Hi there,
I'm trying to find a strategy to call core SNPs in a collection of hundreds of bacterial genomes. I've been able to get some of them using kSNP tool Scalable SNP Analyses of 100+ Bacterial or Viral Genomes, but we think that tool is loosing many of them.
I'm thinking of using a reference-based method.
Thanks so much for your help, Best, Bernardo
Citation: Gardner SN, Slezak T (2010) Scalable SNP Analyses of 100+ Bacterial or Viral Genomes. J Forensic Res 1:107. doi:10.4172/2157-7145.1000107
2 answers
You should definitely use multiple methods to generate results. In comparison between methods you'll be more likely to catch issues generated by a particular approach, and the marginal contribution of very different variant detection methods often far outweighs the trouble of running them.
I develop a reference-based detection method, freebayes, which is designed for and has been used in exactly this kind of context.
Running freebayes should be very easy provided you have a reference and have aligned your data against it. Please ping me or the freebayes mailing list if you have any questions.
There are lots of methods by which to align short reads and call variants, of which GATK performs the best. There can be little concordance between methods at times, particularly with indel calls so it's important.
I've written a short blog post on using GATK with the updated version of haplotype caller which can now be set to ploidy =1.
https://approachedinthelimit.wordpress.com/2015/10/09/variant-calling-with-gatk/
Let me know if this is any use to you!
Log in to answer this question.
