Most proteins fold into 3D structures that determine how they function and orchestrate the biological processes of the cell. Recent developments in computational methods have led to protein structure predictions that have reached the accuracy of experimentally determined models. While this has been independently verified, the implementation of these methods across structural biology applications remains to be tested. Here, we evaluate the use of AlphaFold 2 (AF2) predictions in the study of characteristic structural elements; the impact of missense variants; function and ligand binding site predictions; modelling of interactions; and modelling of experimental structural data. For 11 proteomes, an average of 25% additional residues can be confidently modelled when compared to homology modelling, identifying structural features rarely seen in the PDB. AF2-based predictions of protein disorder and protein complexes surpass state-of-the-art tools and AF2 models can be used across diverse applications equally well compared to experimentally determined structures, when the confidence metrics are critically considered. In summary, we find that these advances are likely to have a transformative impact in structural biology and broader life science research.

Protein-ligand interactions (PLI) are foundational to small molecule drug design. With computational methods striving towards experimental accuracy, there is a critical demand for a well-curated and diverse PLI dataset. Existing datasets are often limited in size and diversity, and commonly used evaluation sets suffer from training information leakage, hindering the realistic assessment of method generalization capabilities. To address these shortcomings, we present PLIN-DER, the largest and most annotated dataset to date, comprising 449,383 PLI systems, each with over 500 annotations, similarity metrics at protein, pocket, interaction and ligand levels, and paired unbound ( apo ) and predicted structures. We propose an approach to generate training and evaluation splits that minimizes task-specific leakage and maximizes test set quality, and compare the resulting performance of DiffDock when retrained with different kinds of splits.

The vast number of protein structures currently available opens exciting opportunities for machine learning on proteins, aimed at predicting and understanding functional properties. In particular, in combination with homology modelling, it is now possible to not only use sequence features as input for machine learning, but also structure features. However, in order to do so, robust multiple structure alignments are imperative.

Protein-protein interactions (PPIs) are fundamental to understanding biological processes and play a key role in therapeutic advancements. As deep-learning docking methods for PPIs gain traction, benchmarking protocols and datasets tailored for effective training and evaluation of their generalization capabilities and performance across real-world scenarios become imperative. Aiming to overcome limitations of existing approaches, we introduce PINDER, a comprehensive annotated dataset that uses structural clustering to derive non-redundant interface-based data splits and includes holo (bound), apo (unbound), and computationally predicted structures. PINDER consists of 2,319,564 dimeric PPI systems (and up to 25 million augmented PPIs) and 1,955 high-quality test PPIs with interface data leakage removed. Additionally, PINDER provides a test subset with 180 dimers for comparison to AlphaFold-Multimer without any interface leakage with respect to its training set. Unsurprisingly, the PINDER benchmark reveals that the performance of existing docking models is highly overestimated when evaluated on leaky test sets. Most importantly, by retraining DiffDock-PP on PINDER interface-clustered splits, we show that interface cluster-based sampling of the training split, along with the diverse and less leaky validation split, leads to strong generalization improvements.

The process of designing biomolecules, in particular proteins, is witnessing a rapid change in available tooling and approaches, moving from design through physicochemical force fields, to producing plausible, complex sequences fast via end-to-end differentiable statistical models. To achieve conditional and controllable protein design, researchers at the interface of artificial intelligence and biology leverage advances in natural language processing (NLP) and computer vision techniques, coupled with advances in computing hardware to learn patterns from growing biological databases, curated annotations thereof, or both. Once learned, these patterns can be leveraged to provide novel insights into mechanistic biology and the design of biomolecules. However, navigating and understanding the practical applications for the many recent protein design tools is complex. To facilitate this, we 1) document recent advances in deep learning (DL) assisted protein design from the last three years, 2) present a practical pipeline that allows to go from de novo -generated sequences to their predicted properties and web-powered visualization within minutes, and 3) leverage it to suggest a generated protein sequence which might be used to engineer a biosynthetic gene cluster to produce a molecular glue-like compound. Lastly, we discuss challenges and highlight opportunities for the protein design field. Availability: pLM generated and UniRef50 sampled sequence sets and predictions are available at http://data.bioembeddings.com/public/design . Code-base and Notebooks for analysis are available at https://github.com/hefeda/PGP . An online version of Table 1 can be found at https://github.com/hefeda/design_tools .

Driven by the development and upscaling of fast genome sequencing and assembly pipelines, the number of protein-coding sequences deposited in public protein sequence databases is increasing exponentially. Recently, the dramatic success of deep learning-based approaches applied to protein structure prediction has done the same for protein structures. We are now entering a new era in protein sequence and structure annotation, with hundreds of millions of predicted protein structures made available through the AlphaFold database. These models cover most of the catalogued natural proteins, including those difficult to annotate for function or putative biological role based on standard, homology-based approaches. In this work, we quantified how much of such “dark matter” of the natural protein universe was structurally illuminated by AlphaFold2 and modelled this diversity as an interactive sequence similarity network that can be navigated at https://uniprot3d.org/atlas/AFDB90v4 . In the process, we discovered multiple novel protein families by searching for novelties from sequence, structure, and semantic perspectives. We added a number of them to Pfam, and experimentally demonstrate that one of these belongs to a novel superfamily of translation-targeting toxin-antitoxin systems, TumE-TumA. This work highlights the role of large-scale, evolution-driven protein comparison efforts in combination with structural similarities, genomic context conservation, and deep-learning based function prediction tools for the identification of novel protein families, aiding not only annotation and classification efforts but also the curation and prioritisation of target proteins for experimental characterisation.