def parse(handle, format):
"""Parses an output file of motif finding programs.
Currently supported formats (case is ignored):
- AlignAce: AlignAce output file format
- MEME: MEME output file motif
- MAST: MAST output file motif
- TRANSFAC: TRANSFAC database file format
- pfm: JASPAR-style position-frequency matrix
- jaspar: JASPAR-style multiple PFM format
- sites: JASPAR-style sites file
As files in the pfm and sites formats contain only a single motif,
it is easier to use Bio.motifs.read() instead of Bio.motifs.parse()
for those.
For example:
>>> from Bio import motifs
>>> for m in motifs.parse(open("Motif/alignace.out"), "AlignAce"):
... print(m.consensus)
TCTACGATTGAG
CTGCAGCTAGCTACGAGTGAG
GTGCTCTAAGCATAGTAGGCG
GCCACTAGCAGAGCAGGGGGC
CGACTCAGAGGTT
CCACGCTAAGAGAGGTGCCGGAG
GCGCGTCGCTGAGCA
GTCCATCGCAAAGCGTGGGGC
GGGATCAGAGGGCCG
TGGAGGCGGGG
GACCAGAGCTTCGCATGGGGG
GGCGTGCGTG
GCTGGTTGCTGTTCATTAGG
GCCGGCGGCAGCTAAAAGGG
GAGGCCGGGGAT
CGACTCGTGCTTAGAAGG
"""
format = format.lower()
if format=="alignace":
from Bio.motifs import alignace
record = alignace.read(handle)
return record
elif format=="meme":
from Bio.motifs import meme
record = meme.read(handle)
return record
elif format=="mast":
from Bio.motifs import mast
record = mast.read(handle)
return record
elif format=="transfac":
from Bio.motifs import transfac
record = transfac.read(handle)
return record
elif format in ('pfm', 'sites', 'jaspar'):
from Bio.motifs import jaspar
record = jaspar.read(handle, format)
return record
else:
raise ValueError("Unknown format %s" % format)
def transfac2ic(args):
"""Compute information content for each filter motif (.transfac)."""
train_samples = np.load(args.train, mmap_mode='r')
probs = np.mean(np.mean(train_samples, axis=1), axis=0)
# background
bg = {'A': probs[0], 'C': probs[1], 'G': probs[2], 'T': probs[3]}
# create output directory
out_dir = os.path.dirname(args.out_file)
if out_dir == "":
out_dir = "."
if not os.path.exists(out_dir):
os.makedirs(out_dir)
# load all filter motifs
with open(args.in_file) as handle:
records = transfac.read(handle)
# for each motif compute IC:
for m in records:
pwm = m.counts.normalize(pseudocounts=bg)
pwm.background = bg
# pssm = pwm.log_odds(background = bg))
ic = compute_mean_ic(pwm)
with open(args.out_file, "a") as file:
file.write(m.get("ID") + "\t" + str(ic) + "\n")
def parse(handle, format):
"""Parses an output file of motif finding programs.
Currently supported formats (case is ignored):
- AlignAce: AlignAce output file format
- MEME: MEME output file motif
- MAST: MAST output file motif
- TRANSFAC: TRANSFAC database file format
- pfm: JASPAR-style position-frequency matrix
- jaspar: JASPAR-style multiple PFM format
- sites: JASPAR-style sites file
As files in the pfm and sites formats contain only a single motif,
it is easier to use Bio.motifs.read() instead of Bio.motifs.parse()
for those.
For example:
>>> from Bio import motifs
>>> for m in motifs.parse(open("Motif/alignace.out"),"AlignAce"):
... print m.consensus
TCTACGATTGAG
CTGCAGCTAGCTACGAGTGAG
GTGCTCTAAGCATAGTAGGCG
GCCACTAGCAGAGCAGGGGGC
CGACTCAGAGGTT
CCACGCTAAGAGAGGTGCCGGAG
GCGCGTCGCTGAGCA
GTCCATCGCAAAGCGTGGGGC
GGGATCAGAGGGCCG
TGGAGGCGGGG
GACCAGAGCTTCGCATGGGGG
GGCGTGCGTG
GCTGGTTGCTGTTCATTAGG
GCCGGCGGCAGCTAAAAGGG
GAGGCCGGGGAT
CGACTCGTGCTTAGAAGG
"""
format = format.lower()
if format == "alignace":
from Bio.motifs import alignace
record = alignace.read(handle)
return record
elif format == "meme":
from Bio.motifs import meme
record = meme.read(handle)
return record
elif format == "mast":
from Bio.motifs import mast
record = mast.read(handle)
return record
elif format == "transfac":
from Bio.motifs import transfac
record = transfac.read(handle)
return record
elif format in ('pfm', 'sites', 'jaspar'):
from Bio.motifs import jaspar
record = jaspar.read(handle, format)
return record
else:
raise ValueError("Unknown format %s" % format)
def parse(handle, format, strict=True):
"""Parse an output file from a motif finding program.
