SOLiD SNP CALL Problem

[pr0t3us]

Dear All,

I'm working with SOLiD data. We have sequenced different strains of E.Coli with shotgun approach and our aim was to find variants (SNPs or INDELs).
I have an average coverage of 150X and I run the spectral correction on the reads to improve their quality.

I tried different softwares (Bioscope SNP calling pipeline, MAQ SNP calling pipeline and PASS). I obtain strange results:
- my bacteria seem to be diploides!!!!
- The three software give total different results. (SNPs in different positions)

Do you have any suggestion on how to approch a SNP calling with SOLiD data?

Best

Alex

[westerman]

1) It is possible that you are sequencing more than one strain of E. coli. Yes, I know that it should be a pure strain but if it isn't then it might appear to be diploid.

2) How variable (or similar) are your SNP calls? If there is 90% overlap between the programs then this may be fine.

[epigen]

1) SNP callers are generally designed for diploid organisms, check if there is an explicit option for haploid ones.
2) Different parameters (cutoff for mapping quality, base quality, coverage, proximity, location in reads ...) = different SNPs. But the "obvious" SNPs should be detected by all methods. I guess 70% overlap would be okay.