def read(handle, format):
alphabet = IUPAC.unambiguous_dna
counts = {}
if format == "pfm":
# reads the motif from Jaspar .pfm file
letters = "ACGT"
for letter, line in zip(letters, handle):
words = line.split()
#if there is a letter in the beginning, ignore it
if words[0] == letter:
words = words[1:]
counts[letter] = map(float, words)
motif = Motif(alphabet, counts=counts)
elif format == "sites":
# reads the motif from Jaspar .sites file
instances = []
for line in handle:
if not line.startswith(">"):
break
# line contains the header ">...."
# now read the actual sequence
line = handle.next()
instance = ""
for c in line.strip():
if c == c.upper():
instance += c
instance = Seq(instance, alphabet)
instances.append(instance)
instances = Instances(instances, alphabet)
motif = Motif(alphabet, instances=instances)
else:
raise ValueError("Unknown format %s" % format)
motif.mask = "*" * motif.length
return motif
def load_from_json(self, file_path):
'''
Sets up all memeber variables based on the stored data in the JSON file.
Parameters
----------
file_path: str
Path to the JSON file that will be loaded in.
Returns
-------
None
'''
file_reader = json.load(open(file_path, 'r'))
self.cluster_id = file_reader['cluster_id']
for op in file_reader['operons']:
temp_op = Operon(operon_id=op['operon_id'],
genome_fragment_name='imported_cluster',
genome_accession=op['genome_accession'],
genome_features='imported',
strand='/')
temp_op.features = op['features']
temp_op.promoter = op['promoter']
self.operons.append(temp_op)
self.filtered_promoters = file_reader['filtered_promoters']
for m in file_reader['motifs']:
self.motifs.append(Motif(instances=Instances(m)))
def get_cluster_motifs(cluster_id, path_to_saved_clusters=cluster_jsons_path):
'''
Parse out the motif for a specific cluster.
Parameters
----------
cluster_id: str
The cluster of interest.
path_to_saved_clusters: str
The path to the directory that is holding the saved JSONs for the saved clusters
Returns
-------
motifs: str
A string that is made of the concatenated consensus sequences of the motifs a part of the cluster.
'''
motifs = []
for file in os.listdir(path_to_saved_clusters):
if file.split('.')[0] == cluster_id:
file_reader = json.load(open(path_to_saved_clusters + file, 'r'))
motifs_from_file = file_reader['motifs']
for m in motifs_from_file:
temp_motif = Motif(instances=Instances(m))
#temp_motif.weblogo(fname=logo_path+cluster_id+"_"+str(motifs_from_file.index(m))+'.png')
motifs.append(str(temp_motif.consensus))
return ' | '.join(motifs)
def main():
seq_list=[]
pos_sequences = SeqIO.parse(open(input_file_sig70_compiled), 'fasta')
count = 0
for seq in pos_sequences:
sequence = seq.seq[28:34]
if sequence[4] == "G":
count+=1
print(count)
print(sequence)
seq_list.append(sequence)
motif = Motif(instances=Instances(instances=seq_list))
PSSM = motif.pssm
print(PSSM.max)
print(PSSM.consensus)
print(PSSM)
amrZ_sequences = SeqIO.parse(open(input_file_AmrZ), 'fasta')
positions = []
for a_seq in amrZ_sequences:
print(a_seq.seq)
prib_10_pos = evaluate_prib10(a_seq.seq, PSSM)
if prib_10_pos != []:
positions.append(prib_10_pos)
print(positions)
def get_cluster_motifs(cluster_id):
'''
Parse out the motif for a specific cluster.
Parameters
----------
cluster_id: str
The cluster of interest.
path_to_saved_clusters: str
The path to the directory that is holding the saved JSONs for the saved clusters
Returns
-------
motifs: str
A string that is made of the concatenated consensus sequences of the motifs a part of the cluster.