Currently supported formats (case is ignored):
- AlignAce: AlignAce output file format
- ClusterBuster: Cluster Buster position frequency matrix format
- XMS: XMS matrix format
- MEME: MEME output file motif
- MINIMAL: MINIMAL MEME output file motif
- MAST: MAST output file motif
- TRANSFAC: TRANSFAC database file format
- pfm-four-columns: Generic position-frequency matrix format with four columns. (cisbp, homer, hocomoco, neph, tiffin)
- pfm-four-rows: Generic position-frequency matrix format with four row. (scertf, yetfasco, hdpi, idmmpmm, flyfactor survey)
- pfm: JASPAR-style position-frequency matrix
- jaspar: JASPAR-style multiple PFM format
- sites: JASPAR-style sites file
As files in the pfm and sites formats contain only a single motif,
it is easier to use Bio.motifs.read() instead of Bio.motifs.parse()
for those.
For example:
>>> from Bio import motifs
>>> with open("motifs/alignace.out") as handle:
... for m in motifs.parse(handle, "AlignAce"):
... print(m.consensus)
...
TCTACGATTGAG
CTGCACCTAGCTACGAGTGAG
GTGCCCTAAGCATACTAGGCG
GCCACTAGCAGAGCAGGGGGC
CGACTCAGAGGTT
CCACGCTAAGAGAAGTGCCGGAG
GCACGTCCCTGAGCA
GTCCATCGCAAAGCGTGGGGC
GAGATCAGAGGGCCG
TGGACGCGGGG
GACCAGAGCCTCGCATGGGGG
AGCGCGCGTG
GCCGGTTGCTGTTCATTAGG
ACCGACGGCAGCTAAAAGGG
GACGCCGGGGAT
CGACTCGCGCTTACAAGG
If strict is True (default), the parser will raise a ValueError if the
file contents does not strictly comply with the specified file format.
"""
format = format.lower()
if format == "alignace":
from Bio.motifs import alignace
return alignace.read(handle)
elif format == "meme":
from Bio.motifs import meme
return meme.read(handle)
elif format == "minimal":
from Bio.motifs import minimal
return minimal.read(handle)
elif format == "clusterbuster":
from Bio.motifs import clusterbuster
return clusterbuster.read(handle)
elif format in ("pfm-four-columns", "pfm-four-rows"):
from Bio.motifs import pfm
return pfm.read(handle, format)
elif format == "xms":
from Bio.motifs import xms
return xms.read(handle)
elif format == "mast":
from Bio.motifs import mast
return mast.read(handle)
elif format == "transfac":
from Bio.motifs import transfac
return transfac.read(handle, strict)
elif format in ("pfm", "sites", "jaspar"):
from Bio.motifs import jaspar
return jaspar.read(handle, format)
else:
raise ValueError("Unknown format %s" % format)
def parse(handle, format, strict=True):
"""Parse an output file from a motif finding program.
Currently supported formats (case is ignored):
- AlignAce: AlignAce output file format
- MEME: MEME output file motif
- MINIMAL: MINIMAL MEME output file motif
- MAST: MAST output file motif
- TRANSFAC: TRANSFAC database file format
- pfm: JASPAR-style position-frequency matrix
- jaspar: JASPAR-style multiple PFM format
- sites: JASPAR-style sites file
As files in the pfm and sites formats contain only a single motif,
it is easier to use Bio.motifs.read() instead of Bio.motifs.parse()
for those.
For example:
>>> from Bio import motifs
>>> with open("Motif/alignace.out") as handle:
... for m in motifs.parse(handle, "AlignAce"):
... print(m.consensus)
...
TCTACGATTGAG
CTGCAGCTAGCTACGAGTGAG
GTGCTCTAAGCATAGTAGGCG
GCCACTAGCAGAGCAGGGGGC
CGACTCAGAGGTT
CCACGCTAAGAGAGGTGCCGGAG
GCGCGTCGCTGAGCA
GTCCATCGCAAAGCGTGGGGC
GGGATCAGAGGGCCG
TGGAGGCGGGG
GACCAGAGCTTCGCATGGGGG
GGCGTGCGTG
GCTGGTTGCTGTTCATTAGG
GCCGGCGGCAGCTAAAAGGG
GAGGCCGGGGAT
CGACTCGTGCTTAGAAGG
If strict is True (default), the parser will raise a ValueError if the
file contents does not strictly comply with the specified file format.
