VEP and Annovar Annotation

Question

[BAGeno]

Hi,

I am trying to figure out which tool VEP or Annovar, is most accurate in terms of SIFT and Polyphen annotation. I annotated my file with both Annovar and VEP. I got about 2386 SIFT deleterious variants and 1547 Polyphen deletrious from annovar. And from VEP I got 37323 SIFT deleterious and 26530 polyphen deleterious. But problem with VEP is that, it annotate different transcripts of a single genes, so there are many records for single snp.

So I am confused as to what tool I should use to find out deleterious variants.

[dyollluap]

which VEP options are you running?

[BAGeno]

I am using these options --sift b, --polyphen b, --ccds, --hgvs, --symbol, --regulatory, --biotype

[dyollluap]

Try with --per_gene and --coding_only to avoid the thousands of transcripts problem.

Answer 1

[Titus]

Hi,

I switched from Annovar to VEP 2 years ago because the database used by VEP is more accurate for human and for some examples i saw. May be you can filter out non coding transcript variation , it reduce a lot your list.

[ATpoint]

Out of curiosity, do you use the web or standalone VEP?

[BAGeno]

I am using standalone VEP

[Titus]

It is standalone VEP annotation.

Answer 2

[ivivek_ngs]

It all depends on what you want to see or what you want to actually fish out in the end. There are also ways in using VEP in more structured way but with some filtrations criteria that gives a better fit. I have moved out of ANNOVAR a long time back and have been using it now for a long time, either you can use with RefSeq or with ENSEMBL, depends entirely upon your choice and the crieterion you want to set. There is also now snpEFF , here are some blog and blog2 that lists out the differences. There is a prioritisation system and ontology for all these tools that you need to understand first.

This paper should also be a good read. It is all about what kind of effect you want and to be honest these values of effects of variants does not really give much of the information unless you really do transcriptomic analysis of these patients/cell lines or your subject of study to understand the real impact.

[BAGeno]

I want two things

First is annotation of whole genome.

Second is to check pharmacogenomically important deleterious variants in whole genome so what do you think which one is most suitable for these tasks.