DPAM: A domain parser for AlphaFold models

DomainFit: Identification of protein domains in cryo-EM maps at intermediate resolution using AlphaFold2-predicted models

Cryoelectron microscopy (cryo-EM) has revolutionized the structural determination of macromolecular complexes. With the paradigm shift to structure determination of highly complex endogenous macromolecular complexes ex vivo and in situ structural biology, there are an increasing number of structures of native complexes. These complexes often contain unidentified proteins, related to different cellular states or processes. Identifying proteins at resolutions lower than 4 Å remains challenging because side chains cannot be visualized reliably. Here, we present DomainFit, a program for semi-automated domain-level protein identification from cryo-EM maps, particularly at resolutions lower than 4 Å. By fitting domains from AlphaFold2-predicted models into cryo-EM maps, the program performs statistical analyses and attempts to identify the domains and protein candidates forming the density. Using DomainFit, we identified two microtubule inner proteins, one of which contains a CCDC81 domain and is exclusively localized in the proximal region of the doublet microtubule in Tetrahymena thermophila.

Deep learning for the PSIPRED Protein Analysis Workbench

The PSIRED Workbench is a long established and popular bioinformatics web service offering a wide range of machine learning based analyses for characterizing protein structure and function. In this paper we provide an update of the recent additions and developments to the webserver, with a focus on new Deep Learning based methods. We briefly discuss some trends in server usage since the publication of AlphaFold2 and we give an overview of some upcoming developments for the service. The PSIPRED Workbench is available at http://bioinf.cs.ucl.ac.uk/psipred.

Structure classification of the proteins from Salmonella enterica pangenome revealed novel potential pathogenicity islands

Salmonella enterica is a pathogenic bacterium known for causing severe typhoid fever in humans, making it important to study due to its potential health risks and significant impact on public health. This study provides evolutionary classification of proteins from Salmonella enterica pangenome. We classified 17,238 domains from 13,147 proteins from 79,758 Salmonella enterica strains and studied in detail domains of 272 proteins from 14 characterized Salmonella pathogenicity islands (SPIs). Among SPIs-related proteins, 90 proteins function in the secretion machinery. 41% domains of SPI proteins have no previous sequence annotation. By comparing clinical and environmental isolates, we identified 3682 proteins that are overrepresented in clinical group that we consider as potentially pathogenic. Among domains of potentially pathogenic proteins only 50% domains were annotated by sequence methods previously. Moreover, 36% (1330 out of 3682) of potentially pathogenic proteins cannot be classified into Evolutionary Classification of Protein Domains database (ECOD). Among classified domains of potentially pathogenic proteins the most populated homology groups include helix-turn-helix (HTH), Immunoglobulin-related, and P-loop domains-related. Functional analysis revealed overrepresentation of these protein in biological processes related to viral entry into host cell, antibiotic biosynthesis, DNA metabolism and conformation change, and underrepresentation in translational processes. Analysis of the potentially pathogenic proteins indicates that they form 119 clusters or novel potential pathogenicity islands (NPPIs) within the Salmonella genome, suggesting their potential contribution to the bacterium's virulence. One of the NPPIs revealed significant overrepresentation of potentially pathogenic proteins. Overall, our analysis revealed that identified potentially pathogenic proteins are poorly studied.

Chainsaw: protein domain segmentation with fully convolutional neural networks

MOTIVATION

Protein domains are fundamental units of protein structure and play a pivotal role in understanding folding, function, evolution, and design. The advent of accurate structure prediction techniques has resulted in an influx of new structural data, making the partitioning of these structures into domains essential for inferring evolutionary relationships and functional classification.

RESULTS

This article presents Chainsaw, a supervised learning approach to domain parsing that achieves accuracy that surpasses current state-of-the-art methods. Chainsaw uses a fully convolutional neural network which is trained to predict the probability that each pair of residues is in the same domain. Domain predictions are then derived from these pairwise predictions using an algorithm that searches for the most likely assignment of residues to domains given the set of pairwise co-membership probabilities. Chainsaw matches CATH domain annotations in 78% of protein domains versus 72% for the next closest method. When predicting on AlphaFold models, expert human evaluators were twice as likely to prefer Chainsaw's predictions versus the next best method.

AVAILABILITY AND IMPLEMENTATION

github.com/JudeWells/Chainsaw.

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

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.

