1000 Genomes Data

[RichardRocca]

I am trying to count short tandem repeats (STR) from 1000 Genomes Project data using the low coverage BAM files, but am getting the same repeats even on known fast STRs. Should I be looking at other files within the project to do this? I am using SAMTools.

[krobison]

Depending on the aligner & parameters used, these may not have aligned well. In the worst case, no single read can contain the entire repeat array AND unique sequence on either side to enable conclusive mapping -- even if the repeat is not strictly too long, very long repeats will be unlikely to have randomly fragmented in such a manner that they can be uniquely mapped.

STRs are definitely findable in the Watson & Venter BAM files & you might test out your method there. These were done with long read (454 & Sanger respectively) technologies that can read through all but the most ginormous repeats.

My intern last summer developed a tool for this purpose; anyone interested in beta testing it can PM me about access. I'm trying to get some additional data generated to test it prior to releasing & publishing it -- the problem is a dearth of data in which the actual repeat length is known as tested by conventional methods. I know we looked at 1K genome data & thought she was able to pull variation out, but I'd need to check.