Comparing sequences between themselves

[ramirob]

Hello,

We ran an experiment where we sequenced WT and mutant isolates from the same particular organism. We have aligned these samples to reference (using BWA) and then found SNPs (using GATK). However, aside from identifying SNPs by comparing to reference, we would like to compare the aligned samples between themselves to look for important differences (eg. WT to Mutant samples). Does anyone have any suggestions on how to approach that? It can certainly be done indirectly by comparing the SNP calls, but I was wondering if there is any existing software that does something like that (e.g. find SNPs between aligned samples as opposed to find SNPs when compared to reference).

Thanks in advance,
Ramiro

[Richard Finney]

Try blseq
You can do it online.

http://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastSearch&PROG_DEF=blastn&BLAST_PROG_DEF=megaBlast&BLAST_SPEC=blast2seq

[ramirob]

Thnaks Richard!

We are talking about huge genome alignments. The samples have already been aligned to reference (this was part of an NGS experiment). My understanding is that bl2seq is more for aligning smaller sequences that have not been aligned before.

[Richard Finney]

How huge?

By the way, there is a command line version for those big batch jobs. It might also handle large jobs that web version can't (maybe?).

[Liam_Gallagher]

Hmm....I would try to use the WT (assembled de-novo) as reference for aligning the mutant and then use a variant caller, to identify variants between two genomes (using the WT as reference).

Also you can take a look at Jspecies. It is a java-based tool that measure the probability if two genomes belonging to the same species or not....and I think you can use your multifasta sequences as input and test your samples for differences. Basically it computes some blast between the samples and for each read you should see the percent of similarity. Then you can investigate the fragments with low similarity....