ClinVar ClinSig updates and resolving submissions ties (4/N)

[pj-sullivan]

Purpose/implementation Section

What feature is being added or bug is being addressed?

Separate PR since there are a lot of edits to scripts/resolve-clinvar-submissions.R
Updated the README since it was out of date.

What was your approach?

• Expand recognized ClinVar significance categories — values now include low-penetrance P/LP, "Established/Likely risk allele" (Reported as "Risk allele"), "Uncertain risk allele" (reported as "Uncertain significance), and VUS-high/mid/low.
• Refactor the consensus/tie-resolution logic (both with and without a concept-ID list). "Latest" and "Most severe" conflict resolution is now performed only for variants from the categories in which there is a tie (e.g., if there are 3 P/LP, 3 VUS, and 1 B/LB, only the P/LP and VUS are considered).
• Output filename is now resolved-clinvar-{date}{-concept-conflict_res}.tsv.
• Validate input file dates: --variant_summary and --submission_summary filenames must reference the same date, or the script errors out.
• Final output filtering: drop rows with NA or non-actionable ClinicalSignificance (-, not provided, association, risk factor, drug response); dedupe by VariationID (keeping the latest evaluated) instead of vcf_id.

What GitHub issue does your pull request address?

Closes #283
Closes #282

Directions for reviewers. Tell potential reviewers what kind of feedback you are soliciting.

Download ClinVar files (I don't feel like this needs to be a data release?)

curl https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/archive/variant_summary_2026-06.txt.gz -o data/variant_summary_2026-06.txt.gz
curl https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/archive/submission_summary_2026-06.txt.gz -o data/submission_summary_2026-06.txt.gz

If you want to verify files:

md5sum data/*2026-06*
dcf6a458aa8fc850542a79ec9565271a  data/submission_summary_2026-06.txt.gz
53c1f539c36b0e7b5314e56788d963ab  data/variant_summary_2026-06.txt.gz

Run resolve clinvar using the cancer concept IDs (file from January is fine, concept IDs don't update often).

Rscript scripts/resolve-clinvar-intepretations.R --variant_summary data/variant_summary_2026-06.txt.gz --submission_summary data/submission_summary_2026-06.txt.gz --outdir refs --conceptID_list refs/clinvar_cancer_concept_ids_20260130.txt --conflict_res "latest"

Which areas should receive a particularly close look?

The output of the scripts/resolve-clinvar-intepretations.R and md5sum:

                           Benign              Benign/Likely benign 
                           209038                             93732 
                    Likely benign                 Likely pathogenic 
                          1098533                            112704 
Likely pathogenic, low penetrance                        Pathogenic 
                               13                            165698 
       Pathogenic, low penetrance      Pathogenic/Likely pathogenic 
                                1                             45791 
                      Risk allele            Uncertain significance 
                               24                           2353012 
                         VUS-high                           VUS-low 
                                1                                 1 

5c4d9e9a04ff05062ab661151369dce8  refs/resolved-clinvar-2026-06-cancer-latest.tsv

Is there anything that you want to discuss further?

Documentation Checklist

• The function has examples to showcase the usage
• Added a vignette

[rjcorb]

I ran resolve-clinvar-interpretations.R and got the same results. Updated code logic also looks good to me.

I've also updated the unadjusted InterVar call substring extraction to close #286.