Intelligent Analysis of Biological Sequences for Protein Function Classification

Abstract

Protein function classification represents an important research area in bioinformatics. With the rapid growth of protein sequence data, there is an increasing need to develop effcient classification systems based on Artificial Intelligence techniques. This work aims to classify protein sequences functionally using the Protein Data Bank database, employing Machine Learning and Deep Learning models, in addition to hybrid and ensemble models, based on different protein sequence representation methods. Furthermore, this study leverages protein language models such as ProtBERT and Evolutionary Scale Modeling to extract rich representations of protein sequences. The models are evaluated using performance metrics such as Accuracy, Precision, Recall, and F1-score, as well as the assessment of model generalization ability. The results demonstrate the effectiveness of the proposed approaches in improving protein function classification performance, particularly the ESM model, which achieved the best performance with an accuracy of 94.95%. This study contributes to the development of protein function classification tools by providing accurate and effective models based on advanced sequence representations, thereby improving classification accuracy and quality