RefSeq and the prokaryotic genome annotation pipeline in the age of metagenomes

The Reference Sequence (RefSeq) project at the National Center for Biotechnology Information (NCBI) contains over 315 000 bacterial and archaeal genomes and 236 million proteins with up-to-date and consistent annotation. In the past 3 years, we have expanded the diversity of the RefSeq collection by including the best quality metagenome-assembled genomes (MAGs) submitted to INSDC (DDBJ, ENA and GenBank), while maintaining its quality by adding validation checks. Assemblies are now more stringently evaluated for contamination and for completeness of annotation prior to acceptance into RefSeq. MAGs now account for over 17000 assemblies in RefSeq, split over 165 orders and 362 families. Changes in the Prokaryotic Genome Annotation Pipeline (PGAP), which is used to annotate nearly all RefSeq assemblies include better detection of protein-coding genes. Nearly 83% of RefSeq proteins are now named by a curated Protein Family Model, a 4.7% increase in the past three years ago. In addition to literature citations, Enzyme Commission numbers, and gene symbols, Gene Ontology terms are now assigned to 48% of RefSeq proteins, allowing for easier multi-genome comparison. RefSeq is found at https://www.ncbi.nlm.nih.gov/refseq/. PGAP is available as a stand-alone tool able to produce GenBank-ready files at https://github.com/ncbi/pgap.

Merizo: a rapid and accurate protein domain segmentation method using invariant point attention

The AlphaFold Protein Structure Database, containing predictions for over 200 million proteins, has been met with enthusiasm over its potential in enriching structural biological research and beyond. Currently, access to the database is precluded by an urgent need for tools that allow the efficient traversal, discovery, and documentation of its contents. Identifying domain regions in the database is a non-trivial endeavour and doing so will aid our understanding of protein structure and function, while facilitating drug discovery and comparative genomics. Here, we describe a deep learning method for domain segmentation called Merizo, which learns to cluster residues into domains in a bottom-up manner. Merizo is trained on CATH domains and fine-tuned on AlphaFold2 models via self-distillation, enabling it to be applied to both experimental and AlphaFold2 models. As proof of concept, we apply Merizo to the human proteome, identifying 40,818 putative domains that can be matched to CATH representative domains.

DomainFit: Identification of Protein Domains in cryo-EM maps at Intermediate Resolution using AlphaFold2-predicted Models

Cryo-electron microscopy (cryo-EM) has revolutionized our understanding of macromolecular complexes, enabling high-resolution structure determination. With the paradigm shift to in situ structural biology recently driven by the ground-breaking development of cryo-focused ion beam milling and cryo-electron tomography, there are an increasing number of structures at sub-nanometer resolution of complexes solved directly within their cellular environment. These cellular complexes often contain unidentified proteins, related to different cellular states or processes. Identifying proteins at resolutions lower than 4 Å remains challenging because the side chains cannot be visualized reliably. Here, we present DomainFit, a program for automated domain-level protein identification from cryo-EM maps at resolutions lower than 4 Å. By fitting domains from artificial intelligence-predicted models such as AlphaFold2-predicted models into cryo-EM maps, the program performs statistical analyses and attempts to identify the proteins forming the density. Using DomainFit, we identified two microtubule inner proteins, one of them, a CCDC81 domain-containing protein, is exclusively localized in the proximal region of the doublet microtubule from the ciliate Tetrahymena thermophila. The flexibility and capability of DomainFit makes it a valuable tool for analyzing in situ structures.

Genetic manipulation of Patescibacteria provides mechanistic insights into microbial dark matter and the epibiotic lifestyle

Patescibacteria, also known as the candidate phyla radiation (CPR), are a diverse group of bacteria that constitute a disproportionately large fraction of microbial dark matter. Its few cultivated members, belonging mostly to Saccharibacteria, grow as epibionts on host Actinobacteria. Due to a lack of suitable tools, the genetic basis of this lifestyle and other unique features of Patescibacteira remain unexplored. Here, we show that Saccharibacteria exhibit natural competence, and we exploit this property for their genetic manipulation. Imaging of fluorescent protein-labeled Saccharibacteria provides high spatiotemporal resolution of phenomena accompanying epibiotic growth, and a transposon-insertion sequencing (Tn-seq) genome-wide screen reveals the contribution of enigmatic Saccharibacterial genes to growth on their hosts. Finally, we leverage metagenomic data to provide cutting-edge protein structure-based bioinformatic resources that support the strain Southlakia epibionticum and its corresponding host, Actinomyces israelii, as a model system for unlocking the molecular underpinnings of the epibiotic lifestyle.

Computational analysis of regulatory regions in human protein kinases

Eukaryotic proteins often feature modular domain structures comprising globular domains that are connected by linker regions and intrinsically disordered regions that may contain important functional motifs. The intramolecular interactions of globular domains and nonglobular regions can play critical roles in different aspects of protein function. However, studying these interactions and their regulatory roles can be challenging due to the flexibility of nonglobular regions, the long insertions separating interacting modules, and the transient nature of some interactions. Obtaining the experimental structures of multiple domains and functional regions is more difficult than determining the structures of individual globular domains. High-quality structural models generated by AlphaFold offer a unique opportunity to study intramolecular interactions in eukaryotic proteins. In this study, we systematically explored intramolecular interactions between human protein kinase domains (KDs) and potential regulatory regions, including globular domains, N- and C-terminal tails, long insertions, and distal nonglobular regions. Our analysis identified intramolecular interactions between human KDs and 35 different types of globular domains, exhibiting a variety of interaction modes that could contribute to orthosteric or allosteric regulation of kinase activity. We also identified prevalent interactions between human KDs and their flanking regions (N- and C-terminal tails). These interactions exhibit group-specific characteristics and can vary within each specific kinase group. Although long-range interactions between KDs and nonglobular regions are relatively rare, structural details of these interactions offer new insights into the regulation mechanisms of several kinases, such as HASPIN, MAPK7, MAPK15, and SIK1B.

