Prediction of MoRFs based on sequence properties and convolutional neural networks

Area of Science:

• Protein bioinformatics
• Computational biology
• Structural bioinformatics

Background:

• Intrinsically disordered proteins (IDPs) are crucial for biological functions due to their flexible structures.
• Molecular Recognition Features (MoRFs) are key functional regions within IDPs that transition from disordered to ordered states upon partner binding.

Purpose of the Study:

• To develop a novel computational method for predicting MoRFs using only protein sequence information.
• To evaluate the performance of the proposed method against existing MoRF prediction techniques.

Main Methods:

• A convolutional neural network (CNN) model named MoRF_CNN was developed.
• Sequence properties, including structural and physicochemical features, were utilized.
• A window-based approach was employed to preprocess sequence properties, emphasizing residue correlations.

Main Results:

• MoRF_CNN effectively predicts MoRFs by analyzing protein sequence properties.
• The method demonstrates superior performance compared to other existing prediction tools.
• The CNN architecture successfully integrates preprocessed sequence features for accurate MoRF identification.

Conclusions:

• MoRF_CNN provides an effective and competitive method for MoRF prediction solely based on protein sequence data.
• The approach does not rely on evolutionary information, simplifying its application.
• The study validates the efficacy of CNNs in identifying functional regions within intrinsically disordered proteins.