"""
format = format.lower()
if format == "alignace":
from Bio.motifs import alignace
record = alignace.read(handle)
return record
elif format == "meme":
from Bio.motifs import meme
record = meme.read(handle)
return record
elif format == "minimal":
from Bio.motifs import minimal
record = minimal.read(handle)
return record
elif format == "mast":
from Bio.motifs import mast
record = mast.read(handle)
return record
elif format == "transfac":
from Bio.motifs import transfac
record = transfac.read(handle, strict)
return record
elif format in ('pfm', 'sites', 'jaspar'):
from Bio.motifs import jaspar
record = jaspar.read(handle, format)
return record
else:
raise ValueError("Unknown format %s" % format)
def parse(handle, format, strict=True):
"""Parse an output file from a motif finding program.
Currently supported formats (case is ignored):
- AlignAce: AlignAce output file format
- MEME: MEME output file motif
- MINIMAL: MINIMAL MEME output file motif
- MAST: MAST output file motif
- TRANSFAC: TRANSFAC database file format
- pfm: JASPAR-style position-frequency matrix
- jaspar: JASPAR-style multiple PFM format
- sites: JASPAR-style sites file
As files in the pfm and sites formats contain only a single motif,
it is easier to use Bio.motifs.read() instead of Bio.motifs.parse()
for those.
For example:
>>> from Bio import motifs
>>> with open("Motif/alignace.out") as handle:
... for m in motifs.parse(handle, "AlignAce"):
... print(m.consensus)
...
TCTACGATTGAG
CTGCAGCTAGCTACGAGTGAG
GTGCTCTAAGCATAGTAGGCG
GCCACTAGCAGAGCAGGGGGC
CGACTCAGAGGTT
CCACGCTAAGAGAGGTGCCGGAG
GCGCGTCGCTGAGCA
GTCCATCGCAAAGCGTGGGGC
GGGATCAGAGGGCCG
TGGAGGCGGGG
GACCAGAGCTTCGCATGGGGG
GGCGTGCGTG
GCTGGTTGCTGTTCATTAGG
GCCGGCGGCAGCTAAAAGGG
GAGGCCGGGGAT
CGACTCGTGCTTAGAAGG
If strict is True (default), the parser will raise a ValueError if the
file contents does not strictly comply with the specified file format.
"""
format = format.lower()
if format == "alignace":
from Bio.motifs import alignace
record = alignace.read(handle)
return record
elif format == "meme":
from Bio.motifs import meme
record = meme.read(handle)
return record
elif format == "minimal":
from Bio.motifs import minimal
record = minimal.read(handle)
return record
elif format == "mast":
from Bio.motifs import mast
record = mast.read(handle)
return record
elif format == "transfac":
from Bio.motifs import transfac
record = transfac.read(handle, strict)
return record
elif format in ('pfm', 'sites', 'jaspar'):
from Bio.motifs import jaspar
record = jaspar.read(handle, format)
return record
else:
raise ValueError("Unknown format %s" % format)
def parse(handle, format, strict=True):
"""Parse an output file from a motif finding program.
Currently supported formats (case is ignored):
- AlignAce: AlignAce output file format
- ClusterBuster: Cluster Buster position frequency matrix format
- XMS: XMS matrix format
- MEME: MEME output file motif
- MINIMAL: MINIMAL MEME output file motif
- MAST: MAST output file motif
- TRANSFAC: TRANSFAC database file format
- pfm-four-columns: Generic position-frequency matrix format with four columns. (cisbp, homer, hocomoco, neph, tiffin)
- pfm-four-rows: Generic position-frequency matrix format with four row. (scertf, yetfasco, hdpi, idmmpmm, flyfactor survey)
- pfm: JASPAR-style position-frequency matrix
- jaspar: JASPAR-style multiple PFM format
- sites: JASPAR-style sites file
As files in the pfm and sites formats contain only a single motif,
it is easier to use Bio.motifs.read() instead of Bio.motifs.parse()
for those.
For example:
>>> from Bio import motifs
>>> with open("motifs/alignace.out") as handle:
... for m in motifs.parse(handle, "AlignAce"):
... print(m.consensus)
...
