Finding antibodies in cryo-EM maps with CrAI

• DOI: 10.1093/bioinformatics/btaf157
• Publication Date: 2025-04-09T17:46:36Z
• Authors: Vincent Mallet; Chiara Rapisarda; Hervé Minoux; Maks Ovsjanikov

Abstract Motivation Therapeutic antibodies have emerged as a prominent class of new drugs due to their high specificity and ability bind several protein targets. Once an initial antibody has been identified, its design characteristics are refined using structural information, when it is available. Cryo-EM currently the most effective method obtain 3D structures. It relies on well-established methods process raw data into map, which may, however, be noisy contain artifacts. To fully interpret these maps number, position structure other proteins present must determined. Unfortunately, existing automated addressing this step limited accuracy, require additional inputs resolution maps, exhibit long running times. Results We propose first automatic efficient dedicated finding in cryo-EM maps: CrAI. This machine learning approach leverages conserved novel database that we built solve problem. Running prediction takes only few seconds, instead hours, requires nothing but seamlessly integrating within analysis pipelines. Our can find location pose both Fabs VHHs at resolutions up 10Å significantly more reliable than approaches. Availability Implementation make our available open source github.com/Sanofi-Public/crai ChimeraX bundle (crai). Supplementary information material Bioinformatics online.