Progress and Research Trends of Artificial Intelligence Incorporation in Protein Structure Prediction

Protein synthesis is closely regulated in nature, and in the field of protein design, scientists have attempted to understand and replicate those processes due to the importance of proteins. Existing protein structure prediction methods include physics-based approaches like molecular dynamics or knowledge-based methods such as energy functions, but these methods are computationally costly and time-consuming. Artificial Intelligence (AI) technology has been applied to protein structure prediction due to its ability to address current challenges more effectively. This paper will review the significant progress made recently in protein structure prediction using both classical machine learning and deep learning methods. Advancements in protein structure prediction using computational methods have been made possible due to the increase of available data, and representative examples of six databases used in protein structure prediction will be provided in this review. These public databases have diverse entries such as protein sequences, protein structures, protein interactions, enzyme nomenclature, and proteomes, which will facilitate more research in the protein design field. Despite the advancements with the incorporation of AI, several challenges remain, and the proposed solutions and research trends will be discussed in this paper.