BQSR: Is it safe to use run_without_dbsnp_potentially_ruining_quality in this case?

johnma

In this particular situation, after calling variants using a genome assembly reference, we would also want to see where there are any variations accumulated in the reporter vector, and as such the reference in this case is the reporter gene sequence (<1.5kb).

I wonder in this case it is safe for me to put on the -run_without_dbsnp_potentially_ruining_quality tag? I haven't noticed any person reporting variants in reporter vectors.

Sheila

@johnma
Hi,

BQSR will not work on such a small number of bases, so you cannot simply run it on the reporter gene by itself. You best option is to re-run the entire analysis for best results. What exactly is your end goal? Do you have reads mapped to the reporter gene? Did you do the mapping after you finished the entire analysis?

-Sheila

Sheila

@johnma
Hi,

It sounds like you want to run on the reporter vector alone, which will not work.

How many samples do you have, and are they whole genome or whole exome?

I don't think you need to use the flag if you run BaseRecalibrator on enough data. The best thing to do is simply add the reporter as an extra contig to your reference. The tool will be able to detect trends based on the other sites where dbSNP does contain variation. The small reporter regions should not be affected.

-Sheila

johnma

The thing is I have already finished calling with the reference genome anyway, so I don't feel like doing the whole calling again.

As for the data, it's a RNA-seq set with 5 samples.

Sheila

@johnma
Hi,

BQSR will not work on such a small number of bases, so you cannot simply run it on the reporter gene by itself. You best option is to re-run the entire analysis for best results. What exactly is your end goal? Do you have reads mapped to the reporter gene? Did you do the mapping after you finished the entire analysis?

-Sheila