Recent advances and future trends for protein-small molecule interaction predictions with protein language models

Area of Science:

• Computational biology
• Bioinformatics
• Artificial intelligence in life sciences

Background:

• Proteins perform critical biological functions via interactions with small molecules, such as enzymes, receptors, and transporters.
• Understanding these protein-small molecule interactions is vital for drug discovery, bioengineering, and cellular metabolism research.
• Protein language models (PLMs) are emerging AI tools analyzing protein sequences to uncover hidden biological patterns.

Purpose of the Study:

• To review state-of-the-art PLMs for analyzing protein sequences.
• To explore the integration of PLMs with small molecule data for interaction prediction.
• To discuss current prediction tasks, limitations, and future directions in this field.

Main Methods:

• Review of current literature on protein language models (PLMs).
• Analysis of methods integrating PLMs with small molecule datasets.
• Examination of prediction tasks for protein-small molecule interactions using PLMs.

Main Results:

• PLMs show significant potential in identifying patterns related to protein structure, function, and stability from amino acid sequences.
• Integration of PLMs with small molecule information enables prediction of protein-small molecule interactions.
• Several prediction tasks and their associated challenges have been identified.

Conclusions:

• PLMs offer a powerful approach to understanding protein behavior and interactions.
• Further development is needed to overcome limitations and enhance the predictive accuracy of PLMs for protein-small molecule interactions.
• This approach holds promise for advancing drug design, bioengineering, and metabolic studies.