Find SNP in 454HCDiffs.txt

[dschika]

Hey there,

I mapped my reads against a reference consisting of the isotigs of the de novo assembly of the same reads. I'm wondering now if the follwoing approach is really sufficient to detect SNPs in the 454HCDiff.txt:
- get the summary line of each diff: grep '>' 454HCDiffs.txt
- check if the start and end position are identical (SNPs need to be at the same position in the reference)
- check if neither the ref nucleotide nor the var nucleotide is only a gap

- check if the var nucleotide length is 1

Regards,
Thomas

[nickloman]

Yes, that approach will give you a list of putative SNP/SNVs.

But you will want to do further filtering (e.g. on read depth, quality) to get a more trusted set of SNPs.

[dschika]

Thanks for your quick reply!

I thought it would be sufficient to take the 454HCDiffs.txt file, because of the High Confidence. That means that (please see the manual for full details):
- there must be at least 3 non-duplicate reads with the difference
- there must be forward and reverse reads with the difference, unless there are at least 5 reads with quality score over 20

Do you think that those filtering options are still too smooth? Can you perhaps suggest some other values?

Btw: I added another step in my first post.

[nickloman]

It will depend on a number of factors. For example if you have greater coverage then you might want to set the read depth cut-off higher. It will depend also on the quality of your reference genome - that might have errors in it. You need to take a view depending on what you are trying to do.