Finding mutations from bam files

Question

[Gene_MMP8]

I have two bam files(~17GB) for both tumor and normal sample in case of pancreatic cancer. Now I want to find the mutations in the tumor sample. How do I do that? Should I check the variants for tumor vs normal sample, or tumor vs reference genome(hg19). Pleas excuse my ignorance as I am relatively new to the field and still has got a lot to catch up. Thanks

Answer 1

[WouterDeCoster]

Have a look at the GATK best practices for somatic mutations as a (good) starting point

[Gene_MMP8]

Thanks for the link. Can you suggest some command line tools?

[WouterDeCoster]

First read the pages, then ask again. The tools from the best practices are command line tools.

Answer 2

[MAPK]

Yes, the first thing you would want to do is create a VCF file using GATK tools as WouterDeCoster has mentioned. You can then use different analyses to call for variants that are statistically significant from your normal sample (so you will be comaparing tumour vs normal samples). If you have many unrelated normal samples to compare with, you could also adopt statistical approaches to find mutations in your samples. Reading this paper could give you some insight: http://www.nejm.org/doi/full/10.1056/NEJMoa1301689#t=article

[Gene_MMP8]

1. SO first I am comparing reference vs tumor and creating the vcf file? and then tumor vs normal to call variants?
2. Just to clarify, when you say, "if you have many unrelated normal samples to compare with", what do you mean? I have only two bam files normal and tumor taken from a single patient. By sample do you mean different patients? Or different sites of the body?

[MAPK]

If you already have bam files, then use GATK protocol to create VCF file. Then do tumour vs normal comparison.

2. Just to clarify, when you say, "if you have many unrelated normal samples to compare with", what do you mean? I have only two bam files normal and tumor taken from a single patient. By sample do you mean different patients? Or different sites of the body?

You don't have an option to work with other statistical tests such as burden and SKAT test since you only have two samples.

[Gene_MMP8]

Thanks a lot for clearing that out. Really appreciate it.

[WouterDeCoster]

If an answer was helpful you should upvote it, if the answer resolved your question you should mark it as accepted.