TCTACGATTGAG
CTGCACCTAGCTACGAGTGAG
GTGCCCTAAGCATACTAGGCG
GCCACTAGCAGAGCAGGGGGC
CGACTCAGAGGTT
CCACGCTAAGAGAAGTGCCGGAG
GCACGTCCCTGAGCA
GTCCATCGCAAAGCGTGGGGC
GAGATCAGAGGGCCG
TGGACGCGGGG
GACCAGAGCCTCGCATGGGGG
AGCGCGCGTG
GCCGGTTGCTGTTCATTAGG
ACCGACGGCAGCTAAAAGGG
GACGCCGGGGAT
CGACTCGCGCTTACAAGG
If strict is True (default), the parser will raise a ValueError if the
file contents does not strictly comply with the specified file format.
"""
format = format.lower()
if format == "alignace":
from Bio.motifs import alignace
return alignace.read(handle)
elif format == "meme":
from Bio.motifs import meme
return meme.read(handle)
elif format == "minimal":
from Bio.motifs import minimal
return minimal.read(handle)
elif format == "clusterbuster":
from Bio.motifs import clusterbuster
return clusterbuster.read(handle)
elif format in ('pfm-four-columns', 'pfm-four-rows'):
from Bio.motifs import pfm
return pfm.read(handle, format)
elif format == "xms":
from Bio.motifs import xms
return xms.read(handle)
elif format == "mast":
from Bio.motifs import mast
return mast.read(handle)
elif format == "transfac":
from Bio.motifs import transfac
return transfac.read(handle, strict)
elif format in ('pfm', 'sites', 'jaspar'):
from Bio.motifs import jaspar
return jaspar.read(handle, format)
else:
raise ValueError("Unknown format %s" % format)
def motif_compare(args):
"""Compare PSSMs of filter motifs."""
# create output directory
if not os.path.exists(args.out_dir):
os.makedirs(args.out_dir)
# load training data to determine background nucleotide content
train_samples = np.load(args.train_data, mmap_mode='r')
probs = np.mean(np.mean(train_samples, axis=1), axis=0)
bg = {'A': probs[0], 'C': probs[1], 'G': probs[2], 'T': probs[3]}
# load all filter motifs from first file
with open(args.in_file1) as handle:
records1 = transfac.read(handle)
# load all filter motifs from second file
with open(args.in_file2) as handle:
records2 = transfac.read(handle)
# convert motifs to pssm's
pssms1 = {}
pssms2 = {}
rc_pssms2 = {}
for idx, m1 in enumerate(records1):
pwm1 = m1.counts.normalize(pseudocounts=bg)
pssm1 = pwm1.log_odds(background=bg)
pssms1[m1.get("ID")] = pssm1
for idx, m2 in enumerate(records2):
pwm2 = m2.counts.normalize(pseudocounts=bg)
pssm2 = pwm2.log_odds(background=bg)
pssms2[m2.get("ID")] = pssm2
# build reverse complement
if args.rc:
rc_pssm2 = pssm2.reverse_complement()
rc_pssms2[idx] = rc_pssm2
result_table = []
# compare motifs
for idx1, pssm1 in pssms1.items():
for idx2, pssm2 in pssms2.items():
if args.extensively or idx1 == idx2:
row = [idx1, idx2]
for measure in [pearsonr, spearmanr]:
cor, p_value, offset = get_motif_similarity(measure, pssm1, pssm2,
args.min_overlap if args.shift else pssm1.length)
orientation = "+"
if args.rc:
rc_pssm2 = rc_pssms2[idx2]
cor_rc, p_value_rc, offset_rc = get_motif_similarity(measure, pssm1, rc_pssm2,
args.min_overlap if args.shift else pssm1.length)
# if cor < cor_rc:
if p_value > p_value_rc:
cor, p_value, offset, orientation = cor_rc, p_value_rc, offset_rc, "-"
row.extend([cor, p_value, offset, orientation])
result_table.append(row)
# write results to output file
out_file_name = args.out_dir + "/correlation_motifs" + ("_extensively" if args.extensively else "") + (
"_rc" if args.rc else "") + ("_shift_min_overlap=" + str(args.min_overlap) if args.shift else "") + ".txt"
with open(out_file_name, 'w') as csv_file:
file_writer = csv.writer(csv_file, delimiter="\t")
file_writer.writerow(["ID1", "ID2", "cor_pearson", "p_value_pearson", "offset_pearson", "orientation_pearson",
"cor_spearman", "p_value_spearman", "offset_spearman", "orientation_spearman"])
for row in result_table:
file_writer.writerow(row)
