Multiple sample comparion or Segregation analysis

[2007lab]

Hi,
Wondering if there has been any reports of multiple samples comparison done for variants?
Say in the TCGA or 100 exomes projects from different centres, has anyone developed scripts to compare multiple samples, weighted on their coverage and confidence scores?
This would be also useful for segregation analyses in Mendelian disorders... to follow up linkage.

[schelhorn]

I would be also interested in comparing SNPs and short Indels from multiple samples (patients with different genotypes) .The GATK and Samtools SNP-discovery modules do not seem to include a comparison of multiple samples. Perhaps they shouldn't, and this should better be done externally. Does anyone know of an R package or some in-house scripts that can simplify the job? Should be a pretty common task in variant discovery. I basically would like to throw the pileups of two patient-groups at a tool and get back the SNPs/indels that vary significantly between the groups (given a suitable statistical model).

[ntremblay]

Hi,

you can use any scripts or tools that intersect parameter such as rsID or chromosomal positions to compare a set of snp called from different sample...

Excel can do the job for exome sequencing as you should call about 20000 snps from a sample ...

Otherwise, a script like compare from VarScan will return the intersect, difference, etc. from a file, but it takes only 2 files at a time,

If you want to do segregation analysis, start with the last generation and move up the pedegree with your interesting snp ...

Might also want to run all your snp trough polyphen or SIFT to do this analysis on snps that are predicted to be damaging,

Hope that helps,