AlphaMissense

Help

Search page

Identifier: For searching proteins you can use various identifiers including UniProt ACC, Entry name, Gene name, and Ensembl transcript ID. You can enter one identifier in the search box.

Filters: You can limit your search based on the predicted pathogenicity. You can select single or multiple labels (likely benign, likely pathogenic, and ambiguous, as defined in the original AlphaMissense paper). You can use high-pass and low-pass filters to select predictions with very high or very low scores.

Results page

If you do not get any hits, you can use the browser's back button to go back to the Search page and check the query.

Result tables: The table contains the variant name in p.AA1_POSITION_AA2 format (according to HGVS nomenclature), the predicted AlphaMissense score and class. The last column is colored according to the pathogenicity score.

Download: You can download the results by clicking the 'Download' button at the top. This results in a file named AlphaMissenseSearch-{UNIPROT_ACC}.zip which contains AlphaMissenseSearch-{UNIPROT_ACC}.tsv (tab separated values) file. This file can be imported into Excel or processed in other ways. This file contains the following columns: Uniprot ACC, Entry name, Gene name, Ensemble id, protein variant, a.a.1, position, a.a.2, pathogenicity score, and pathogenicity class.

Hotspot page

Protein: The structure and the pathogenicity table for CFTR are displayed by default. You can get another protein with one of the supported identifiers.

Structure: The AlphaFold predicted structure of the portein is displayed using the PDBe-Mol* viewer based on Mol* . The loading and rendering may be slow depending on your computer. You can rotate, click, highlight, etc.

The pathogenicity scores are placed in the occupancy and also in the B-factor columns of the pdb file. Therefore, if you hoover a residue then the occupancy value in the right bottom corner of the labels is the pathogenicity score.

Result table: The first row contains the WT residue and its position, color coded as:

blue, likely bening, 0-0.34
no color, ambiguous, 0.34-0.564
light red, liekly pathogenic, 0.564-0.78
dark red, likely pathogenic, 0.78-1.0

You can click the residues in this column to highlight the residue in the structure.
The columns #2-#4 contain the number of possible single nucleotide substitutions and the corresponding amino acids (e.g. 5:K,R,L,P,E). The last column displays the mean of AlphaMissense pathogenicity scores for each position.

Structures, PyMOL plugin

You can color the downloaded structures according to pathogenicity score based on the B-factor column in PyMOL. You can also download individual or all structures - see details at "Download".

You can also download a simple PyMOL plugin from here to make the coloring easy: coloram.py Install as a plugin or run as run https://alphamissense.hegelab.org/coloram.py. Finally issue the coloram command for coloring the structure.

Contact

If you have some question, feedback, or request: contact us.