Background: AlphaFold2—awarded the 2024 Nobel Prize in Chemistry—has created the extraordinary opportunity of evaluating the impact of millions of missense variants directly in three-dimensional (3D) structural space rather than relying solely on sequence-based information. In 2019, we developed the Missense3D algorithm to predict the impact of amino-acid substitutions. Unlike many variant-effect predictors, including AlphaMissense, Missense3D provides explicit structural explanations rather than simply benign/damaging predictions, helping users understand the mechanistic basis of variant effects. The publication (doi: 10.1016/j.jmb.2019.04.009) has >500 citations and the webserver supports >8,000 unique users annually. Missense3D is free for academic and commercial users at https://missense3d.bc.ic.ac.uk/ Missense3D performs in-silico structural mutagenesis by introducing the amino-acid change into the wild-type structure and identifies structural disruptions, such as cavity formation, steric clashes, or the breakage of key chemical bonds, that can compromise protein stability. Key features of Missense3D:

1. Flexible input: variants can be submitted using genomic (HGVS, Ensembl annotations, dbSNP rsIDs), protein-level and structure-level coordinates.

2. Modelling variants on single-protein and protein–protein complex: analysis performed on experimental or predicted structures, including AlphaFold2.

3. Automated sequence–structure mapping: dynamic mapping of the query sequence to thousands of 3D structures, supporting non-experts in 3D-structure.

4. Structure quality control: built-in checks guide users towards reliable 3D coordinates for variant analysis. Missense3D is a user-friendly, interpretable framework for integrating protein structure insights into variant interpretation for any organism without requiring prior expertise in 3D structures.