half_life_chr14_correct.hdf5')
f = h5py.File('half_life_chr14_correct.hdf5', 'r+')
# Get trans interaction data
trans_dict = {}
with open(args.t, 'r') as t:
for line in t:
splitline = line.split()
trans_dict[splitline[0]] = float(splitline[1])
# trans graph
trans_barplot(args.i, trans_dict)
# List of timecourse file objects in order
f_obj_list = []
for a in args.i:
f_obj_list.append(h5py.File(a, 'r'))
# List of chromosome 14 matrices for timecourse data
# in order
chr14_list = []
for o in f_obj_list:
chr14_list.append(mf.get_cis_matrix(o, 'chr14'))
if not check_shape(chr14_list):
print 'ERROR unequal dimensions for cis matrices'
sys.exit()
median_trans = []
for file in args.i:
median_trans.append(trans_dict[file])
median_trans = np.array(median_trans)
chr14_list = subtract_trans(chr14_list, median_trans)
interactions = []
time = np.array([0, 5, 60, 120, 180, 240, 960])
for h in chr14_list:
interactions.append(h[80, 90])
# polyfit
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
hl, pm = get_half_life(interactions, time)
scatter_plot(interactions, time, xp, p, 'test_10_correct.png', hl, pm)
interactions = []
for h in chr14_list:
interactions.append(h[80,180])
# polyfit
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
hl, pm = get_half_life(interactions, time)
scatter_plot(interactions, time, xp, p, 'test_100_correct.png', hl, pm)
print interactions
interactions = []
for h in chr14_list:
interactions.append(h[80,85])
# polyfit
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
hl, pm = get_half_life(interactions, time)
scatter_plot(interactions, time, xp, p, 'test_5_correct.png', hl, pm)
print interactions
interactions = []
for h in chr14_list:
interactions.append(h[80,120])
# polyfit
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
hl, pm = get_half_life(interactions, time)
scatter_plot(interactions, time, xp, p, 'test_40_correct.png', hl, pm)
hl, pred_mid = get_half_life(interactions, time)
print '*****hl*****'
print hl
print '************'
#scatter_plot(interactions, time, xp, p, 'test_10_lines.png', hl, pred_mid)
chrom = 'chr14'
chr_idx, = np.where(f['chrs'][:] == chrom)
chr_idx = chr_idx[0]
bins = f['chr_bin_range'][chr_idx]
print bins
print len(range(bins[0], bins[1] + 1))
print f['interactions'][bins[0] + 80, bins[0] + 90]
for i in range((bins[1] - bins[0]) + 1):
print 'on row: ' + str(i)
for j in range((bins[1] - bins[0]) + 1):
interactions = []
for h in chr14_list:
interactions.append(h[i, j])
hl, pred_mid = get_half_life(interactions, time)
f['interactions'][bins[0] + i, bins[0] + j] = hl
print f['interactions'][:]
print f['interactions'][bins[0] + 80, bins[0] + 90]
f.close()
def main():
parser = argparse.ArgumentParser(
description='Create half-life matrix from timecourse')
parser.add_argument(
'-i',
help='timecourse hdf5 files in order Mock -> Overnight digestion',
nargs=7)
args = parser.parse_args()
# Create copy of mock and use that to write over
# with slope data
shutil.copy(args.i[0], 'slope_chr14_zscore.hdf5')
f = h5py.File('slope_chr14_zscore.hdf5', 'r+')
# List of timecourse file objects in order
f_obj_list = []
for a in args.i:
f_obj_list.append(h5py.File(a, 'r'))
# List of chromosome 14 matrices for timecourse data
# in order
chr14_list = []
for o in f_obj_list:
chr14_list.append(mf.get_cis_matrix(o, 'chr14'))
if not check_shape(chr14_list):
print 'ERROR unequal dimensions for cis matrices'
sys.exit()
# Get expected matrices
chr14_expected = []
for c in chr14_list:
chr14_expected.append(mf.expected(c))
# Write expected matrices to file
for i, c in enumerate(chr14_expected):
colnames = map(str, range(c.shape[0]))
colnames = ['x' + x for x in colnames]
rownames = list(colnames)
mf.numpy_matrix_2_dekker(
c, colnames, rownames,
'expected/' + args.i[i][:-5] + '_expected.matrix.gz')
# Get observed-expected matrices
chr14_obs_exp = []
for i, c in enumerate(chr14_list):
oe = c - chr14_expected[i]
chr14_obs_exp.append(oe)
# Write observed - expected matrices to file
for i, c in enumerate(chr14_obs_exp):
colnames = map(str, range(c.shape[0]))
colnames = ['x' + x for x in colnames]
rownames = list(colnames)
mf.numpy_matrix_2_dekker(
c, colnames, rownames,
'obs-exp/' + args.i[i][:-5] + '_obs-exp.matrix.gz')
# Get z-score matrices
chr14_zscore = []
for c in chr14_list:
chr14_zscore.append(mf.z_score(c))
# Write z-score matrices to file
for i, c in enumerate(chr14_zscore):
colnames = map(str, range(c.shape[0]))
colnames = ['x' + x for x in colnames]
rownames = list(colnames)
mf.numpy_matrix_2_dekker(
c, colnames, rownames,
'zscore/' + args.i[i][:-5] + '_zscore.matrix.gz')
interactions = []
time = np.array([0, 5, 60, 120, 180, 240, 960])
for h in chr14_zscore:
interactions.append(h[80, 90])
# polyfit
s = get_slope(interactions, time, 150)
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
scatter_plot_slope(interactions, time, xp, p, s, 150, 'test_10_slope.png')
interactions = []
for h in chr14_zscore:
interactions.append(h[80, 180])
# polyfit
s = get_slope(interactions, time, 150)
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
scatter_plot_slope(interactions, time, xp, p, s, 150, 'test_100_slope.png')
print interactions
interactions = []
for h in chr14_zscore:
interactions.append(h[80, 85])
# polyfit
s = get_slope(interactions, time, 150)
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
scatter_plot_slope(interactions, time, xp, p, s, 150, 'test_5_slope.png')
print interactions
interactions = []
for h in chr14_zscore:
interactions.append(h[80, 120])
# polyfit
s = get_slope(interactions, time, 150)
z = np.polyfit(time, interactions, 2)
p = np.poly1d(z)
xp = np.linspace(-10, 1000, 100)
scatter_plot_slope(interactions, time, xp, p, s, 150, 'test_40_slope.png')
quit()
chrom = 'chr14'
chr_idx, = np.where(f['chrs'][:] == chrom)
chr_idx = chr_idx[0]
bins = f['chr_bin_range'][chr_idx]
print bins
print len(range(bins[0], bins[1] + 1))
print f['interactions'][bins[0] + 80, bins[0] + 90]
for i in range((bins[1] - bins[0]) + 1):
print 'on row: ' + str(i)
for j in range((bins[1] - bins[0]) + 1):
interactions = []
for h in chr14_zscore:
interactions.append(h[i, j])
s = get_slope(interactions, time, 150)
if s < 1e-15 and s > -1e-15:
s = 0.0
f['interactions'][bins[0] + i, bins[0] + j] = s
print f['interactions'][:]
print f['interactions'][bins[0] + 80, bins[0] + 90]
f.close()