How to Identify the pathogenicity of a set of variants (by using ACMG guidelines)?

Question

[DanielC]

Dear Friends,

Am trying to find the pathogenicity of a set of variants (by using ACMG guidelines: https://www.ncbi.nlm.nih.gov/pubmed/25741868 ). The vcf file of the variants looks like this:

CHROM   POS     ID      REF     ALT     GT
22      11005678        .       G       A       0|1
22      16052167        .       A       AAAAC   0|1

I went through the paper but did not find a clear idea - could you please let me know the steps to predict the pathogenicity of such variants?

Thanks, DK

[finswimmer]

Hello DK ,

is this more a general question how to work with the ACMG guidelines or are you hopping that there is a computational solution for it? The later one will be quite hard to implement as it requires literature research.

If it's more a general question: Go through the criteria list in table 3 and 4 and classify by the rule sin table 5.

fin swimmer

[DanielC]

Thanks Fin! I am looking for computational approach to predict the pathogenicity of the variants in the vcf file using the ACMG guidelines. I am finding it hard to start, like what is the first step to implement? I know that vcf files contain SNP information and I should be looking at heterozygous alternate(0/1) and homozygous alternate(1/1), but then what are the steps to detect the pathogenicity? The table 3 and 4 give information on the criterion to name the variants either strong or very strong etc, but how to get to that step? I hope I put the question clearly, please let me know if I am not clear.

Thanks!

[finswimmer]

The first thing you should and can do is, is to annotate your vcf file file with

• population data like ExAC, gnomAD
• if available: desease database information like clniVar
• effect of the variant (nonsense, missense, ...)
• In Silico Predictive Algorithms

The most easiest way to do this is VEP by ensembl. Also snpEff/snpSift is very good.

Bu this is only the beginning in classifying by ACMG. For example functional data is missing.

fin swimmer

[Eugene A]

I'd like to refresh the discussion here (instead of making a new post) and ask for some advice:

The clinVar database has a field corresponding to ACMG (if I understand this correctly), for example mutation 140749365 in BRAF has a tag "Pathogenic" in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar?term=((174176[AlleleID])OR(29020[AlleleID])) ). I assume that this info could be integrated in VEP output. Nevertheless, I'm getting the empty CLIN_SIG field in vep output, even using flag "--everything" (as well as flag --check_existing). I'm not sure what is the problem here and actually out of good guesses :(

Also other output fields of VEP are quite cryptic: for example PHENO field has a value of "1&1,A" which suppose to be linked to "Existing_variation" field ("CM092083&COSV56058494"), but I'm not sure how to interpret it (seems that it is key -> value code, but I do not know where it is described)

Answer 1

[kuan-lin.huang]

Please check out our implementation of ACMG guideline: https://github.com/ding-lab/CharGer

Best, Kuan Huang Lab | Computational Omics @ MSSM

[brunobsouzaa]

Documentation is very sparse... A tutorial would be nice!