ADPredict: ADP-ribosylation site prediction based on physicochemical and structural descriptors.

Area of Science:

• Biochemistry and Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• ADP-ribosylation is a critical post-translational modification (PTM) involved in DNA repair, chromatin structure, metabolism, and stress responses.
• A comprehensive understanding of ADP-ribosylation targets and modification sites across various tissues and diseases remains incomplete.
• Identifying specific ADP-ribosylation sites is crucial for elucidating the molecular mechanisms governed by this modification.

Purpose of the Study:

• To develop a computational tool for predicting ADP-ribosylated sites, specifically on aspartic and glutamic acids.
• To provide researchers with a tool to advance the study of ADP-ribosylation in biological systems.

Main Methods:

• Developed ADPredict, a predictive algorithm utilizing physicochemical properties, novel secondary structure descriptors (HM-ratio), and 3D protein features.
• Employed principal component analysis and machine learning techniques for model development.
• Validated performance through internal bootstrapping and prediction on external experimental datasets.

Main Results:

• ADPredict demonstrates superior performance compared to existing tools like ModPred for predicting ADP-ribosylated sites.
• The novel HM-ratio descriptor significantly contributed to the predictive model's accuracy.
• The tool successfully predicts ADP-ribosylated aspartic and glutamic acids.

Conclusions:

• ADPredict is the first dedicated computational tool for predicting ADP-ribosylated aspartic and glutamic acids.
• The tool represents a valuable asset for bioinformatics research, particularly in PTM prediction.
• ADPredict is freely accessible, facilitating broader research in ADP-ribosylation.