StackDPPred: a stacking based prediction of DNA-binding protein from sequence.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Identifying DNA-binding proteins solely from sequence information presents a significant challenge in genome annotation.
• DNA-binding proteins are vital for fundamental biological processes including DNA replication, repair, transcription, and splicing.
• Current experimental methods for identifying DNA-binding proteins are costly and time-consuming, necessitating efficient computational approaches.

Purpose of the Study:

• To develop an effective computational method for predicting DNA-binding proteins using only sequence information.
• To improve the accuracy of DNA-binding protein prediction beyond existing methods that rely solely on Position-Specific Scoring Matrix (PSSM) profiles.
• To provide a tool that can accelerate genome annotation and guide experimental validation.

Main Methods:

• Proposed StackDPPred, a stacking-based machine learning method.
• Utilized features extracted from PSSM profiles and residue-specific contact energy.
• Employed jackknife validation on a benchmark dataset of 1063 proteins (518 DNA-binding, 545 non DNA-binding).

Main Results:

• StackDPPred achieved high performance metrics: 89.96% accuracy (ACC), 0.799 Matthews Correlation Coefficient (MCC), and 94.50% Area Under the Curve (AUC).
• The method demonstrated superior performance compared to several state-of-the-art approaches.
• Consistent outperformance was observed on two independent test datasets, validating its robustness.

Conclusions:

• StackDPPred offers an effective computational solution for predicting DNA-binding proteins directly from their amino acid sequences.
• The integration of PSSM and contact energy features enhances prediction accuracy.
• The developed method serves as a valuable tool for rapid annotation and experimental guidance in genomics research.