Description

This track is part of the MaveMD collection. It shows one item per measured variant per score set, so a variant measured in several experiments appears once for each of them.

Each item carries the assay score, the functional class that score falls into, and, where the score set has been calibrated against variants of known clinical significance, the ACMG/AMP functional evidence code that follows: PS3 for evidence of a damaging effect, BS3 for evidence of a normal one, at a strength set by the calibration's odds of pathogenicity following the ClinGen recommendations for the PS3/BS3 criterion (Brnich et al.). ClinVar significance, gnomAD allele frequency and the ClinGen allele ID are attached where they exist.

Each item also carries the assay metadata: what the assay measured, in what model system, the molecular mechanism assessed, and how the variant library was built. That last field matters clinically. An assay built on an in vitro construct library introduces a synthetic copy of the target sequence, so it cannot detect an effect on splicing or on nonsense-mediated decay and can read falsely normal for such a variant. An assay that edits the endogenous locus can.

Two cautions when comparing variants. Many of the evidence codes come from calibrations MaveDB marks research use only; each item names its calibration and flags this. And the functional score has no common scale or direction between score sets, so a higher score does not always mean a more normal protein. Compare by functional class or evidence code, which mean the same thing everywhere.

Display Conventions and Configuration

Items are colored by the ACMG functional evidence code from the calibration named in the item details. Red shades carry evidence toward pathogenic, blue toward benign, darkening with the strength of the evidence.

Score sets that classify their variants without assigning an ACMG code are a measurement rather than a clinical claim, and use a separate palette:

Filters are available on the ACMG evidence code, the functional class, the gene, the ClinVar significance and the assay metadata, and none are applied by default. Because every substitution at a codon is a separate item, a well studied gene stacks deeply and the track falls back to a density graph in all but narrow windows; filtering to one gene or one evidence code makes individual variants readable. Item names, ClinGen allele IDs and MaveDB variant identifiers are searchable from the position box.

Methods

MaveMD curates MaveDB score sets for clinical relevance, integrates them with ClinVar and the ClinGen Allele Registry, and exports evidence structured for ACMG/AMP variant classification (McEwen et al.). Candidates were drawn from MaveDB, from publications cited in ClinVar for functional evidence, and from community recommendation, then restricted to genes with a moderate or stronger gene-disease association in ClinGen or GenCC.

Data are downloaded from the MaveDB API, one request per score set. MaveDB resolves some score sets to the genome and others only to a protein sequence, so a variant is placed from its genomic term where one exists, otherwise by projecting its protein term onto the corresponding codon through RefSeq or GENCODE, otherwise by converting the submitter's transcript term. Haplotypes, several substitutions measured as a single unit, have no one position and are excluded.

The build scripts are in the kent source tree, the steps that run them and the counts for each build in the makeDoc, and the track configuration in human/hg38/mavemd.ra.

Data Access

The data can be explored interactively in table format with the Table Browser or the Data Integrator and exported from there to spreadsheet or tab-sep tables. From scripts, the data can be accessed through our API, track=mavemdVar.

For automated download and analysis, the genome annotation is stored in a bigBed file that can be downloaded from our download server. The file for this track is called mavemdVar.bb. Individual regions or the whole genome annotation can be obtained using our tool bigBedToBed, which can be compiled from the source code or downloaded as a precompiled binary for your system. Instructions for downloading source code and binaries can be found here. The tool can also be used to obtain features within a given range, e.g. bigBedToBed http://hgdownload.soe.ucsc.edu/gbdb/$db/mavemd/mavemdVar.bb -chrom=chr17 -start=7668402 -end=7687550 stdout

The original annotation source data can be downloaded from the MaveDB API at https://api.mavedb.org.

Credits

Thanks to Alan Rubin, Benjamin Capodanno and Jeremy Stone of the MaveDB team for curating the MaveMD collection and for advice on retrieving it, and to the many groups whose experiments make up the underlying score sets.

References

McEwen AE, Stone J, Tejura M, Gupta P, Capodanno BJ, Da EY, Grindstaff SB, Moore N, Reinhart D, Snyder AE et al. MaveMD: A functional data resource for genomic medicine. medRxiv. 2025 Nov 19;. PMID: 41332838; PMC: PMC12668102

Rubin AF, Stone J, Bianchi AH, Capodanno BJ, Da EY, Dias M, Esposito D, Frazer J, Fu Y, Grindstaff SB et al. MaveDB 2024: a curated community database with over seven million variant effects from multiplexed functional assays. Genome Biol. 2025 Jan 21;26(1):13. PMID: 39838450; PMC: PMC11753097

Arbesfeld JA, Da EY, Stevenson JS, Kuzma K, Paul A, Farris T, Capodanno BJ, Grindstaff SB, Riehle K, Saraiva-Agostinho N et al. Mapping MAVE data for use in human genomics applications. Genome Biol. 2025 Jun 25;26(1):179. PMID: 40563119; PMC: PMC12188674

Brnich SE, Abou Tayoun AN, Couch FJ, Cutting GR, Greenblatt MS, Heinen CD, Kanavy DM, Luo X, McNulty SM, Starita LM et al. Recommendations for application of the functional evidence PS3/BS3 criterion using the ACMG/AMP sequence variant interpretation framework. Genome Med. 2019 Dec 31;12(1):3. PMID: 31892348; PMC: PMC6938631