Deep Learning-Based Advances in Protein Structure Prediction.

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• Determining accurate protein structures is crucial for understanding biological mechanisms.
• Experimental methods for protein structure determination have advanced but cannot keep pace with sequence data.
• A gap exists between the number of known protein sequences and experimentally determined structures.

Purpose of the Study:

• To highlight key milestones and progress in computational protein structure prediction driven by Deep Learning (DL).
• To review DL advancements in various stages of the protein structure prediction pipeline.
• To discuss DL's impact on experimental techniques like Cryo-Electron Microscopy (Cryo-EM) and outline future research directions.

Main Methods:

• Review of Deep Learning (DL) based approaches in protein structure prediction.
• Analysis of DL successes in Critical Assessment of protein Structure Prediction (CASP) experiments.
• Examination of DL applications in contact map prediction, distogram prediction, real-valued distance prediction, and quality assessment/refinement.
• Inclusion of end-to-end DL models and DL advancements in Cryo-Electron Microscopy (Cryo-EM).

Main Results:

• Deep Learning (DL) approaches, exemplified by AlphaFold2, have revolutionized protein structure prediction.
• Significant progress has been made in DL-based prediction of protein contact maps, distograms, and distances.
• DL is enhancing protein structure quality assessment, refinement, and enabling end-to-end prediction pipelines.
• DL is also contributing to advancements in Cryo-Electron Microscopy (Cryo-EM) for structure determination.

Conclusions:

• Deep Learning (DL) has become a transformative force in computational protein structure prediction.
• DL-based methods are rapidly improving the accuracy and efficiency of predicting protein structures.
• Future research directions involve further refining DL models and integrating them with experimental structural biology techniques.