Testing: Ensure consistency between the output of the Variant-Linker tool and the VEP online version by testing various variants.

berntpopp commented 2 weeks ago

Current Behavior

There may be discrepancies between the annotations generated by the Variant-Linker tool and the VEP online version.

Proposed Solution

Select a set of representative variants to test.
Compare the annotations generated by the Variant-Linker tool with those from the VEP online version.
Identify any discrepancies and investigate their causes.
Adjust the tool's implementation if necessary to ensure consistency.

Acceptance Criteria

[ ] Select a diverse set of variants for testing.
[ ] Compare the Variant-Linker output with the VEP online version for these variants.
[ ] Document any discrepancies and their potential causes.
[ ] Implement necessary changes to ensure consistency between the tool and the VEP online version.
[ ] Comprehensive tests are added to maintain consistency in future updates.

berntpopp commented 2 weeks ago

also relates to #5

berntpopp commented 2 weeks ago

@ninarank please select ~20 representative variants from the ClinVar training set (all types from frameshifts, splcie, missense, deeper intronic) and run them through the online VEP version.

Post here the input and output files:

test_variants_input_2024-06-20.txt test_variants_vcf_format_2024-06-20.txt VEP_online_output_test_variants_2024-06-20.txt VEP_online_settings_2024-06-20

(VEP online tool sorts variants by chromosome, position, etc. The original input was not sorted.)

relevant for Nephro Variant Score are the following columns: gnomADg_AF gnomADe_AF Consequence CADD_PHRED IMPACT

berntpopp / variant-linker