I'm a newbie to cancer genomics, so pardon me if my question sounds a bit vague. I have a situation wherein I have a sequenced patient tumor sample (min. 100x coverage) and I have to detect novel mutations from it. I may or may not have a matched normal in each case. So how do I proceed with the analysis, if I don't have a matched-normal genome for it?
Any help or suggested reading is appreciated. Thanks a lot in advance.
you could check for known polymorphisms by comparing vs the 1000 genomes project and check for known somatic mutations by comparing vs Cosmic
You will be limited for the cases having no normal. For example, when a tumor base differs from the reference, it will not be readily possible to distinguish whether this represents a private SNP or a somatic change. I think Catriona makes a good case for what to do here.
Dear Catriona,
Thanks for your reply but what about other mutation types besides somatic mutations? I would try out the 1000 genomes data and see if I could come up with a "consensus normal".
There are also increasing numbers of germline polymorphisms being reported, such as those from Wang et al., Nat Genet 46, 573, 2014, which may add to those already mentioned by Catriona.
To pick up on polymorphisms and mutations mentioned in publications, could be worth simply googling HGVS-style descriptions of the variants - depends on how much data you're dealing with
Hey Sandeep,
Once you call out the variants, you can compared them using a reference like 1000 genomes or dbSNP annotation. You may use Seattleseq server for annotating which gives you dbSNP and non-dbSNP variants. :)
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics