filtering mpileup results

Hit · 06-30-2011, 07:20 PM

Does anyone have any suggestions on filtering the mpileup results? We have obtained SNPs/indels according to the command lines on the mpileup website:

samtools mpileup -uf ref.fa aln1.bam aln2.bam | bcftools view -bvcg - > var.raw.bcf (1)
bcftools view var.raw.bcf | vcfutils.pl varFilter -D100 > var.flt.vcf (2)

As in pileup where there is "awk '($3=="*"&&$6>=50)||($3!="*"&&$6>=20)' sample1.flt.txt > sample1.final.txt" suggested to filter the results, are there similar rules necessary to filter mpileup results?

For example, is the 6th column (named QUAL) in the vcf file from (2) the same as the 6th column in the pileup file, and is it appropriate to apply "$6>20" for the vcf file too?

Any suggestion is appreciated.

swbarnes2 · 07-01-2011, 08:46 AM

If you only have one sample in your vcf, filtering on depth might be more useful. You can use the DP value, but the DP4 values will omit poor quality reads, which probably makes it more useful. But it's a trickier awk statement.

If you have multiple sample, QUAL won't help you figure out which samples really have the SNP, and which don't. The best you can do is to use the GQ quality, and unfortunately, you won't have depths of coverage for individual samples.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
filtering the variants	kjaja	Bioinformatics	1	08-07-2013 02:38 AM
sequence filtering	yaximik	Bioinformatics	2	03-13-2012 05:10 AM
mpileup results using -q	Hit	Bioinformatics	0	03-31-2011 08:18 PM
Can pre-filtering reads affect your analysis results?	PFS	Bioinformatics	0	03-24-2011 11:08 AM
mpileup and pileup results	rururara	Bioinformatics	1	03-23-2011 05:05 AM

06-30-2011, 07:20 PM	#1
Hit Member Location: United States, TN Join Date: Jul 2010 Posts: 15	filtering mpileup results Does anyone have any suggestions on filtering the mpileup results? We have obtained SNPs/indels according to the command lines on the mpileup website: samtools mpileup -uf ref.fa aln1.bam aln2.bam \| bcftools view -bvcg - > var.raw.bcf (1) bcftools view var.raw.bcf \| vcfutils.pl varFilter -D100 > var.flt.vcf (2) As in pileup where there is "awk '($3==""&&$6>=50)\|\|($3!=""&&$6>=20)' sample1.flt.txt > sample1.final.txt" suggested to filter the results, are there similar rules necessary to filter mpileup results? For example, is the 6th column (named QUAL) in the vcf file from (2) the same as the 6th column in the pileup file, and is it appropriate to apply "$6>20" for the vcf file too? Any suggestion is appreciated.

07-01-2011, 08:46 AM	#2
swbarnes2 Senior Member Location: San Diego Join Date: May 2008 Posts: 912	If you only have one sample in your vcf, filtering on depth might be more useful. You can use the DP value, but the DP4 values will omit poor quality reads, which probably makes it more useful. But it's a trickier awk statement. If you have multiple sample, QUAL won't help you figure out which samples really have the SNP, and which don't. The best you can do is to use the GQ quality, and unfortunately, you won't have depths of coverage for individual samples.