PageRenderTime 43ms CodeModel.GetById 18ms app.highlight 19ms RepoModel.GetById 1ms app.codeStats 0ms

/tools/solid_tools/solid_qual_stats.py

https://bitbucket.org/cistrome/cistrome-harvard/
Python | 140 lines | 114 code | 24 blank | 2 comment | 46 complexity | 5560d37474ff2b3d8ebf369c696fb229 MD5 | raw file
  1#!/usr/bin/env python
  2#Guruprasad Ananda
  3
  4import sys, os, zipfile, tempfile
  5
  6QUAL_UPPER_BOUND = 41
  7QUAL_LOWER_BOUND = 1
  8
  9def stop_err( msg ):
 10    sys.stderr.write( "%s\n" % msg )
 11    sys.exit()
 12 
 13def unzip( filename ):
 14    zip_file = zipfile.ZipFile( filename, 'r' )
 15    tmpfilename = tempfile.NamedTemporaryFile().name
 16    for name in zip_file.namelist():
 17        file( tmpfilename, 'a' ).write( zip_file.read( name ) )
 18    zip_file.close()
 19    return tmpfilename
 20   
 21def __main__():
 22
 23    infile_score_name = sys.argv[1].strip()
 24    fout = open(sys.argv[2].strip(),'r+w')
 25
 26    infile_is_zipped = False
 27    if zipfile.is_zipfile( infile_score_name ):
 28        infile_is_zipped = True
 29        infile_name = unzip( infile_score_name )
 30    else:
 31        infile_name = infile_score_name
 32    
 33    readlen = None
 34    invalid_lines = 0
 35    j = 0
 36    for line in file( infile_name ):
 37        line = line.strip()
 38        if not(line) or line.startswith("#") or line.startswith(">"):
 39            continue
 40        elems = line.split()
 41        try:
 42            for item in elems:
 43                int(item)
 44            if not readlen:
 45                readlen = len(elems)
 46            if len(elems) != readlen:
 47                print "Note: Reads in the input dataset are of variable lengths."
 48            j += 1
 49        except ValueError:
 50            invalid_lines += 1
 51        if j > 10:
 52            break
 53        
 54    position_dict = {}
 55    print >>fout, "column\tcount\tmin\tmax\tsum\tmean\tQ1\tmed\tQ3\tIQR\tlW\trW"
 56    for k,line in enumerate(file( infile_name )):
 57        line = line.strip()
 58        if not(line) or line.startswith("#") or line.startswith(">"):
 59            continue
 60        elems = line.split()
 61        if position_dict == {}:
 62            for pos in range(readlen):
 63                position_dict[pos] = [0]*QUAL_UPPER_BOUND
 64        if len(elems) != readlen:
 65            invalid_lines += 1
 66            continue
 67        for ind,item in enumerate(elems):
 68            try:
 69                item = int(item)
 70                position_dict[ind][item]+=1
 71            except:
 72                pass
 73    
 74    invalid_positions = 0
 75    for pos in position_dict:
 76        carr = position_dict[pos] #count array for position pos
 77        total = sum(carr) #number of bases found in this column.
 78        med_elem = int(round(total/2.0))
 79        lowest = None   #Lowest quality score value found in this column.
 80        highest = None  #Highest quality score value found in this column.
 81        median = None   #Median quality score value found in this column.
 82        qsum = 0.0      #Sum of quality score values for this column.
 83        q1 = None       #1st quartile quality score.
 84        q3 = None       #3rd quartile quality score.
 85        q1_elem = int(round((total+1)/4.0))
 86        q3_elem = int(round((total+1)*3/4.0))
 87        
 88        try:
 89            for ind,cnt in enumerate(carr):
 90                qsum += ind*cnt
 91                
 92                if cnt!=0:
 93                    highest = ind
 94                
 95                if lowest==None and cnt!=0:  #first non-zero count
 96                    lowest = ind
 97                
 98                if q1==None:
 99                    if sum(carr[:ind+1]) >= q1_elem:
100                        q1 = ind
101                           
102                if median==None:
103                    if sum(carr[:ind+1]) < med_elem:
104                        continue
105                    median = ind
106                    if total%2 == 0: #even number of elements
107                        median2 = median
108                        if sum(carr[:ind+1]) < med_elem+1:
109                            for ind2,elem in enumerate(carr[ind+1:]):
110                                if elem != 0:
111                                    median2 = ind+ind2+1
112                                    break
113                        median = (median + median2)/2.0
114    
115                
116                if q3==None:
117                    if sum(carr[:ind+1]) >= q3_elem:
118                        q3 = ind
119                 
120                
121            mean = qsum/total    #Mean quality score value for this column.
122            iqr = q3-q1
123            left_whisker = max(q1 - 1.5*iqr,lowest)
124            right_whisker = min(q3 + 1.5*iqr,highest)
125            
126            print >>fout,"%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s" %(pos+1,total,lowest,highest,qsum,mean,q1,median,q3,iqr,left_whisker,right_whisker)
127        except:
128            invalid_positions += 1
129            nullvals = ['NA']*11
130            print >>fout,"%s\t%s" %(pos+1,'\t'.join(nullvals))
131
132    if invalid_lines:
133        print "Skipped %d reads as invalid." %invalid_lines
134    if invalid_positions:
135        print "Skipped stats computation for %d read positions." %invalid_positions
136        
137if __name__=="__main__":
138    __main__()
139        
140