TransFun: A Tool of Integrating Large Language Models, Transformers, and Equivariant Graph Neural Networks to Predict Protein Function.

Authors

Frimpong Boadu, Ahhyun Lee, Jianlin Cheng

Published in

Methods in molecular biology (Clifton, N.J.). Volume 2941. Pages 101-111.

Abstract

Experimentally determining the functions of proteins is a complex and time-consuming process. This challenge contributes to a gap, where many proteins have known sequences, predicted structures, and other crucial information, yet lack functional annotations. This gap underscores the critical importance of automated function prediction (AFP) methods, which aim to develop computational techniques dedicated to predicting protein functions. Most AFP methods leverage the wealth of diverse protein information available, such as sequences, structures, protein-protein interactions, and domain characteristics. These methods often utilize individual features or integrate multiple features to enhance the accuracy of function prediction. In this chapter, we focus on TransFun, a structure-based protein function prediction technique. TransFun leverages the embeddings provided by the ESM-1b pretrained protein language models to distill intricate sequence features and combines them with AlphaFold's predicted structures to predict protein functions. Availability: https://github.com/jianlin-cheng/TransFun.

PMID:
40601253
Bibliographic data and abstract were imported from PubMed on 02 Jul 2025.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.