crossfilt 0.1.1__py3-none-any.whl

crossfilt/identical_reads.py

@@ -0,0 +1,139 @@
+#!/usr/bin/python3
+
+import sys
+import argparse
+import pysam
+import logging
+import array
+from timeit import default_timer as timer
+import math
+import os
+from collections import defaultdict
+import importlib.metadata
+
+__version__ = importlib.metadata.version('crossfilt')
+
+def main():
+  parser = argparse.ArgumentParser(
+                      prog='crossfilt-filter',
+                      description='Outputs reads from bam1 that that have identical contig, position, CIGAR string, and XF tag (optional) in bam2')
+  
+  parser.add_argument("-x", "--xf", required=False, help="Require identical XF tag", action="store_true")
+  parser.add_argument("bam1", help="Input bam files.")
+  parser.add_argument("bam2", help="Input bam files.")
+  parser.add_argument('--version', action='version',
+                    version='CrossFilt v{version}'.format(version=__version__))
+  
+  args = parser.parse_args()
+  use_xf = args.xf
+  
+  name_sorted = False
+  
+  if not os.path.exists(args.bam1 + ".bai"):
+    print("Warning: " + args.bam1 + ".bai not found. Assuming files are filtered and sorted by read name.", file=sys.stderr)
+    name_sorted = True
+    
+  if not os.path.exists(args.bam2 + ".bai"):
+    print("Warning: " + args.bam2 + ".bai not found. Assuming files are filtered and sorted by read name.", file=sys.stderr)
+    name_sorted = True
+    
+  
+  # This function will pull read pairs from two position sorted files. It will 
+  # cache reads from the second argument 
+  
+  def read_pair_generator(bam1, bam2, chrom):
+    
+    read2_iter = bam2.fetch(chrom)
+    read2_dict = defaultdict()
+    last_read2_pos = 0
+    
+    for read1 in bam1.fetch(chrom):
+      
+      qname = read1.query_name
+      pos   = read1.reference_start
+      r1    = read1.is_read1
+      
+      read1_id = qname + str(r1)
+      
+      # add reads from file2 till the position in that file is greater than this read
+      while last_read2_pos <= pos:
+        try:
+          read2 = next(read2_iter)
+          last_read2_pos = read2.reference_start
+          read2_dict[read2.query_name + str(read2.is_read1)] = read2
+        except StopIteration:
+          break
+      
+      if read1_id in read2_dict:
+        yield read1, read2_dict[read1_id]
+        del read2_dict[read1_id]
+        
+      # clear out the cache of reads that we have passed
+      for read2_id, read2 in read2_dict.items():
+        if read2.reference_start < pos:
+          del read2_dict[read2_id]
+    
+  
+  def get_read_count(file):
+    contig_list = []
+    total_reads = 0
+    index_stats = file.get_index_statistics()
+    
+    for i in index_stats:
+      if i[3] != 0:
+        contig_list.append(i[0])
+        total_reads += i[3]
+  
+    return total_reads, contig_list
+  
+  
+  SAMFILE1 = pysam.AlignmentFile(args.bam1, "rb")
+  SAMFILE2 = pysam.AlignmentFile(args.bam2, "rb")
+  OUTFILE  = pysam.AlignmentFile('-', "wb", template=SAMFILE1)
+  
+  if not name_sorted:
+    file1_total_reads, file1_contigs = get_read_count(SAMFILE1)
+    file2_total_reads, file2_contigs = get_read_count(SAMFILE2)
+     
+    # this will be more efficient if bam2 is the smaller file
+    i = matched = 0
+    for contig in file1_contigs:
+      for read1, read2 in read_pair_generator(SAMFILE1, SAMFILE2, contig):
+        i += 1
+        if not read1.reference_start == read2.reference_start: continue
+        if not read1.reference_name  == read2.reference_name: continue
+        if not read1.cigarstring     == read2.cigarstring: continue
+        
+        if use_xf:
+          if not read1.has_tag("XF"): continue
+          if not read2.has_tag("XF"): continue
+          if not read1.get_tag("XF") ==  read2.get_tag("XF"): continue
+    
+        matched += 1
+        OUTFILE.write(read1)
+    
+    print(str(matched) + ' (' + str(round(100*matched/i,2)) + '%) successfully matched', file=sys.stderr)
+  
+  else:
+    iter1 = SAMFILE1.fetch(until_eof = True)
+    iter2 = SAMFILE2.fetch(until_eof = True)
+      
+    i = matched = 0
+    for read1, read2 in zip(iter1, iter2):
+      i += 1
+      # check read names to make sure they match
+      if not read1.query_name == read2.query_name:
+        sys.exit("Error: Read number " + str(i) + " query names are not identical (" + read1.query_name + " and " + read2.query_name + ")\nUse position sorted files or filter and sort your bam files by name.")
+      
+      if use_xf:
+        if not read1.has_tag("XF"): continue
+        if not read2.has_tag("XF"): continue
+        if not read1.get_tag("XF") ==  read2.get_tag("XF"): continue
+    
+      matched += 1
+      OUTFILE.write(read1)
+    
+    print(matched,"(",round(100*matched/i,2),"%) successfully matched",file=sys.stderr)
+  
+if __name__ == '__main__':
+    main()