Opinion on Quality basecall threshold to use

Question

[ricfoz]

Hello everyone

I want to pose everyone a question, i know filtering options depend on the strictness threshold anyone would like, depending on his/her research question,

still...

On a vcf, the quality column, is expressed in phred parameters? ... if not so, which parameters? and, the actual question..

what filtering quality would be the most appropiate to set as lowest in your analysis?, i am interested on the evolution of a bone growth-related gene in humans.

thanks in advance for any interaction

[WouterDeCoster]

That's the quality of the variant call, not of the basecall. The basecall quality is in the fastq files.

[ricfoz]

Alright, and how is the meassure of the quality of the variant call in vcfs expressed?, in some cases i haven't got acces to fastq files, although it is possible to convert bams to fastq as far as i know.

Maybe this conversion allows to see the basecall quality when comming from a bam ?

[WouterDeCoster]

The variant call is a combination of the basecall, the coverage, the strand bias, the sequence context, the cycle number,... among other things, which all add evidence to the variant as being true. Essentially, all metrics the variant caller software uses to determine if this position has a variant yes or no.

Answer 1

[ricfoz]

Alright, i've been searching for further information about what you post, and i have filtered my files to get a consensus using only the variants above QUAL 30 ... Would you say 30 is a fair number to filter on, or should i raise the standard?

Cheers