'''
path_to_saved_clusters = '/Users/ichaudr/Documents/UMBC/Lab-Erill/Isaac/Vibrio_SOA/Saved Clusters/complete_clusters/'
motifs = []
for file in os.listdir(path_to_saved_clusters):
if file.split('.')[0] == cluster_id:
file_reader = json.load(open(path_to_saved_clusters + file, 'r'))
motifs_from_file = file_reader['motifs']
for m in motifs_from_file:
temp_motif = Motif(instances=Instances(m))
motifs.append(str(temp_motif.consensus))
return motifs
def ic_at(motif, other, offset):
'''
Caculates the information content, ic, for a specific alignment. The approach makes a temporary motif object containing the overlapping sequences in the alignemnt and taking the average of the pssm.
Parameters
----------
motif, other: Motif objects
The motifs of interest
offset: int
The offset value that results in the alignment of interest.
'''
#Pull the sequences containined in the aligned region of the motifs from each of the motif instances.
alignment_len = min(len(motif) - offset, len(other))
motif_seqs = [
site[offset:alignment_len + offset] for site in motif.instances
]
other_seqs = [site[:alignment_len] for site in other.instances]
# Create the motif and compute the IC
amotif = Motif(instances=Instances(motif_seqs + other_seqs))
amotif.pseudocounts = dict(A=0.25, C=0.25, G=0.25, T=0.25)
#print('Motif Seqs: ' , motif_seqs)
#print('Other Seqs: ' , other_seqs)
#print('Offset ', offset)
#print('IC: ' , amotif.pssm.mean(), '\n\n')
return amotif.pssm.mean()
def build_pssm_35(sequences):
seq_list = []
for seq in sequences:
seq35 = seq[28:34]
seq_list.append(seq35)
motif = Motif(instances=Instances(instances=seq_list))
PSSM = motif.pssm
return PSSM
def build_pssm_10(sequences):
seq_list = []
for seq in sequences:
seq10 = seq[52:58]
seq_list.append(seq10)
motif = Motif(instances=Instances(instances=seq_list))
PSSM = motif.pssm
return PSSM
def build_pssm(sequences):
seq_list = []
for seq in sequences:
seq_list.append(seq)
motif = Motif(instances=Instances(instances=seq_list))
motif = motif.counts.normalize(0.01)
PSSM = PositionSpecificScoringMatrix(alphabet=IUPAC.unambiguous_dna,
values=motif)
return PSSM
def printCons10(sequences):
seq_list = []
for seq in sequences:
seq10 = seq[52:58]
seq_list.append(seq10)
motif = Motif(instances=Instances(instances=seq_list))
PSSM = motif.pssm
print(PSSM.consensus)
print(PSSM)
def printCons35(sequences):
seq_list = []
for seq in sequences:
seq35 = seq[28:34]
seq_list.append(seq35)
motif = Motif(instances=Instances(instances=seq_list))
PSSM = motif.pssm
print(PSSM.consensus)
print(PSSM)
def read(handle):
"""Parse an AlignACE format handle as a Record object."""
