AI Tool ‘Disobind’ to Decode Elusive Protein Interactions

Recently, researchers developed Disobind, an open-source deep-learning tool to predict interactions involving intrinsically disordered proteins (IDPs).

• Disobind focuses on intrinsically disordered regions (IDRs), flexible protein segments that lack a fixed structure but play crucial roles in cell signalling, gene regulation, cancer progression and neurodegenerative diseases.
• Unlike traditional structural biology tools, which struggle with shape-shifting proteins, Disobind uses protein language models trained on millions of protein sequences to identify which regions of a disordered protein bind to a specific partner.
• A key strength of Disobind is that it is partner-aware, analysing both interacting proteins, recognising that the binding behaviour of IDRs depends heavily ....