MAQ vs BWA/SAMTools - differences in SNP calling algorithm?

mard · 10-19-2010, 04:12 PM

I was wondering if anyone has come across this before. Basically we’re getting SNPs that were identified using MAQ, and subsequently validated by Sanger sequencing, not being called by BWA/SAMTools on the same set of reads (same sequence.txt files) yet the BWA/SAMTools pileup shows that there is an alternate allele present at the SNP location.

In the example below, for both BWA and MAQ a similar number of reads mapped over the SNP (134 for BWA and 130 for MAQ). The SAMTools pileup shows that there is an alternate allele present (T), yet no SNP is called. I thought that MAQ’s SNP caller and the SAMtools SNP caller used the same algorithm so I’m confused as to why the SNP was not called by SAMTools. It is not being filtered out by varFilter as adding the –p flag doesn’t identify it either.

BWA + SAMtools pileup:
C C 73 0 58 134 ,$,$,$,$,$,$TT,,,T.A..,,,,.t,,,,,,.T.T.......,,,..........t............,........TTTT........tttttt........ttt,tt.............TTt,,t,,,,t,,,t

MAQ + MAQ’s SNP caller output:
C Y 37 130 1.00 63 62 C 250 T

MAQ Pileup:
C 130 @......TT...A,,,....,.....,T,,,,T,,,,...,,,,,,,,,,t,,,,,,,,,,,,.,,,,,,,,TTT,,,,,,,,tttttt,,,,,,,,tttt.t,,,,,,,,,,,,,TTt.....t...t.t

kmcarr · 10-20-2010, 06:26 AM

Did you run samtools.pl varFilter after samtools pileup? The varFilter has a maximum read depth cutoff parameter for SNP calling (-D). The default value is 100. With a read depth of 134 the SNP would be rejected.

mard · 10-20-2010, 03:17 PM

Thanks for the suggestion but I ran varFilter with a read depth cutoff of 1,000,000 so that can't be the reason. Also I ran varFilter with the -p option aswell which prints the SNPs that are filtered out by varFilter and that SNP does not make it into that file either.
So for some reason it's not getting recognised as a SNP by SAMTools yet MAQ was able to call it.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
MAQ SNP calling	bioinfosm	Bioinformatics	10	01-03-2012 01:03 AM
annotation file for maq snp calling	johnsequence	Bioinformatics	0	10-26-2011 11:39 AM
Any software for snp calling other than MAQ?	ericfit	Bioinformatics	1	01-21-2011 10:40 PM
percentage of SNP calling in MAQ	johnsequence	Bioinformatics	1	01-08-2010 10:54 AM
MAQ and SNP calling	michy	Bioinformatics	3	12-11-2009 09:23 AM

10-19-2010, 04:12 PM	#1
mard Member Location: Melbourne Join Date: Jan 2010 Posts: 21	MAQ vs BWA/SAMTools - differences in SNP calling algorithm? I was wondering if anyone has come across this before. Basically we’re getting SNPs that were identified using MAQ, and subsequently validated by Sanger sequencing, not being called by BWA/SAMTools on the same set of reads (same sequence.txt files) yet the BWA/SAMTools pileup shows that there is an alternate allele present at the SNP location. In the example below, for both BWA and MAQ a similar number of reads mapped over the SNP (134 for BWA and 130 for MAQ). The SAMTools pileup shows that there is an alternate allele present (T), yet no SNP is called. I thought that MAQ’s SNP caller and the SAMtools SNP caller used the same algorithm so I’m confused as to why the SNP was not called by SAMTools. It is not being filtered out by varFilter as adding the –p flag doesn’t identify it either. BWA + SAMtools pileup: C C 73 0 58 134 ,$,$,$,$,$,$TT,,,T.A..,,,,.t,,,,,,.T.T.......,,,..........t............,........TTTT........tttttt........ttt,tt.............TTt,,t,,,,t,,,t MAQ + MAQ’s SNP caller output: C Y 37 130 1.00 63 62 C 250 T MAQ Pileup: C 130 @......TT...A,,,....,.....,T,,,,T,,,,...,,,,,,,,,,t,,,,,,,,,,,,.,,,,,,,,TTT,,,,,,,,tttttt,,,,,,,,tttt.t,,,,,,,,,,,,,TTt.....t...t.t

10-20-2010, 06:26 AM	#2
kmcarr Senior Member Location: USA, Midwest Join Date: May 2008 Posts: 1,178	Did you run samtools.pl varFilter after samtools pileup? The varFilter has a maximum read depth cutoff parameter for SNP calling (-D). The default value is 100. With a read depth of 134 the SNP would be rejected.

10-20-2010, 03:17 PM	#3
mard Member Location: Melbourne Join Date: Jan 2010 Posts: 21	Thanks for the suggestion but I ran varFilter with a read depth cutoff of 1,000,000 so that can't be the reason. Also I ran varFilter with the -p option aswell which prints the SNPs that are filtered out by varFilter and that SNP does not make it into that file either. So for some reason it's not getting recognised as a SNP by SAMTools yet MAQ was able to call it.