Abstract:AlphaFold2 (AF2) and RoseTTaFold (RF) have revolutionized structural biology, serving as highly reliable and effective methods for predicting protein structures. This article explores their impact and limitations, focusing on their integration into experimental pipelines and their application in diverse protein classes, including membrane proteins, intrinsically disordered proteins (IDPs), and oligomers. In experimental pipelines, AF2 models help X-ray crystallography in resolving the phase problem, while complementarity with mass spectrometry and NMR data enhances structure determination and protein flexibility prediction. Predicting the structure of membrane proteins remains challenging for both AF2 and RF due to difficulties in capturing conformational ensembles and interactions with the membrane. Improvements in incorporating membrane-specific features and predicting the structural effect of mutations are crucial. For intrinsically disordered proteins, AF2's confidence score (pLDDT) serves as a competitive disorder predictor, but integrative approaches including molecular dynamics (MD) simulations or hydrophobic cluster analyses are advocated for accurate dynamics representation. AF2 and RF show promising results for oligomeric models, outperforming traditional docking methods, with AlphaFold-Multimer showing improved performance. However, some caveats remain in particular for membrane proteins. Real-life examples demonstrate AF2's predictive capabilities in unknown protein structures, but models should be evaluated for their agreement with experimental data. Furthermore, AF2 models can be used complementarily with MD simulations. In this Perspective, we propose a "wish list" for improving deep-learning-based protein folding prediction models, including using experimental data as constraints and modifying models with binding partners or post-translational modifications. Additionally, a meta-tool for ranking and suggesting composite models is suggested, driving future advancements in this rapidly evolving field.

Overview of AlphaFold2 and breakthroughs in overcoming its limitations

AlphaFold2 and its applications in the fields of biology and medicine

AlphaFold 2: Why It Works and Its Implications for Understanding the Relationships of Protein Sequence, Structure, and Function

AlphaFold2 knows some protein folding principles

AlphaFold2 for protein structure prediction: Best practices and critical analyses

A Perspective on the Prospective Use of AI in Protein Structure Prediction

AlphaFold2 in biomedical research: facilitating the development of diagnostic strategies for disease

The power and pitfalls of AlphaFold2 for structure prediction beyond rigid globular proteins

Dissecting AlphaFolds Capabilities with Limited Sequence Information

AI-Based Protein Structure Prediction in Drug Discovery: Impacts and Challenges

Protein structure prediction in the deep learning era

AlphaFold-latest: revolutionizing protein structure prediction for comprehensive biomolecular insights and therapeutic advancements

AlphaFold2 transmembrane protein structure prediction shines

Prediction of peptide structural conformations with AlphaFold2

Recent Advances and Challenges in Protein Structure Prediction

Assessing AF2's ability to predict structural ensembles of proteins

Recent Progress of Protein Tertiary Structure Prediction

Prediction of protein structure and AI

AI told you so: navigating protein structure prediction in the era of machine learning

Ins and outs of AlphaFold2 transmembrane protein structure predictions

Highly accurate protein structure prediction with AlphaFold