Machine learning-based model for accurate identification of druggable proteins using light extreme gradient boosting

Area of Science:

• Computational biology
• Drug discovery
• Bioinformatics

Background:

• Identifying druggable proteins (DPs) is vital for advancing drug development, personalized medicine, and understanding disease mechanisms.
• Machine learning approaches offer a more efficient and cost-effective alternative to traditional methods for DP identification.

Purpose of the Study:

• To introduce Drug-LXGB, a novel computational predictor designed to enhance the identification of druggable proteins.
• To explore and integrate various feature descriptors for improved DP prediction accuracy.

Main Methods:

• Utilized feature descriptors including composition, transition, and distribution (CTD), composition of K-spaced amino acid pair (CKSAAP), pseudo-position-specific scoring matrix (PsePSSM), and multi-block pseudo amino acid composition (MB-PseAAC).
• Employed sequential forward selection for feature selection and evaluated machine learning models such as random forest, extreme gradient boosting, and light eXtreme gradient boosting (LXGB).
• Performance was assessed using 10-fold cross-validation.

Main Results:

• The integrated feature set, selected via sequential forward selection, provided robust characteristics for prediction.
• The light eXtreme gradient boosting (LXGB)-based model demonstrated superior predictive performance compared to other existing predictors.
• The Drug-LXGB model achieved the highest accuracy in identifying druggable proteins.

Conclusions:

• The developed Drug-LXGB protocol significantly enhances the identification of druggable proteins, aiding in novel drug design.
• This computational approach offers a more universal view of potential drug targets, accelerating therapeutic development.
• The findings are expected to be fruitful in exploring potential drug targets and advancing personalized medicine.