Multiplexed Assays of Variant Effect Database

ChimeraX can fetch or open local files containing deep mutational scanning data from MaveDB, the Multiplexed Assays of Variant Effect Database, for visualization and analysis with the Mutation Scores tool and/or mutationscores command.

Data can be fetched using the MaveDB experiment set identifier, for example to open alpha-synuclein data entry with identifier 45, use ChimeraX command

  open 45 from mavedb

Comma-separated values files (.csv) containing mutational scan data can also be downloaded from MaveDB and opened in ChimeraX with the File / Open... menu or ChimeraX command

  open ~/Downloads/urn_mavedb_00001222-a-2_scores.csv

Opening a mutation scores data automatically shows the Mutation Scores tool allowing display of heatmaps, scatterplots, histograms, and coloring atomic structures.

Issues when viewing MaveDB data

The MaveDB database contains diverse data, only some of which can be visualized in ChimeraX. Here are some of the current limitations.

1. Protein coding sequences. MaveDB contains mutational scans for protein coding DNA sequences and also non-coding sequences such as enhancers or promoters. ChimeraX currently only reads protein variant HGVS notation. Opening entries for non-protein coding sequences will give an error message saying that data is not supported.
2. Nucleic acid variants. Some MaveDB entries for proteins specify the sequence variants only using DNA HGVS notation. This has advantages because specifying variants with protein HGVS notation cannot distinguish between alternate DNA codons that result in the same amino acid mutation. ChimeraX does not currently read the nucleic acid HGVS so cannot display these MaveDB entries. We hope to support this in the future. One difficulty is the MaveDB entries do not always include the DNA sequence, necessary for converting DNA variants to protein variants. Instead they provide a reference sequence identifier (e.g. NM_004006.3) and it will be necessary for ChimeraX to be able to query other databases to obtain those sequences.
3. Synonymous mutations. Synonymous mutations are non-wildtype DNA codons that code for the same wildtype amino acid. While these don't change the protein sequence they can effect protein expression levels and hence activity assay scores. Some MaveDB entries include protein and DNA HGVS notations but the protein descriptors don't include the synonymous mutations while the DNA descriptors do include them. ChimeraX expects to see protein HGVS that includes synonymous mutations (e.g. p.(Ala123Ala) or p.(Ala123=)). Some MaveDB entries contain protein variants p.= meaning no change to amino acid sequence, while the DNA variant will indicate the synonymous change. Since ChimeraX only currently reads the protein variant notation it does not know which residue has the synonymous mutation. Opening such data files may produce a warning, but also in cases produces no warning, for instance when the DNA variant indicates a synonymous and a non-synonymous mutant and the protein notation indicates only the non-synonymous mutation, ChimeraX will incorrectly conclude this is a single position mutant.
4. Long score names. MaveDB entries often have multiple assays that are given identifiers a-1, a-2, ..., b-1, b-2, .... Since these identifiers convey no information about what was assayed ChimeraX instead displays the database score titles to name the different assays. These titles can be long (e.g. "Deep Mutational Scanning of alpha-Synuclein based on Toxicity in Yeast Treated with Brefeldin A") which make them harder to see in the ChimeraX visualization control panels where limited screen space is available to show the names. Unfortunately MaveDB does not currently have short descriptive names for the different assays of an entry. We plan to allow renaming the scores in ChimeraX to work around this problem.