MetaDome 2027: ClinVar variants of uncertain significance at homologous Pfam domain positions

MetaDome 2027: ClinVar variants of uncertain significance at homologous Pfam domain positions Missense variants of uncertain significance (VUS) from ClinVar that sit at Pfam domain positions where pathogenic or likely pathogenic missense variation is observed at the evolutionarily equivalent position in a homologous domain elsewhere in the genome. There is one table per genome build, with one row per variant per domain placement. These are the candidate sets behind the meta-domain evidence analysis in the MetaDome 2027 paper, and they are the paper's Supplementary Data S2 and S3. The copies attached to the paper hold the same tables in zip archives. Everything here is derived from the published MetaDome 2027 data release and a single ClinVar VCF per build, using only bcftools, bedtools and awk, so the whole result can be rebuilt by anyone holding those two inputs. Files File Content metadome2027_clinvar-vus-metadomain_GRCh38.p14_clinvar-20251006_gencode-v45_uniprot-2025-01_pfam-37.4.tsv.gz Candidate set, GRCh38.p14 metadome2027_clinvar-vus-metadomain_GRCh37.p13_clinvar-20251006_gencode-v19_uniprot-2025-01_pfam-37.4.tsv.gz Candidate set, GRCh37.p13 metadome2027_clinvar-vus-metadomain_GRCh38.p14_clinvar-20251006_gencode-v45_uniprot-2025-01_pfam-37.4_summary.tsv Summary counts, GRCh38.p14 metadome2027_clinvar-vus-metadomain_GRCh37.p13_clinvar-20251006_gencode-v19_uniprot-2025-01_pfam-37.4_summary.tsv Summary counts, GRCh37.p13 Provenance Input GRCh38.p14 GRCh37.p13 Genome build GRCh38.p14 GRCh37.p13 GENCODE v45 v19 UniProtKB/Swiss-Prot 2025_01 2025_01 Pfam 37.4 37.4 ClinVar clinvar_20251006.vcf.gz (GRCh38 release), md5 b3737280b4b5a9e4a3e187ece89a7581 clinvar_20251006.vcf.gz (GRCh37 release), md5 fa7cc30e4ffaf85682e8e7a4fc64be12 MANE Select v1.2, as carried by the GENCODE v45 annotation Not available in GENCODE v19 Analysis code scripts/variant_analysis/clinvar_analysis.sh at github.com/laurensvdwiel/metadome Same Counts GRCh38.p14 GRCh37.p13 Missense VUS in ClinVar 1,920,156 1,920,209 Distinct VUS inside a Pfam domain 786,926 765,455 Distinct VUS with any ClinVar missense record at their meta-domain position 286,505 271,746 Distinct VUS with homologous pathogenic evidence (the candidate set) 147,835 140,374 Rows, one per variant per placement 368,438 332,525 Class 1, same protein position 3,337 3,179 Class 2, homologous positions only 52,463 51,479 Class 3, conflicting 92,035 85,716 Restricted to MANE Select placements, GRCh38.p14 only: Count Distinct VUS inside a Pfam domain 759,840 Candidate set 142,726 Rows 208,416 Class 1 3,283 Class 2 50,709 Class 3 88,734 The _summary.tsv files carry these denominators as key and value pairs, including counts that the tables cannot express, since variants without homologous evidence are not written to the tables. Each candidate is counted once, in the strongest class across its placements: class 1 has pathogenic evidence at the same protein position, class 2 at homologous positions only, and class 3 at homologous positions together with benign variation. Columns # Column Type Description 1 chrom text Chromosome, chr1 to chrY. The chr prefix is added where the ClinVar VCF omits it, so it matches the release tracks and metadome_url. 2 pos integer Position on that chromosome, 1-based, as ClinVar reports it. 3 clinvar_id text ClinVar Variation ID, the ID field of the VCF. 4 ref text Reference allele. 5 alt text Alternate allele. 6 clnsig text ClinVar clinical significance, Uncertain_significance throughout. 7 clnrevstat text ClinVar review status. 8 clnhgvs text ClinVar genomic HGVS on the RefSeq chromosome accession. 9 gene_symbol text Gene symbol for the Swiss-Prot accession, taken from MetaDome rather than from ClinVar's GENEINFO. Accessions shared by several genes keep all symbols, joined with a vertical bar. Empty where the accession has no symbol in the release. 10 mane_select boolean True or False: whether this placement's accession is the one the gene's MANE Select transcript maps to. False on every GRCh37.p13 row, since MANE is not part of GENCODE v19. 11 uniprot_ac text Swiss-Prot accession. Isoform accessions carry a -N suffix, so one protein can appear under several accessions. 12 uniprot_pos integer Residue position in that accession, 1-based. 13 pfam_id text Pfam accession of the domain containing the residue. 14 consensus_pos integer Position in the Pfam HMM consensus, 1-based: the meta-domain position. 15 homolog_missense_P_count integer Distinct ClinVar Pathogenic missense records at this meta-domain position, excluding records at this variant's codon. 16 homolog_missense_LP_count integer The same for Likely_pathogenic. 17 clinvar_P_accessions text ClinVar Variation IDs behind column 15, comma separated without spaces. Empty where the count is zero. 18 clinvar_LP_accessions text The same for column 16. 19 homolog_benign_count integer Distinct ClinVar benign missense records at this meta-domain position, excluding records at this residue. 20 same_residue_PLP_count integer Distinct ClinVar pathogenic or likely pathogenic missense records at this residue in this accession. 21 metadome_url text Opens the MetaDome position-lookup view for the codon carrying the variant, from which the dashboard opens at the residue. The coordinate in the URL is a position within that codon and can differ from pos Missing values are the empty string, used only in columns 9, 17 and 18. Counts are never empty, and a count of zero is written as 0. Every row has 21 fields. The homologous Pathogenic and Likely pathogenic counts correspond to the homologous-domain ClinVar records in the dashboard's per-position information panel, which, like the file, exclude records at the variant's own codon. The homologous benign counts are derived from ClinVar by the analysis only; the platform shows population variation from gnomAD at homologous positions instead. Conventions Tab separated, UTF-8, LF line endings, one header row, no index column. Rows are sorted by chrom, then pos; the order among rows sharing a position is not meaningful. Positions in pos are 1-based, following ClinVar, whereas the release BED tracks these tables are derived from are 0-based half-open, as the format specifies. uniprot_pos and consensus_pos are 1-based. Licence Creative Commons Attribution 4.0 International (CC BY 4.0). ClinVar is in the public domain. GENCODE, UniProtKB/Swiss-Prot, Pfam and MANE carry their own terms, which apply to any redistribution of the underlying annotation. Contact Issues and questions: https://github.com/laurensvdwiel/metadome/issues