Estimating NGS paired-end read insert size (or fragment length) from SAM/BAM files

getinsertsize.py

#!/usr/bin/env python
'''
Automatically estimate insert size of the paired-end reads for a given SAM/BAM file.
Usage: getinsertsize.py <SAM file> or samtools view <BAM file> | getinsertsize.py -
Author: Wei Li

Copyright (c) <2015> <Wei Li>



Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:



The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.



THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.


'''


from __future__ import print_function
import sys;
import pydoc;
import os;
import re;
import fileinput;
import math;
import argparse;

parser=argparse.ArgumentParser(description='Automatically estimate the insert size of the paired-end reads for a given SAM/BAM file.');
parser.add_argument('SAMFILE',type=argparse.FileType('r'),help='Input SAM file (use - from standard input)');
parser.add_argument('--span-distribution-file','-s',type=argparse.FileType('w'),help='Write the distribution of the paired-end read span into a text file with name SPAN_DISTRIBUTION_FILE. This text file is tab-delimited, each line containing two numbers: the span and the number of such paired-end reads.');
parser.add_argument('--read-distribution-file','-r',type=argparse.FileType('w'),help='Write the distribution of the paired-end read length into a text file with name READ_DISTRIBUTION_FILE. This text file is tab-delimited, each line containing two numbers: the read length and the number of such paired-end reads.');

args=parser.parse_args();

plrdlen={};
plrdspan={};

def getmeanval(dic,maxbound=-1):
  nsum=0;  n=0;
  for (k,v) in dic.items():
    if maxbound!=-1 and k>maxbound:
      continue;
    nsum=nsum+k*v;
    n=n+v;
  meanv=nsum*1.0/n;
  nsum=0; n=0;
  for (k,v) in dic.items():
    if maxbound!=-1 and k>maxbound:
      continue;
    nsum=nsum+(k-meanv)*(k-meanv)*v;
    n=n+v;
  varv=math.sqrt(nsum*1.0/(n-1));
  return (meanv,varv);

objmrl=re.compile('([0-9]+)M$');
objmtj=re.compile('NH:i:(\d+)');

nline=0;
for lines in args.SAMFILE:
  field=lines.strip().split();
  nline=nline+1;
  if nline%1000000==0:
    print(str(nline/1000000)+'M...',file=sys.stderr);
  if len(field)<12:
    continue;
  try:
    mrl=objmrl.match(field[5]);
    if mrl==None: # ignore non-perfect reads
      continue;
    readlen=int(mrl.group(1));
    if readlen in plrdlen.keys():
      plrdlen[readlen]=plrdlen[readlen]+1;
    else:
      plrdlen[readlen]=1;
    if field[6]!='=':
      continue;
    dist=int(field[8]);
    if dist<=0: # ignore neg dist
      continue;
    mtj=objmtj.search(lines);
    # if mtj==None:
    #   continue;
    # if int(mtj.group(1))!=1:
    #   continue;
    #print(field[0]+' '+str(dist));
    if dist in plrdspan.keys():
      plrdspan[dist]=plrdspan[dist]+1;
    else:
      plrdspan[dist]=1;
  except ValueError:
    continue;

#print(str(plrdlen));
#print(str(plrdspan));

# get the maximum value
readlenval=getmeanval(plrdlen);
print('Read length: mean '+str(readlenval[0])+', STD='+str(readlenval[1]));
# print('Possible read lengths and their counts:');
# print(str(plrdlen));

if args.span_distribution_file is not None:
  for k in sorted(plrdspan.keys()):
    print(str(k)+'\t'+str(plrdspan[k]),file=args.span_distribution_file);

if args.read_distribution_file is not None:
  for k in sorted(plrdlen.keys()):
    print(str(k)+'\t'+str(plrdlen[k]),file=args.read_distribution_file);


if len(plrdspan)==0:
  print('No qualified paired-end reads found. Are they single-end reads?');
else:
  maxv=max(plrdspan,key=plrdspan.get);
  spanval=getmeanval(plrdspan,maxbound=maxv*3);
  print('Read span: mean '+str(spanval[0])+', STD='+str(spanval[1]));
  # print('maxv:'+str(maxv));

Awesome! The span/length writeouts are super useful for potting distribution graphs - thanks so much :)

Am trying it, thank you...

Hello,
Thank you for this tool !
I am feeding getinsertsize.py with overlapping paired-end reads. I get two metrics: "read size" and "read span". Is the "read span" the insert size + the 2x the number of cycles? I am asking because I find it confusing that in the first case "read" really means sequenced read, but in the second it would mean the distance between first base sequenced for reads 1 and 2 right? The outer distance in some sense.

Hello,
Thank you for this tool !
I am feeding getinsertsize.py with overlapping paired-end reads. I get two metrics: "read size" and "read span". Is the "read span" the insert size + the 2x the number of cycles? I am asking because I find it confusing that in the first case "read" really means sequenced read, but in the second it would mean the distance between first base sequenced for reads 1 and 2 right? The outer distance in some sense.

I have the same question

Read his original blog post and comments there for more details.