SNP filtering

[gfmgfm]

I ran samtools pileup and varfilter on human capture Ilumina data.
Does anyone has suggestions on how to filter the output in order to get reliable homozygous and heterozygous SNPs?

[nickloman]

I'd suggest looking at VarScan and playing with the settings.

http://varscan.sourceforge.net/

[drio]

pileup is deprecated. Use mpileup instead. It computes BAQ (check same link) and uses it for the SNP calling. It seems BAQ helps reduce false positives (check alignment example at the bottom in the link above).

[gfmgfm]

Thanks a lot for the suggestion. I already tried VarScan. The problem is that it does not give columns with total coverage for the position, so it is difficult to decide whether it is a reliable heterozygous/homozygous SNP. For example here are two lines from an output of VarScan:
1.chrM 16236 C A 0 35 100% 0 1 0 93 0.98
2. chrM 16236 C T 0 275 100% 0 2 0 93 0.98

I thought that in order to distinguish b/w reliable homozygous/heterozygous SNPs, I need to know the ratio of the "A" coverage (line 1) relative to the total coverage for that position. I would also like to know what is the ratio between the most frequent nucleotide to the second frequent nucleotide.
without this info how can I tell, for example in line 1, if it is a reliable SNP and what kind of SNP?

[gfmgfm]

drio, thanks a lot for the reply. I looked at the mpileup. If I have already results and conclusions from pileup results - is it OK to use them?

[drio]

As soon as you used reasonable filters (check protocol in FAQ for a starting point) yes. pileup has been used with great level of success in various papers.

[gfmgfm]

ok, thanks.

[bioman1]

Varscan SNV

Dear all,
I am new to NGS analysis. I have used bowtie (ver:bowtie-0.12.7) for aligning reference sequence (fastq format) with two paired end files of illumina reads (fastq format). Then I used SAM tools (ver:samtools-0.1.18) and made a 'mpileup' file. Then I have used Varscan (ver:.v2.2.11) for variant calling. I used "pileup2snp' command (with default parameters) to determine SNV and for heterozygosity & homozygosity.

1. The output gives in colums and I have below as rows for easy reading
Output:
Chrom:gi|53564564|gb|JH556356.3
Position:1781287
Ref:T
Cons:Y
Reads1:7
Reads2:2
VarFreq:22.22%
Strands1:2
Strands2:2
Qual1:27
Qual2:26
Pvalue:0.98
MapQual1:1
MapQual2:1
Reads1Plus:5
Reads1Minus:2
Reads2Plus:1
Reads2Minus:1
VarAllele:c


2. Any any one can tell me how identify SNV (how many of them are heterzygous & homozygous) with the above output ?. I have searched this forum, I could not find any help.

[soban]

SNP Filtering

Dear Fellows,
I am new to NGS technologies, i ran BWA for mapping my reads then i used GATK-Tool for SNP-Calling, now i want to filter the SNPs, i dont know how to proceed further, please name some tools to filter the SNPs, also how to use that one?.