Automated use of mutagenesis data in structure prediction

Area of Science:

• Computational Biology
• Structural Bioinformatics
• Molecular Modeling

Background:

• Experimental structural determination is often challenging.
• Protein sequence modifications (mutagenesis) provide valuable data for understanding molecular structure and function.
• Current methods for incorporating mutagenesis data into modeling are often manual or indirect.

Purpose of the Study:

• To develop a computational approach for predicting molecular structures by directly incorporating mutation information into the fitness score.
• To enhance the accuracy of protein structure prediction, especially when using less precise force fields.
• To accelerate conformational search algorithms used in structure prediction.

Main Methods:

• Developed a method based on statistical lattice models to integrate mutation data into a fitness score.
• Accounted for mutation phenotypes (neutral or disruptive) and their impact on molecular stability.
• Calculated energy for a given structure over an ensemble of sequences, including wild type and mutated variants.
• Utilized three types of sequence ensembles: saturation mutagenesis, scanning mutagenesis, and homologous proteins.

Main Results:

• The approach successfully uses mutation information to improve the accuracy of structure prediction, even with suboptimal force fields.
• Incorporating multiple sequences into a statistical ensemble energetically distinguishes native states from misfolded structures.
• The ensemble energy calculation effectively speeds up conformational search algorithms like Monte Carlo methods.

Conclusions:

• Mutation data can be directly and effectively integrated into computational structure prediction workflows.
• This method enhances the reliability and efficiency of predicting protein structures.
• The approach offers a valuable tool for molecular modeling and understanding protein behavior.