Prediction of protein stability changes upon single-point variant using 3D structure profile.

Area of Science:

• Biophysics
• Computational Biology
• Genomics

Background:

• Accurate prediction of protein thermodynamic stability changes from single-point variants is critical for understanding mutation impacts in biophysics, genomics, and disease.
• Existing computational methods often require tertiary protein structures, limiting their application due to structural data availability.
• Current prediction accuracy for single-point variants is insufficient for practical use.

Purpose of the Study:

• To review existing computational approaches for predicting single-point variant effects on protein stability.
• To introduce MU3DSP, a novel computational framework for predicting protein thermodynamic stability changes from sequence data without experimental structures.
• To enhance the accuracy and applicability of computational predictions for protein variants.

Main Methods:

• A comprehensive review of current computational tools for predicting single-point variant stability changes.
• Development of MU3DSP, a framework utilizing point-mutated protein 3D structure profiles.
• Integration of sequence-level features and averaged 3D structure features derived from PDB alignments within MU3DSP.

Main Results:

• MU3DSP demonstrates superior performance compared to existing methods on various benchmarks.
• The framework effectively assesses thermodynamic stability changes induced by single-point substitutions.
• MU3DSP provides reliable predictions using only protein sequence as input.

Conclusions:

• MU3DSP offers a significant advancement in predicting protein thermodynamic stability changes from single-point variants.
• This tool can reliably assess both somatic and germline substitution variants.
• MU3DSP serves as a valuable asset for protein design and the study of mutation-related diseases.