How to select a "representative" transcript in multiple transcript variants from annotation with ANNOVAR and gencode

Question

[svlachavas]

Dear Biostars,

based on a recent annotation of resulted vcf files regarding the identification of somatic point alterations in a specific cancer of interest, I used ANNOVAR default settings with gencode annotation. Based on the relative output of files, I noticed the following pattern in a specific column (ANNOVAR_TRANSCRIPTS) regarding multiple variant transcripts for each variant/specific gene, for example:

BRAF:ENST00000479537.1:exon2:c.T83A:p.V28E,BRAF:ENST00000288602.6:exon15:c.T1799A:p.V600E,

However, this poses a general problem in which of the transcripts to select/consider as the “main transcript variant”, as in this example we have two "different" mutations (exon 2 mutation distinct from exon 15 and V600E mutation)-this is for really important for my downstream analysis, as we are searching specific mutations regarding KRAS/BRAF genes-

from previous posts (Only the longest transcript while annotating with Annovar?) I saw that a suggestion would be to choose the longest transcript-however, are there any additional selection criteria to choose one transcript per mutation ?

Thank you in advance,

Efstathios

[GenoMax]

NCBI/EBI's MANE project is trying to address this for specific reasons. One happens to be a MANE plus clinical variant which would be used for

MANE Plus Clinical: Transcripts chosen to supplement MANE Select when needed for clinical variant reporting.

[svlachavas]

Dear GenoMax,

thank you very much for your interesting suggestion- I have never heard it before-do you know or you have used it ? I could use gencode annotations, or you would go alternatively with selected gene symbols to check which are the most "representative" transcripts ?

[GenoMax]

I have not MANE specifically for clinical variant calling but the project is being worked on precisely to provide a clinically stable locations. While that may not strictly reflect biology it should help the situation you are describing in original question.

[Ram]

I don't think so. The closest I've gotten is with the canonical transcript option in VEP.

[svlachavas]

Dear Ram, thank you for your reply-unfortunately already I know the VEP option, we utilize ANNOVAR with an additional custom pipeline, thus we could not change in this current phase of the project