def test_quality_plot(self):
df3 = parse_rosetta_fragments(self.frag3).sample_top_neighbors()
df9 = parse_rosetta_fragments(self.frag9)
# auto-load
df3 = df3.add_quality_measure(None)
# load target quality file
with pytest.raises(ValueError):
df9 = df9.add_quality_measure(self.frag3q)
df9 = df9.add_quality_measure(self.frag9q)
assert df3.is_comparable(df9) is False
assert 'rmsd' in df3
assert 'rmsd' in df9
consensus_seq = df9.select_quantile().quick_consensus_sequence()
consensus_sse = df9.select_quantile(
).quick_consensus_secondary_structure()
assert consensus_seq == "KIPVPVVVNGKIVAVVVVPPENLEEALLEALKELGLIKDPEEVKAVVVSPDGRLELSF"
assert consensus_sse == "EEEEEEEELLEEEEEEEELLLLHHHHHHHHHHHHLLLLLLLLLLEEEEELLLEEEEEE"
fig = plt.figure(figsize=(25, 10))
plot_fragment_profiles(fig, df3, df9, consensus_seq, consensus_sse)
plt.tight_layout()
return fig
def test_frequency_matrices_and_networks(self):
df3 = parse_rosetta_fragments(self.frag3)
df9 = parse_rosetta_fragments(self.frag9)
# auto-load
df3 = df3.add_quality_measure(None)
# load target quality file
df9 = df9.add_quality_measure(self.frag9q)
matrix = df3.select_quantile(0.1).make_sequence_matrix()
assert matrix.min().min() == -9
matrix = df9.select_quantile(0.1).make_sequence_matrix(frequency=True)
G = df9.select_quantile(0.1).make_per_position_frequency_network()
Gf = df9.select_quantile(0.1).make_frequency_network()
assert matrix.shape == (58, 20)
assert G.number_of_edges() > Gf.number_of_edges()
value = 1 - G.get_edge_data("0X", "1A")['weight']
assert matrix["A"].values[0] == pytest.approx(value)
n = 6
target = str(n + 1) + "R"
for aa in list("ARNDCQEGHILKMFPSTWYV"):
origin = str(n) + aa
if origin in G:
value = 1 - G.get_edge_data(origin, target)['weight']
assert matrix["R"].values[n] == pytest.approx(value)
def test_concat_fragments(self):
# load fragments
_3mers = parse_rosetta_fragments(self.frag3)
# make chunks
_3mers_1 = _3mers[(_3mers['frame'] >= 9) & (_3mers['frame'] <= 12)]
_3mers_2 = _3mers[(_3mers['frame'] >= 22) & (_3mers['frame'] <= 24)]
_3mers_3 = _3mers[(_3mers['frame'] >= 45) & (_3mers['frame'] <= 46)]
# mix fragments
m = concat_fragments([_3mers_3, _3mers_1, _3mers_2])
# checkpoints
assert len(m) == 5400
assert list(m.drop_duplicates('frame')['frame']) == list(range(1, 10))
f1 = NamedTemporaryFile(delete=False)
f1.close()
nonstrict = write_rosetta_fragments(m, 3, 300, f1.name, False)
f2 = NamedTemporaryFile(delete=False)
f2.close()
isstrict = write_rosetta_fragments(m.renumber(10).top_limit(30),
prefix=f2.name,
strict=True)
assert not parse_rosetta_fragments(nonstrict).is_comparable(
parse_rosetta_fragments(isstrict))
def main( options ):
# Read Fragment Files
small_f = parse_rosetta_fragments(options.fsmall)
large_f = parse_rosetta_fragments(options.flarge)
# Read or calculate Fragment Quality
small_f = small_f.add_quality_measure(options.qsmall, options.pdb)
large_f = large_f.add_quality_measure(options.qlarge, options.pdb)
# Plot
fig = plt.figure(figsize=(40, 10) if options.format == "h" else (20, 20))
grid = (1, 2) if options.format == "h" else (2, 1)
ax00 = plt.subplot2grid(grid, (0, 0), fig=fig)
ax01 = plt.subplot2grid(grid, (0, 1) if options.format == "h" else (1, 0), fig=fig)
ax00.yaxis.set_major_formatter(FormatStrFormatter('%.1f'))
ax01.yaxis.set_major_formatter(FormatStrFormatter('%.1f'))
plot_fragments(small_f, large_f, small_ax=ax00, large_ax=ax01,
showfliers=False, titles="top" if options.format == "h" else "right")
if options.format == "h":
plt.tight_layout(pad=2)
else:
plt.tight_layout(rect=(0.037, 0, 1, 1))
# Write to file
if options.ofile is not None:
plt.savefig(options.ofile)
# Show on screen
if not options.silent:
plt.show()
return fig
def test_add_fragments_append(self):
df = parse_rosetta_fragments(self.frag3)
xx = df[(df['frame'] <= 10) & (df['neighbor'] <= 100)]
dfrep = df.add_fragments(xx, 10, 'append')
fig = plt.figure(figsize=(25, 10))
ax0 = plt.subplot2grid((2, 1), (0, 0))
prange = range(len(dfrep.groupby('frame')))
ax0.bar(prange,
[max(y['neighbor']) for x, y in dfrep.groupby('frame')])
ax1 = plt.subplot2grid((2, 1), (1, 0))
ax1.bar(
prange,
[y['neighbors'].unique()[0] for x, y in dfrep.groupby('frame')])
plt.tight_layout()
return fig
def make_fragment_files(self,
dfloop: pd.DataFrame,
edges: Dict,
masfile: Path,
no_loop: Optional[bool] = True) -> Dict:
"""Combin the fragments from the different matches.
