search

jts / nanopolish

Signal-level algorithms for MinION data
MIT License
559 stars 159 forks source link

vcf2fasta min QUAL #565

Closed MaestSi closed 5 years ago

MaestSi commented 5 years ago

Dear Jared, I would like to know if there are any plans to add a vcf2fasta option for filtering out variants based on quality score (QUAL field) in the vcf file produced by nanopolish. Thanks in advance, Simone

jts commented 5 years ago

Hi Simone,

vcf2fasta is designed to make a new consensus sequence so nanopolish will use all positive-scoring variants so I don't think we should apply a filter here. Do you find that you get a better consensus if you filter the variants?

In general better variant filtering is on our TODO list but it won't happen for awhile.

Jared

MaestSi commented 5 years ago

I am running nanopolish for polishing consensus sequences from amplicons. Based on a few tests I did, I found out that if coverage is about 200X, there is no need for filtering out variants, while if coverage is about 500X, I get sligthly better results when filtering out variants with QUAL < 100. Probably, the increase of coverage above a certain threshold results in a bit of overpolishing. What do you think about it? Simone