Protein structure-based evaluation of missense variants: Resources, challenges and future directions

Area of Science:

• Genomics and Bioinformatics
• Structural Biology
• Computational Biology

Background:

• Missense variants are crucial in genetic diseases.
• Accurate interpretation of missense variants is challenging.
• Protein structure provides valuable insights into variant effects.

Purpose of the Study:

• To provide an overview of protein structure-based methods for missense variant evaluation.
• To categorize algorithms based on their prediction approach (e.g., ΔΔG calculation vs. feature-based prediction).
• To discuss the impact of recent deep learning advancements on variant interpretation.

Main Methods:

• Review of existing algorithms for missense variant effect prediction.
• Categorization of methods into free energy difference (ΔΔG) calculators and structural feature predictors.
• Discussion of machine learning and deep learning approaches used in these algorithms.

Main Results:

• Algorithms can be broadly divided into those calculating ΔΔG and those using structural features.
• Machine learning is widely employed in developing these predictive algorithms.
• Recent deep learning breakthroughs offer new opportunities for variant interpretation.

Conclusions:

• Protein structure-based methods are essential for missense variant evaluation.
• Machine learning and deep learning are transforming the field of variant interpretation.
• Further development and application of these methods will improve understanding of genetic variants.