iThermo: A Sequence-Based Model for Identifying Thermophilic Proteins Using a Multi-Feature Fusion Strategy.

Area of Science:

• Biotechnology and protein engineering
• Computational biology and bioinformatics
• Biochemistry and molecular biology

Background:

• Thermophilic proteins are vital for industrial and biotechnological applications.
• Current biochemical identification methods are inefficient, costly, and time-consuming.
• Accurate identification of thermophilic proteins is essential for protein engineering.

Purpose of the Study:

• To develop a fast and accurate computational method for identifying thermophilic proteins.
• To establish a reliable benchmark dataset for thermophilic protein identification.
• To create a user-friendly software package for practical application.

Main Methods:

• Construction of a benchmark dataset with 1,368 thermophilic and 1,443 non-thermophilic proteins.
• Development of a multi-layer perceptron (MLP) model utilizing a multi-feature fusion strategy.
• Validation of the model's performance on an independent dataset.

Main Results:

• The proposed MLP model achieved a high accuracy of 96.26% in discriminating thermophilic proteins.
• A user-friendly software package, iThermo, was developed for convenient model utilization.
• The model demonstrates significant potential for accelerating the discovery and engineering of thermostable proteins.

Conclusions:

• The developed computational model offers a fast and accurate alternative to traditional biochemical identification methods.
• The iThermo software package enhances the practical applicability of the model in research and industry.
• This study facilitates advancements in the field of thermally stable protein discovery and engineering.