AlphaFind was published in an article

Prediction of protein structure from the sequence is a long-standing problem. The AlphaFold machine learning algorithm recently provided structure for more than 200 million proteins that would otherwise be available only as a sequence. All predicted structures are stored in the AlphaFold DB database.

But how can you effectively search for a relevant structure (or a set of relevant structures) in such a vast database? One way is to query the database with an input structure and a similarity search algorithm. This approach is implemented in the AlphaFind web-based search engine, which uses a machine learning model to search for proteins with a tertiary structure similar to the one the user provides as the search query.

AlphaFind was recently published in the Web Server issue of the Nucleic Acids Research journal. Its article is titled AlphaFind: discover structure similarity across the proteome in AlphaFold DB. The software tool and the article result from our cooperation with the Faculty of Informatics and the Institute of Computer Science, both at Masaryk University.