Samtools (sequence alignment/mapping)

[wisosonic]

hi all,

i am working on a project that i should find SNPs in the exome. so i installed Bowtie and i run it normaly with a set of reads. the result is in text format.

Now i want to work with Samtools to map this reads on the genome and to find SNPs. I took a look on samtools manual, but its so complicated and i dont know how do i start.
Its mentioned that samtools supports SAM/BAM format as input, but i think i didnt understand the right way to proceed with samtools, like : what format should i get after using Bowtie to be the samtools's input, and then what should i do with that input ....

so i will be very thankfull if someone could tell me those steps to do, cz seaching on the web its not always the best way.

thanks in advance
wassim

[n00c]

It seems that most people prefer BWA over Bowtie for resequencing and targeted/capture sequencing like in your case, and Bowtie over BWA for RNA-seq (reason being that Bowtie does not handle indels correctly which may give rise to higher false discover rate with SNPs). So you may want to look at BWA if you haven't yet -- but perhaps you already chose Bowtie for some other specific reason.

Getting back to your question -- you would need to get Bowtie (if that's what you want to use) to output BAM format. For SNP calling, the "mpileup" command in Samtools is the way to go. You may also want to remove PCR duplicates before you run Samtools. So I would recommend: (1) Map reads using BWA (or if you have a specific reason to use Bowtie, then Bowtie) producing BAM files, (2) Run Picard MarkDuplicates on the BAM files, (3) Run Samtools mpileup to produce VCF files containing SNP calls, (4) Validate your SNP calls by looking at Ti/Tv ratio, dbSNP concordance etc.

Some info about mpileup:

http://samtools.sourceforge.net/mpileup.shtml

[wisosonic]

Thank you so much ... i will check BWA, i heard about it but i didn't tested it yet.