TargetMM: Accurate Missense Mutation Prediction by Utilizing Local and Global Sequence Information with Classifier...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Missense mutations (MM) can cause human diseases by altering protein function.
• Accurate prediction of MM is vital for protein annotation and drug design.
• Existing computational methods for MM prediction have limitations, necessitating improved approaches.

Purpose of the Study:

• To develop a novel computational method for accurate prediction of missense mutations.
• To enhance the identification of pathogenic mutations from neutral ones.

Main Methods:

• A new feature extraction method was designed, considering the impact of residues within the microenvironment of a mutation site.
• Three heterogeneous prediction models were trained and ensembled for final prediction.
• The efficacy of the method was evaluated using stringent cross-validation and independent testing on benchmark datasets.

Main Results:

• The developed predictor, TargetMM, demonstrated superior performance compared to existing advanced methods on independent test data.
• Statistical evaluation confirmed the effectiveness of the proposed feature representation and classifier ensemble techniques.

Conclusions:

• TargetMM offers an improved approach for identifying pathogenic missense mutations.
• The developed tool and datasets are available for academic research, facilitating further advancements in the field.