Why do these two variants appear to result in the same amino acid change when annotated by snpeff, but are annotated differently in clinvar

Question

[curious]

I am running the following command for two vcfs, each containing one variant

 snpEff -Xmx64G  'hg38' my_vcf | SnpSift annotate -noId clinvar.vcf.gz

This is the result of vcf1

13  32354919    rs397507894 TTT T   .   .   AF=2.52525e-05;MAF=2.52525e-05;;ANN=T|frameshift_variant|HIGH|BRCA2|BRCA2|transcript|NM_000059.3|protein_coding|14/27|c.7067_7068delTT|p.Phe2356fs|7294/11386|7067/10257|2356/3418||;LOF=(BRCA2|BRCA2|1|1.00)

Here is the result of vcf2:

13  32354918    52262   ATT A   .   .   ANN=A|frameshift_variant|HIGH|BRCA2|BRCA2|transcript|NM_000059.3|protein_coding|14/27|c.7067_7068delTT|p.Phe2356fs|7294/11386|7067/10257|2356/3418||INFO_REALIGN_3_PRIME;LOF=(BRCA2|BRCA2|1|1.00);ALLELEID=66930;CLNDISDB=MONDO:MONDO:0012933,MedGen:C2675520,OMIM:612555,Orphanet:145|MONDO:MONDO:0016419,MedGen:C0346153,OMIM:114480,Orphanet:227535;CLNDN=Breast-ovarian_cancer,_familial,_susceptibility_to,_2|Familial_cancer_of_breast;CLNHGVS=NC_000013.11:g.32354920_32354921del;CLNREVSTAT=reviewed_by_expert_panel;CLNSIG=Pathogenic;CLNVC=Deletion;CLNVCSO=SO:0000159;CLNVI=ClinGen:CA024828;GENEINFO=BRCA2:675;MC=SO:0001589|frameshift_variant;ORIGIN=1;RS=397507894

It seems like both of these result in the same change NM_000059.3:c.7067_7068delTT and p.Phe2356fs, but one is annotated in Clinvar as pathogenic and the other one isnt? Why is this?

I can see the Clinvar VCF is like this:

13  32354918    52262   ATT A   .   .   ALLELEID=66930;CLNDISDB=MONDO:MONDO:0012933,MedGen:C2675520,OMIM:612555,Orphanet:145|MONDO:MONDO:0016419,MedGen:C0346153,OMIM:114480,Orphanet:227535;CLNDN=Breast-ovarian_cancer,_familial,_susceptibility_to,_2|Familial_cancer_of_breast;CLNHGVS=NC_000013.11:g.32354920_32354921del;CLNREVSTAT=reviewed_by_expert_panel;CLNSIG=Pathogenic;CLNVC=Deletion;CLNVCSO=SO:0000159;CLNVI=ClinGen:CA024828;GENEINFO=BRCA2:675;MC=SO:0001589|frameshift_variant;ORIGIN=1;RS=397507894

Is the issue that SnpSift is niave to the protein change and just adds the annotation to vcf2 because the chrom:pos:ref:alt aligns perfectly with Clinvar?

Answer 1

[Jeremy Leipzig]

I think you need to normalize (left-align your VCFs before annotating): https://genome.sph.umich.edu/wiki/Variant_Normalization

[curious]

yup, quick check with bcftools norm seems to make the vcfs the same