'''

Created on Apr 13, 2011

@author: cambray

'''

import os

import re

from basic_seq_manip import revcomp

def parse_features(file="/Users/cambray/Dropbox/BIOFAB/Wet Lab/APE/Features/Default_Features.txt"):

"""

parse txt file containing name and sequence of features

for now, tab separated

"""

h = open(file, 'r')

features = []

for l in h:

tabs = l.split("\t")

features.append((tabs[0], tabs[1].lower()))

h.close()

return features

def find_features(sequences, features, direction="+/-",internal=False):

"""

find only complete features

Does not take gap or incomplete features into account

- sequence can be a simple sequence or a list of sequences

- features is a list of (name, sequence) 2-uple

"""

isstring=False

if isinstance(sequences, str):

isstring=True

sequences=[sequences]

# compile feature pattern for efficiency

nfeatures=[]

for i in range(len(features)):

if "+" in direction:

nfeatures.append(('%s'%features[i][0], features[i][1], re.compile(features[i][1])))

if "-" in direction:

revcompfeat = revcomp(features[i][1])

nfeatures.append(('(-)%s'%features[i][0], revcompfeat, re.compile(revcompfeat)))

# find position of all occurences

mapping = []

for seq in sequences:

matches = []

seq = seq.lower()

for feat in nfeatures:

for match in feat[2].finditer(seq):

matches.append((match.start(), match.end(), feat[0]))

rm = []

if not internal:

# get rid of internal features

for i in range(len(matches)-1):

if not i in rm:

for j in range(i+1, len(matches)):

if not j in rm:

# feature j inside i

if matches[j][0] >= matches[i][0]\

and matches[j][1] <= matches[i][1]:

rm.append(j)

# feature i inside j

elif matches[i][0] >= matches[j][0]\

and matches[i][1] <= matches[j][1]:

rm.append(i)

break

mapping.append([])

for i in range(len(matches)):

if not i in rm:

mapping[-1].append(matches[i])

mapping[-1] = sorted(mapping[-1], key=lambda matches: matches[0])

if isstring:

mapping = mapping[0]

return mapping

def format_annotation(mapping, html=True):

"""

generate formated string to print

"""

transform=False

if isinstance(mapping[0], tuple):

mapping = [mapping]

transform=True

lines=[]

for map in mapping:

# populate sublists according to overlap

lines.append([])

topass = []

while len(topass)!=len(map):

current_stop=0

lines[-1].append('')

for i in range(len(map)):

if i in topass:

continue

if map[i][0]>=current_stop:

feature_len = map[i][1]-map[i][0]-2

if len(map[i][2]) <= feature_len:

feature_len -= len(map[i][2])

feature_str='='*(feature_len/2)+map[i][2]+'='*(feature_len/2+feature_len%2)

else:

feature_str=map[i][2][:feature_len]

if html:

lines[-1][-1]+="""%s<a title="%s">&#60;%s&#62;</a>""" % (" "*(map[i][0]-current_stop), map[i][2], feature_str)

else:

lines[-1][-1]+="%s<%s>" % (" "*(map[i][0]-current_stop), feature_str)

topass.append(i)

current_stop = map[i][1]

lines[-1][-1]+="\n"

if transform:

lines = lines[0]

return lines

def annotate(seq_path, feat_path):

"""

extract and annotate sequences in a folder

"""

seq_files = [file for file in os.listdir(seq_path) if (re.search("\.seq$", file))]

sequences = []

for file in seq_files:

sequences.append("")

h = open("%s/%s" %(seq_path,file), "r")

for l in h:

if not ">" in l:

sequences[-1] += l.strip("\n\s").lower()

h.close()

features = parse_features(feat_path)

mapping = find_features(sequences, features)

for i in range(len(seq_files)):

print seq_files[i]

for j in range(len(mapping[i])):

print "\t%s..%s\t%s" % (mapping[i][j][0],mapping[i][j][1],mapping[i][j][2])

return

if __name__ == '__main__':

s= "ATATCATTTAAAATTTATCAAAAAGAGTATTGACTTAAAGTCTAACCTATAGGATACTTACAGCCATCGAGAGTTTGGAATTCATTAAAGAGGAGAAAGGTACCATGAGCAAA"

feat = find_features(s, parse_features(), internal=True)

f = format_annotation(feat, html=False)

print s

for l in f:

print l