Provide the command used or report the bug here
I noticed a potential issue in scripts/update_intervar.R related to PM5 assignment.
PM5 should be based on a different pathogenic/likely pathogenic missense variant at the same amino-acid residue. However, the current code appears to require the ClinVar comparator to be an SNV, but does not explicitly check whether it is missense.
For TP53 NM_000546.6:c.1015G>A (p.Glu339Lys), the original InterVar result was VUS with PM5=0, but after the AutoGVP update it became Likely pathogenic with PM5=1.
At the same residue, ClinVar contains p.Glu339Ter (Variation ID 438708), which is Pathogenic but is a nonsense variant. Another missense variant, p.Glu339Gln (Variation ID 141893), is classified as Likely benign.
Therefore, p.Glu339Ter may be incorrectly used as the comparator for PM5 because the current implementation compares the amino-acid position without explicitly excluding nonsense variants.
Would it be appropriate to require the ClinVar comparator used for PM5 to be a missense variant?
It may also be useful to retain the ClinVar Variation ID(s) supporting PM5 in the output for auditability.
What version are you using?
AutoGVP v.2.0.1
Add error message here (if applicable)
N/A
Add Session info
Run sessionInfo() and post the output below
N/A
Provide the command used or report the bug here
I noticed a potential issue in scripts/update_intervar.R related to PM5 assignment.
PM5 should be based on a different pathogenic/likely pathogenic missense variant at the same amino-acid residue. However, the current code appears to require the ClinVar comparator to be an SNV, but does not explicitly check whether it is missense.
For TP53 NM_000546.6:c.1015G>A (p.Glu339Lys), the original InterVar result was VUS with PM5=0, but after the AutoGVP update it became Likely pathogenic with PM5=1.
At the same residue, ClinVar contains p.Glu339Ter (Variation ID 438708), which is Pathogenic but is a nonsense variant. Another missense variant, p.Glu339Gln (Variation ID 141893), is classified as Likely benign.
Therefore, p.Glu339Ter may be incorrectly used as the comparator for PM5 because the current implementation compares the amino-acid position without explicitly excluding nonsense variants.
Would it be appropriate to require the ClinVar comparator used for PM5 to be a missense variant?
It may also be useful to retain the ClinVar Variation ID(s) supporting PM5 in the output for auditability.
What version are you using?
AutoGVP v.2.0.1
Add error message here (if applicable)
N/A
Add Session info
Run
sessionInfo()and post the output belowN/A