Question: Problems while annotating with snpEFF
0
gravatar for kumarnaren13
4.5 years ago by
kumarnaren1310
kumarnaren1310 wrote:

Hi I am trying to annotate the variants in a vcf file using snpEFF, but its giving multiple annotations for a single position. Any suggestions on how to choose the best or most appropriate one?

Thanks in advance

snp • 1.8k views
ADD COMMENTlink modified 4.5 years ago by Denise CS5.1k • written 4.5 years ago by kumarnaren1310
1

but its giving multiple annotations for a single position

life... again...

Any suggestions on how to choose the best or most appropriate one?

use the "worst" one. e.g: non-synonymous > synonymous. But it depends of your needs...

ADD REPLYlink written 4.5 years ago by Pierre Lindenbaum131k

on your needs

and the context. For example, the example you provided will most probably only be seen with alternative splicing or multiple reading frames. I wonder how one picks the best fit annotation without falling into the cherry-picking conundrum.

ADD REPLYlink written 4.5 years ago by RamRS30k

Exactly, I was wondering whether there is a way to figure this out as I have to parse almost 100 vcf files so that I can write a script?

ADD REPLYlink written 4.5 years ago by kumarnaren1310

Not really. Scripts can automate monotonous tasks, but they cannot do stuff that need subjective perspectives - not basic tools at least.

ADD REPLYlink written 4.5 years ago by RamRS30k
2
gravatar for Denise CS
4.5 years ago by
Denise CS5.1k
UK, Hinxton, EMBL-EBI
Denise CS5.1k wrote:

Firstly, I'd suggest checking some of the results with another annotation tool. Perhaps get a few of those variants giving multiple annotations and confirm the pattern with the rather popular and powerful Variant Effect Predictor, or VEP. As suggested already in the comments here, the multiple annotation will be due to alternatively splicing and there is nothing much we can do apart from trying to chose one or a few "best" ones based on certain parameters and based on the question you want to answer, really. Let's suppose you are interested in the most severe consequence of a given variant. If so, choose the most severe consequence with the VEP(--most_severe flag). If you want to focus on the reliability of a transcript, you can pick the annotations for Ensembl genes that are golden (identical annotation between Ensembl and HAVANA). You can also filter out those that have lower TSL (use the --tsl flag on the VEP script) or choose the 'canonical' (find out what is deemed as canonical transcript by Ensembl) transcript (use the --canonical flag), and so on so forth...

ADD COMMENTlink modified 4.5 years ago • written 4.5 years ago by Denise CS5.1k
1

Thanks for the suggestion, it was indeed helpful to try VEP at least for my case.

ADD REPLYlink written 4.5 years ago by kumarnaren1310
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1959 users visited in the last hour