SLAM: Structure-aware lysine β-hydroxybutyrylation prediction with protein language model

Area of Science:

• Biochemistry
• Proteomics
• Computational Biology

Background:

• Post-translational modifications (PTMs) are crucial for protein function diversity.
• Lysine β-hydroxybutyrylation (Kbhb) is a newly identified PTM with significant implications for functional proteomics.
• Experimental methods for PTM site identification are often costly and time-consuming.

Purpose of the Study:

• To develop the first computational method for predicting lysine β-hydroxybutyrylation (Kbhb) sites.
• To provide an accurate and efficient tool for identifying novel Kbhb sites.
• To facilitate the study of protein function and stability through PTM analysis.

Main Methods:

• Development of a deep learning-based method named SLAM.
• In silico prediction of lysine β-hydroxybutyrylation sites.
• Performance evaluation using 5-fold cross-validation and independent test sets.

Main Results:

• SLAM achieved high AUROC values (0.890-0.923) on general and species-specific independent test sets.
• The method successfully predicted known Kbhb sites in human S-adenosyl-L-homocysteine hydrolase.
• Demonstrated accuracy and efficiency in identifying potential Kbhb sites.

Conclusions:

• SLAM is the first deep learning tool for in silico prediction of lysine β-hydroxybutyrylation sites.
• The method enables accurate and efficient characterization of novel Kbhb sites.
• SLAM can aid in understanding protein function and stability and guide the identification of other PTM sites.