AlphaFold in protein structure and function prediction

The introduction of AlphaFold2 [1], its subsequent AlphaFold3 version [2], and the public release of hundreds of millions of protein structures predicted by this landmark method [3], [4] have fundamentally changed the fields of structural biology and protein bioinformatics. Structural biology, which once relied on painstaking and relatively slow experimental determination of protein structure, has become one of abundance, with structural models now available for nearly every known protein sequence. Protein bioinformatics has evolved from developing structure prediction tools to analyzing and applying structural models. This special issue of the Computational and Structural Biotechnology Journal commemorates these transitions by exploring the many diverse ways in which AlphaFold is used. We initiated this special issue in July of 2024 and were flooded with submissions. We extended the original submission deadline from October 31, 2024, to January 31, 2025, to accommodate this strong interest. The resulting special issue includes 32 articles that collectively define the current state of structural bioinformatics. We are deeply grateful to the authors who entrusted us with their studies, the reviewers who provided their timely and insightful advice, and the editorial team for their help with this process and encouragement to organize this special issue. Following, we provide a brief overview of our special issue.

A large group of articles in our special issue focuses on critical and rigorous benchmarking and corresponding limitations of AlphaFold. Lu et al. [5] review advances and challenges in multimeric complex prediction. Thanu et al. [6] show that traditional template-based modelling still outperforms AlphaFold 3 for insect odorant receptors, while Surya et al. [7] reveal erroneous homodimer predictions by AlphaFold-Multimer for flavivirus NS4B proteins. Yu et al. [8] provide a large-scale proteome-wide assessment of secondary structure and solvent accessibility predictions generated from AlphaFold results, revealing that they outperform other methods. Some works highlight limitations in conformational sampling [9] and in the prediction of intrinsically disordered binding regions [10].

Another cluster of contributions centers on innovations by introducing new tools and predictors built upon AlphaFold outputs. These resources include SpatialPPIv2 for protein-protein interface prediction [11], VarMeter2 for pathogenic variant prediction [12], Struct2SL for synthetic-lethality prediction [13], bindNode24 for binding-site prediction [14], and ToxDL 2.0 for toxicity prediction [15]. Articles in this group also address the prediction of enzymatic functions that combines structure and sequence-derived information [16], physics-aware approach for evaluation of protein complex models [17], improvements to the prediction of structures for chimeric proteins using windowed multiple sequence alignments [18], and improvements to the modelling of the protein folding process [19].

This special issue also encompasses several hybrid approaches that investigate AlphaFold structures together with molecular dynamics, docking, and/or experimental validation. Examples include accelerated molecular dynamics-based exploration of fungal hydrophobins [20], conformational sampling of the mitochondrial ADP/ATP carrier [21], LisghtDock-based antibody optimization [22], HADDOCK/MD refinement of antibody-antigen interfaces [23], and mass spectrometry-guided modelling of nanobody-antigen interactions [24]. Authors also explore topoisomerase substrate binding with AlphaFold 3 models [25], pLDDT-weighted flexibility simulations [26], and the use of molecular dynamics in membrane positioning for 7-transmembrane RGS proteins [27] and in the context of predicting receptor-ligand binding affinity in plant-microbe interfaces [28].

Several studies concentrate on tackling challenging systems where solutions are facilitated by new structural models. These works cover drug-target screening in Trypanosoma cruzi [29], structural characterization of novel bacteriophage components [30], LSm-like phage proteins [31], bacterial DinG-like proteins [32], prokaryotic PolB2 and Pol V mutasomes [33], and disease-associated variants, such as the Alzheimer’s-linked TREM2 R47H [34], wound-healing molecules from fish proteins [35], and plasticity of the interleukin-6 protein family [36].

The articles in this collection demonstrate that we have entered a "post-structure" era where the protein structure itself is no longer the final goal, but rather it is an input or variable in a larger equation. By exploring the breadth and quality of the AlphaFold predictions in the context of the physical rigor of molecular dynamics, benchmarking and experimental validation efforts, and building a new generation of tools for protein folding and function modelling, the research in this special issue paves the way for a deeper, more accurate, dynamic, and actionable understanding of biological systems.