Structural homology screens reveal host-derived poxvirus protein families impacting inflammasome activity

Viruses acquire host genes via horizontal transfer and can express them to manipulate host biology during infections. Some homologs retain sequence identity, but evolutionary divergence can obscure host origins. We use structural modeling to compare vaccinia virus proteins with metazoan proteomes. We identify vaccinia A47L as a homolog of gasdermins, the executioners of pyroptosis. An X-ray crystal structure of A47 confirms this homology, and cell-based assays reveal that A47 interferes with caspase function. We also identify vaccinia C1L as the product of a cryptic gene fusion event coupling a Bcl-2-related fold with a pyrin domain. C1 associates with components of the inflammasome, a cytosolic innate immune sensor involved in pyroptosis, yet paradoxically enhances inflammasome activity, suggesting differential modulation during infections. Our findings demonstrate the increasing power of structural homology screens to reveal proteins with unique combinations of domains that viruses capture from host genes and combine in unique ways.

Genetic manipulation of candidate phyla radiation bacteria provides functional insights into microbial dark matter

The study of bacteria has yielded fundamental insights into cellular biology and physiology, biotechnological advances and many therapeutics. Yet due to a lack of suitable tools, the significant portion of bacterial diversity held within the candidate phyla radiation (CPR) remains inaccessible to such pursuits. Here we show that CPR bacteria belonging to the phylum Saccharibacteria exhibit natural competence. We exploit this property to develop methods for their genetic manipulation, including the insertion of heterologous sequences and the construction of targeted gene deletions. Imaging of fluorescent protein-labeled Saccharibacteria provides high spatiotemporal resolution of phenomena accompanying epibiotic growth and a transposon insertion sequencing genome-wide screen reveals the contribution of enigmatic Saccharibacterial genes to growth on their Actinobacteria hosts. Finally, we leverage metagenomic data to provide cutting-edge protein structure-based bioinformatic resources that support the strain Southlakia epibionticum and its corresponding host, Actinomyces israelii , as a model system for unlocking the molecular underpinnings of the epibiotic lifestyle.

Structural homology screens reveal poxvirus-encoded proteins impacting inflammasome-mediated defenses

Viruses acquire host genes via horizontal gene transfer and can express them to manipulate host biology during infections. Some viral and host homologs retain sequence identity, but evolutionary divergence can obscure host origins. We used structural modeling to compare vaccinia virus proteins with metazoan proteomes. We identified vaccinia A47L as a homolog of gasdermins, the executioners of pyroptosis. An X-ray crystal structure of A47 confirmed this homology and cell-based assays revealed that A47 inhibits pyroptosis. We also identified vaccinia C1L as the product of a cryptic gene fusion event coupling a Bcl-2 related fold with a pyrin domain. C1 associates with components of the inflammasome, a cytosolic innate immune sensor involved in pyroptosis, yet paradoxically enhances inflammasome activity, suggesting a benefit to poxvirus replication in some circumstances. Our findings demonstrate the potential of structural homology screens to reveal genes that viruses capture from hosts and repurpose to benefit viral fitness.

Classification of domains in predicted structures of the human proteome

Recent advances in protein structure prediction have generated accurate structures of previously uncharacterized human proteins. Identifying domains in these predicted structures and classifying them into an evolutionary hierarchy can reveal biological insights. Here, we describe the detection and classification of domains from the human proteome. Our classification indicates that only 62% of residues are located in globular domains. We further classify these globular domains and observe that the majority (65%) can be classified among known folds by sequence, with a smaller fraction (33%) requiring structural data to refine the domain boundaries and/or to support their homology. A relatively small number (966 domains) cannot be confidently assigned using our automatic pipelines, thus demanding manual inspection. We classify 47,576 domains, of which only 23% have been included in experimental structures. A portion (6.3%) of these classified globular domains lack sequence-based annotation in InterPro. A quarter (23%) have not been structurally modeled by homology, and they contain 2,540 known disease-causing single amino acid variations whose pathogenesis can now be inferred using AF models. A comparison of classified domains from a series of model organisms revealed expansions of several immune response-related domains in humans and a depletion of olfactory receptors. Finally, we use this classification to expand well-known protein families of biological significance. These classifications are presented on the ECOD website (http://prodata.swmed.edu/ecod/index_human.php).