"""
data = {
'loop_length': int(dfloop.iloc[0]['loop_length']),
'abego': list(dfloop['loop'].values),
'edges': edges,
'fragfiles': [],
'match_count': 0
}
dfs3 = []
dfs9 = []
sample = math.ceil(200 / dfloop.shape[0])
if not no_loop:
for i, row in dfloop.iterrows():
# Remember: MASTER match starts with 0!
dfs3.append((parse_rosetta_fragments(
str(row['3mers']),
source=f'{row["pdb"]}_{row["chain"]}').slice_region(
row['start'] + 1,
row['stop'] + 1).sample_top_neighbors(sample).renumber(
edges['ini']).top_limit(edges['end'])))
dfs9.append((parse_rosetta_fragments(
str(row['9mers']),
source=f'{row["pdb"]}_{row["chain"]}').slice_region(
row['start'] + 1,
row['stop'] + 1).sample_top_neighbors(sample).renumber(
edges['ini']).top_limit(edges['end'])))
else:
for i, row in dfloop.iterrows():
# Remember: MASTER match starts with 0!
dfs3.append((parse_rosetta_fragments(
str(row['3mers']),
source=f'{row["pdb"]}_{row["chain"]}').slice_region(
row['start'] + 1,
row['stop'] + 1).sample_top_neighbors(sample).renumber(
edges['ini']).top_limit(edges['end'])))
dfs9.append((parse_rosetta_fragments(
str(row['9mers']),
source=f'{row["pdb"]}_{row["chain"]}').slice_region(
row['start'] + 1,
row['stop'] + 1).sample_top_neighbors(sample).renumber(
edges['ini']).top_limit(edges['end'])))
# Merge Fragments
dfs3all = dfs3[0]
dfs9all = dfs9[0]
for i in range(1, len(dfs3)):
dfs3all = dfs3all.add_fragments(dfs3[i],
ini=edges['ini'],
how='append')
dfs9all = dfs9all.add_fragments(dfs9[i],
ini=edges['ini'],
how='append')
dfs3all = dfs3all.sample_top_neighbors(200)
dfs9all = dfs9all.sample_top_neighbors(200)
# set up
lord = int(dfloop.order.drop_duplicates().values[0])
nfolder = masfile.parent.absolute().joinpath(f'loop{int(lord):02d}')
nfolder.mkdir(parents=True, exist_ok=True)
masfile2 = str(nfolder.joinpath(f'jump{int(lord):02d}'))
self.log.debug('Writing 3mers fragfile\n')
#data['fragfiles'].append(write_rosetta_fragments(dfs3all, prefix=str(masfile.with_suffix('')), strict=True))
data['fragfiles'].append(
write_rosetta_fragments(dfs3all, prefix=masfile2, strict=True))
self.log.debug(f'3mers fragfile: {data["fragfiles"][-1]}\n')
self.log.debug('Writing 9mers fragfile\n')
#data['fragfiles'].append(write_rosetta_fragments(dfs9all, prefix=str(masfile.with_suffix('')), strict=True))
data['fragfiles'].append(
write_rosetta_fragments(dfs9all, prefix=masfile2, strict=True))
self.log.debug(f'9mers fragfile: {data["fragfiles"][-1]}\n')
dfs3all.drop(columns=[
'pdb', 'frame', 'neighbors', 'neighbor', 'aa', 'sse', 'phi', 'psi',
'omega'
]).to_csv(data['fragfiles'][0] + '.csv', index=False)
dfs9all.drop(columns=[
'pdb', 'frame', 'neighbors', 'neighbor', 'aa', 'sse', 'phi', 'psi',
'omega'
]).to_csv(data['fragfiles'][1] + '.csv', index=False)
imageprefix = Path(masfile2).with_suffix('.fragprofile')
TBPlot.plot_fragment_templates(self.log, dfs3all, dfs9all, imageprefix)
return data, nfolder
def make_fragment_files(dfloop: pd.DataFrame, edges: Dict,
masfile: Path) -> Dict:
"""
"""
data = {
'loop_length': int(dfloop.iloc[0]['loop_length']),
'abego': list(dfloop['loop'].values),
'edges': edges,
'fragfiles': [],
'match_count': 0
}
dfs3 = []
dfs9 = []
sample = math.ceil(200 / dfloop.shape[0])
for i, row in dfloop.iterrows():
