Variant Calling

Question

[Maryam]

Hello,

First of all thank you so much, this platform is so helpful.

My question is I have a list of variants along with Qual, DP(INFO), DP(FORMAT) and AF values for all of my variants. Also I have Bam% for my alt allele. Now I want to select my best variants based on these values.

What would be the best criteria??

I really need good suggestions because every variant has different Qual, DP(INFO), DP(FORMAT) and AF values.

Thank you!

Answer 1

[Fernando y un desagüe]

I am doing my graduation project about VC, so if I am wrong someone could correct me.

1) I recommend you to get your vcf files annotated if you have a reference genome using for example SNPeff. This gives you a lot of useful information, like the impact of your Variant Calling since at this point, you could have on that file, for example synonimous variants that cant tell much about whatever you are looking for.

2) Right now, the filter I use to manage my .vcf annotated file is: Quality > 30 and Impact == HIGH.

http://pcingola.github.io/SnpEff/

Keep in mind that Quality is based on phred score and it is not really important nowadays because it is usually very high due to the advances in sequencing and once it is over 20, it is acceptable because this means that whatever it qualifies is 99% accurate, with a 1% chance of error to give the wrong ALT so you might look for another key-factor too.

What I am trying to say is there is not a perfect criteria, you might need a combination depending of what you wanna do with the vcf file.

I hope this is at least a starting point tho :)

Answer 2

[Prash]

Yes, it is. You could also check the depth and set it to minimum 10 as a pilot.

BTW, it us Q30 not 20 :-)