InterVar and Varsome discrepancies in ACMG classification

Question

[cocchi.e89]

I am working on ACMG classification of variants in a cohort of patients, I found that InterVar classification differs (generally more strict for what does it seems to me at the moment) form Varsome, e.g. var X-107939570-C-T is classified as "Pathogenic" by Varsome (https://varsome.com/variant/hg19/X-107939570-C-T) and Uncertain Significance by InterVar (command line tool: "InterVar: Uncertain significance PVS1=0 PS=[0, 0, 0, 0, 0] PM=[0, 1, 0, 0, 0, 0, 0] PP=[0, 0, 1, 0, 1, 0] BA1=0 BS=[0, 0, 0, 0, 0] BP=[0, 0, 0, 0, 0, 0, 0, 0]")

At the same manner var: X-107898624-G-T. Varsome: Likely Pathogenic (https://varsome.com/variant/hg19/X-107898624-G-T) and Intervar: "InterVar: Uncertain significance PVS1=0 PS=[0, 0, 0, 0, 0] PM=[0, 1, 0, 0, 0, 0, 0] PP=[0, 0, 1, 0, 1, 0] BA1=0 BS=[0, 0, 0, 0, 0] BP=[1, 0, 0, 0, 0, 0, 0, 0]"

Does somebody has got a clue for this difference? There are other command line tools to make this annotation automated (make more similar to Varsome "relaxed" interpretation?

Thanks a lot in advance for any help!

[zankov]

Hey I am also working on ACMG classification and was wondering what command you used with InterVar and also how your vcf is structured since I am struggling to get an output in general.

Thanks in advance for your help!

[cocchi.e89]

Just clone the InterVar repo on GitHub:

git clone https://github.com/WGLab/InterVar.git

and then access the InterVar directory in which you find the python script InterVar.py Call it on a normal VCF file with, as example (I use python3.7 but any distribution >3.3 should work properly):

python3.7 InterVar.py  \
    -b hg19 \
    -i <your_input.vcf> \
     --input_type=VCF \
     -o <your_output>

And then you will find your output files where you specified. Please note, in the same InterVar directory you must download Annovar executables (annotate_variation.pl, convert2annovar.pl, retrieve_seq_from_fasta.pl, coding_change.pl, table_annovar.pl, variants_reduction.pl) and the first time you run it will take long in order for Annovar to download all its reference databases in the "humandb" directory that will appear in the same "InterVar" directory where you launch the script.

Answer 1

[manuel.belmadani]

You can actually replicate what InterVar tells you with the VarSome UI.

Example:

PVS1=0 PS=[0, 0, 0, 0, 0] PM=[0, 1, 0, 0, 0, 0, 0] PP=[0, 0, 1, 0, 1, 0] BA1=0 BS=[0, 0, 0, 0, 0] BP=[1, 0, 0, 0, 0, 0, 0, 0]

Why do they get different information? I'm not too familiar with InterVar, but from the looks of it the data is static from: https://github.com/WGLab/InterVar/tree/master/intervardb

Some files have been recently updated (~3 months ago) while others are a couple years old. Whereas VarSome uses ClinVar from this month's release. My guess is that one of difference is just that VarSome uses more recent information, though presumably you would get the same thing if you updated the InterVar files to equivalent versions.

There's seems to be a difference with BP1:

"Missense variant in a gene for which primarily truncating variants are known to cause disease"

Intevar considers this to be true while Varsome does not, because:

Reason it failed
Missense variant in gene COL4A5 that has 488 pathogenic variants of which 392 pathogenic missense variants = 80.3% which is more than maximum of 10.0%.

[mrodrigues.fernanda]

Please check out our automated tool for clinical characterization of germline variants following ACMG guidelines: https://github.com/ding-lab/CharGer

Fernanda Martins Rodrigues | Washington University in St. Louis | Ding Lab