record = Record()
line = next(handle)
record.version = line.strip()
line = next(handle)
record.command = line.strip()
mask = None
number = None
for line in handle:
line = line.strip()
if line == "":
pass
elif line[:4] == "Para":
record.parameters = {}
elif line[0] == "#":
seq_name = line.split("\t")[1]
record.sequences.append(seq_name)
elif "=" in line:
par_name, par_value = line.split("=")
par_name = par_name.strip()
par_value = par_value.strip()
record.parameters[par_name] = par_value
elif line[:5] == "Input":
record.sequences = []
elif line[:5] == "Motif":
words = line.split()
assert words[0] == "Motif"
number = int(words[1])
instances = []
elif line[:3] == "MAP":
alphabet = "ACGT"
instances = Instances(instances, alphabet)
motif = Motif(alphabet, instances)
motif.score = float(line.split()[-1])
motif.number = number
motif.mask = mask
record.append(motif)
elif len(line.split("\t")) == 4:
seq = Seq(line.split("\t")[0])
instances.append(seq)
elif "*" in line:
mask = line.strip("\r\n")
else:
raise ValueError(line)
return record
def transcription_factor_matrices(self):
if not os.path.exists(
self.cache_location / 'transcription_factor_matrices.json'):
tms = taxonid_to_jaspar_matrix_ids(self.taxonid)
write_tms = {k: v.counts for k, v in tms.items()}
with open(
self.cache_location / 'transcription_factor_matrices.json',
'w+') as f:
json.dump(write_tms, f)
self._transcription_factor_matrices = tms
elif not hasattr(self, '_transcription_factor_matrices'):
with open(self.cache_location /
'transcription_factor_matrices.json') as f:
write_tms = json.load(f)
tms = {k: Motif(counts=v) for k, v in write_tms.items()}
self._transcription_factor_matrices = tms
return self._transcription_factor_matrices
#Get the counts to fill the above dictionaries
for file in os.listdir(path_to_clusters):
if file[-4:] != 'json':
continue
cluster_id = file[:-5]
cluster_reader = json.load(open(path_to_clusters + file, 'r'))
#Add info about the number of motifs
num_motifs_per_cluster[cluster_id] = len(cluster_reader['motifs'])
#Add info about the IC per motif
ic_per_motif_per_cluster[cluster_id] = {}
for m in range(len(cluster_reader['motifs'])):
temp_motif = Motif(instances=Instances(cluster_reader['motifs'][m]))
temp_motif_seq = str(temp_motif.consensus)
temp_motif_ic = temp_motif.pssm.mean()
ic_per_motif_per_cluster[cluster_id][temp_motif_seq] = temp_motif_ic
#All stats to be calculated:
avg_num_motifs_per_cluster = statistics.mean(
list(num_motifs_per_cluster.values()))
stdv_num_motifs_per_cluster = statistics.stdev(
list(num_motifs_per_cluster.values()))
max_num_motifs_per_cluster = max(list(num_motifs_per_cluster.values()))
min_num_motifs_per_cluster = min(list(num_motifs_per_cluster.values()))
median_num_motifs_per_cluster = statistics.median(
list(num_motifs_per_cluster.values()))
mode_num_motifs_per_cluster = statistics.mode(
all_motifs = []
#Holds motifs per cluster
motifs_per_cluster = {}
#Import the motifs
for file in tqdm(list(os.listdir(operon_clusters_path)),
desc='Reading in motifs from saved clusters'):
if file.split('.')[-1] != 'json':
continue
cluster_info = dict(json.load(open(operon_clusters_path + file, 'r')))
motifs_per_cluster[cluster_info['cluster_id']] = []
for motif in cluster_info['motifs']:
motif_obj = Motif(instances=Instances(motif))
all_motifs.append(motif_obj)
motifs_per_cluster[cluster_info['cluster_id']].append(motif_obj)
print('-' * 5, ' Total motifs: ', len(all_motifs))
### STEP 2: Bin the motifs based off some threshold weight.
#Holds all motif bins
print('Binning motifs...')
sim_threshold = 0.26
motif_bins = {}
motif_assignments = {}
for m in all_motifs:
def jaspar_matrix_id_to_motif(matrix_id):
return Motif(counts=requests.get(
f'http://jaspar.genereg.net/api/v1/matrix/{matrix_id}/').json()['pfm'])
''' A copy from https://github.com/ErillLab/Transfer_method_analysis/blob/71a042cc8ca0ce03a1d44c84d87f9bb509b6ec4c/src/motif.py#L81 Being used for testing purposes. ''' from Bio.motifs import Motif, Instances import math test_motif = Motif(instances=Instances(["ATCAGTCA", "ATCAGTAA", "ATCTGTCA"])) print(test_motif.consensus)
