AnnoPRO: a strategy for protein function annotation based on multi-scale protein representation and a hybrid deep...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics
• Proteomics

Background:

• Protein function annotation is a critical challenge in biological sciences.
• Existing computational methods struggle with the 'long-tail problem,' where many Gene Ontology (GO) families have limited annotated proteins.
• This limitation hinders comprehensive understanding of protein roles and biological pathways.

Purpose of the Study:

• To develop an innovative computational strategy, AnnoPRO, to overcome the limitations of current protein function annotation methods.
• To improve the accuracy and coverage of protein function prediction, particularly for underrepresented GO families.
• To provide a robust and efficient tool for large-scale protein annotation.

Main Methods:

• AnnoPRO employs sequence-based multi-scale protein representation to capture diverse sequence features.
• A dual-path protein encoding strategy utilizes pre-training for enhanced feature extraction.
• Function annotation is achieved through a long short-term memory (LSTM)-based decoding mechanism.

Main Results:

• Case studies and benchmark evaluations demonstrate AnnoPRO's superior performance compared to existing methods.
• The method effectively addresses the long-tail problem, improving annotation for sparse GO families.
• AnnoPRO shows significant advancements in predicting protein functions accurately.

Conclusions:

• AnnoPRO represents a significant advancement in computational protein function annotation.
• The developed strategy offers a powerful solution for the longstanding challenge of the long-tail problem in GO annotation.
• Freely available source code and models facilitate broader adoption and further research in the field.