Researchers at the National Centre for Biological Sciences, under the Tata Institute of Fundamental Research, Bengaluru, have developed a deep-learning tool that can predict how intrinsically disordered proteins bind to their partners. The breakthrough addresses a long-standing challenge in molecular biology and opens new possibilities in disease research and drug discovery.
Why Intrinsically Disordered Proteins Matter
Unlike most proteins that fold into stable three-dimensional structures, intrinsically disordered proteins (IDPs) lack a fixed shape. These shapeshifting molecules are central to cellular life. They regulate signalling networks, control gene expression, assist protein folding and quality control, and help form dynamic cellular structures known as condensates. Their flexibility, while biologically advantageous, has made them difficult to analyse using conventional structural biology techniques.
The Disobind Deep-Learning Tool
The newly developed tool, named Disobind, analyses protein sequences to predict which regions of a disordered protein will interact with a binding partner. It uses protein language models, a form of artificial intelligence trained on millions of known protein sequences. Crucially, Disobind does not require prior structural information or sequence alignments and explicitly accounts for the binding partner, a key factor in IDP interactions.
Performance and Benchmarking
The research team, led by Kartik Majila, benchmarked Disobind against existing predictors, including AlphaFold-Multimer and AlphaFold3. Disobind consistently showed higher accuracy, particularly when tested on previously unseen protein pairs. When used alongside AlphaFold-Multimer predictions, its performance improved further, highlighting its complementary value to structure-based approaches.
Imporatnt Facts for Exams
Intrinsically disordered proteins lack a stable 3D structure.
Protein language models are AI systems trained on large protein sequence datasets.
NCBS is a premier life sciences research institute under TIFR.
AlphaFold tools are used for protein structure and interaction prediction.
Applications in Disease and Drug Design
|