# Remember: MASTER match starts with 0!
dfs3.append((parse_rosetta_fragments(
str(row['3mers']),
source='{}_{}'.format(row['pdb'], row['chain'])).slice_region(
row['match'][0][0] + 1,
row['match'][1][1] + 1).sample_top_neighbors(sample).renumber(
edges['ini']).top_limit(edges['end'])))
dfs9.append((parse_rosetta_fragments(
str(row['9mers']),
source='{}_{}'.format(row['pdb'], row['chain'])).slice_region(
row['match'][0][0] + 1,
row['match'][1][1] + 1).sample_top_neighbors(sample).renumber(
edges['ini']).top_limit(edges['end'])))
# Merge Fragments
dfs3all = dfs3[0]
dfs9all = dfs9[0]
for i in range(1, len(dfs3)):
dfs3all = dfs3all.add_fragments(dfs3[i],
ini=edges['ini'],
how='append')
dfs9all = dfs9all.add_fragments(dfs9[i],
ini=edges['ini'],
how='append')
dfs3all = dfs3all.sample_top_neighbors(200)
dfs9all = dfs9all.sample_top_neighbors(200)
if TBcore.get_option('system', 'debug'):
sys.stdout.write('Writing 3mers fragfile\n')
data['fragfiles'].append(
write_rosetta_fragments(dfs3all,
prefix=str(masfile.with_suffix('')),
strict=True))
if TBcore.get_option('system', 'debug'):
sys.stdout.write('3mers fragfile: {}\n'.format(data['fragfiles'][-1]))
sys.stdout.write('Writing 9mers fragfile\n')
data['fragfiles'].append(
write_rosetta_fragments(dfs9all,
prefix=str(masfile.with_suffix('')),
strict=True))
if TBcore.get_option('system', 'debug'):
sys.stdout.write('9mers fragfile: {}\n'.format(data['fragfiles'][-1]))
dfs3all.drop(columns=[
'pdb', 'frame', 'neighbors', 'neighbor', 'aa', 'sse', 'phi', 'psi',
'omega'
]).to_csv(data['fragfiles'][0] + '.csv', index=False)
dfs9all.drop(columns=[
'pdb', 'frame', 'neighbors', 'neighbor', 'aa', 'sse', 'phi', 'psi',
'omega'
]).to_csv(data['fragfiles'][1] + '.csv', index=False)
imageprefix = masfile.with_suffix('.fragprofile')
TButil.plot_fragment_templates(dfs3all, dfs9all, imageprefix)
return data
TButil.checkpoint_out(checkpoint, data)
return case
def loop_master_protocol(case: Case, folders: Path) -> Tuple[str, str]:
"""
"""
lf = case['metadata.loop_fragments']
if lf is None:
raise TButil.PluginOrderError(
'Data that should be loaded through loop_master is not found.')
for i, loop in enumerate(lf):
if i == 0:
ff3 = parse_rosetta_fragments(loop['fragfiles'][0])
ff9 = parse_rosetta_fragments(loop['fragfiles'][1])
df3 = [
pd.read_csv(str(loop['fragfiles'][0]) + '.csv'),
]
df9 = [
pd.read_csv(str(loop['fragfiles'][1]) + '.csv'),
]
else:
df3.append(pd.read_csv(str(loop['fragfiles'][0]) + '.csv'))
df9.append(pd.read_csv(str(loop['fragfiles'][1]) + '.csv'))
ff3 = ff3.add_fragments(parse_rosetta_fragments(
loop['fragfiles'][0]),
ini=int(loop['edges']['ini']),
how='append')
ff9 = ff9.add_fragments(parse_rosetta_fragments(
# Checkpoint save
TButil.checkpoint_out(self.log, checkpoint, data)
return case
def loop_master_protocol(self, case: Case,
folders: Path) -> Tuple[str, str]:
"""
"""
lf = case['metadata.loop_fragments']
if lf is None:
raise NodeMissingError(
'Data that should be loaded through loop_master is not found.')
for i, loop in enumerate(lf):
if i == 0:
ff3 = parse_rosetta_fragments(loop['fragfiles'][0])
ff9 = parse_rosetta_fragments(loop['fragfiles'][1])
df3 = [
pd.read_csv(str(loop['fragfiles'][0]) + '.csv'),
]
df9 = [
pd.read_csv(str(loop['fragfiles'][1]) + '.csv'),
]
else:
df3.append(pd.read_csv(str(loop['fragfiles'][0]) + '.csv'))
df9.append(pd.read_csv(str(loop['fragfiles'][1]) + '.csv'))
ff3 = ff3.add_fragments(parse_rosetta_fragments(
loop['fragfiles'][0]),
ini=int(loop['edges']['ini']),
how='append')
ff9 = ff9.add_fragments(parse_rosetta_fragments(