Fine tuned personalized machine learning models to detect insomnia risk based on data from a smart bed platform.

Introduction: Insomnia causes serious adverse health effects and is estimated to affect 10-30% of the worldwide population. This study leverages personalized fine-tuned machine learning algorithms to detect insomnia risk based on questionnaire and longitudinal objective sleep data collected by a smart bed platform. Methods: Users of the Sleep Number smart bed were invited to participate in an IRB approved study which required them to respond to four questionnaires (which included the Insomnia Severity Index; ISI) administered 6 weeks apart from each other in the period from November 2021 to March 2022. For 1,489 participants who completed at least 3 questionnaires, objective data (which includes sleep/wake and cardio-respiratory metrics) collected by the platform were queried for analysis. An incremental, passive-aggressive machine learning model was used to detect insomnia risk which was defined by the ISI exceeding a given threshold. Three ISI thresholds (8, 10, and 15) were considered. The incremental model is advantageous because it allows personalized fine-tuning by adding individual training data to a generic model. Results: The generic model, without personalizing, resulted in an area under the receiving-operating curve (AUC) of about 0.5 for each ISI threshold. The personalized fine-tuning with the data of just five sleep sessions from the individual for whom the model is being personalized resulted in AUCs exceeding 0.8 for all ISI thresholds. Interestingly, no further AUC enhancements resulted by adding personalized data exceeding ten sessions. Discussion: These are encouraging results motivating further investigation into the application of personalized fine tuning machine learning to detect insomnia risk based on longitudinal sleep data and the extension of this paradigm to sleep medicine.

RCSB Protein Data Bank: powerful new tools for exploring 3D structures of biological macromolecules for basic and applied research and education in fundamental biology, biomedicine, biotechnology, bioengineering and energy sciences.

The Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB), the US data center for the global PDB archive and a founding member of the Worldwide Protein Data Bank partnership, serves tens of thousands of data depositors in the Americas and Oceania and makes 3D macromolecular structure data available at no charge and without restrictions to millions of RCSB.org users around the world, including >660 000 educators, students and members of the curious public using PDB101.RCSB.org. PDB data depositors include structural biologists using macromolecular crystallography, nuclear magnetic resonance spectroscopy, 3D electron microscopy and micro-electron diffraction. PDB data consumers accessing our web portals include researchers, educators and students studying fundamental biology, biomedicine, biotechnology, bioengineering and energy sciences. During the past 2 years, the research-focused RCSB PDB web portal (RCSB.org) has undergone a complete redesign, enabling improved searching with full Boolean operator logic and more facile access to PDB data integrated with >40 external biodata resources. New features and resources are described in detail using examples that showcase recently released structures of SARS-CoV-2 proteins and host cell proteins relevant to understanding and addressing the COVID-19 global pandemic.

Real time structural search of the Protein Data Bank.

Detection of protein structure similarity is a central challenge in structural bioinformatics. Comparisons are usually performed at the polypeptide chain level, however the functional form of a protein within the cell is often an oligomer. This fact, together with recent growth of oligomeric structures in the Protein Data Bank (PDB), demands more efficient approaches to oligomeric assembly alignment/retrieval. Traditional methods use atom level information, which can be complicated by the presence of topological permutations within a polypeptide chain and/or subunit rearrangements. These challenges can be overcome by comparing electron density volumes directly. But, brute force alignment of 3D data is a compute intensive search problem. We developed a 3D Zernike moment normalization procedure to orient electron density volumes and assess similarity with unprecedented speed. Similarity searching with this approach enables real-time retrieval of proteins/protein assemblies resembling a target, from PDB or user input, together with resulting alignments (http://shape.rcsb.org).

Lateral A11 type tetramerization in lamins.

The assembly of intermediate filaments (IFs) including nuclear lamins is driven by specific interactions of the elementary coiled-coil dimers in both lateral and longitudinal direction. The assembly mode A11 is dependent on lateral tetramerization of the second coiled-coil segment (coil1b) in antiparallel fashion. Recent cryo-electron microscopy studies pointed to 3.5 nm lamin filaments built from two antiparallel threads of longitudinally associated dimers but little molecular detail is available to date. Here we present the 2.6 Šresolution X-ray structure of a lamin A fragment including residues 65-222 which reveals the molecular basis of the A11 interaction. The crystal structure also indicates a continuous α-helical structure for the preceding linker L1 region. The middle part of the antiparallel tetramer reveals unique interactions due to the lamin-specific 42-residue insert in coil1b. At the same time, distinct characteristics of this insert provide for the preservation of common structural principles shared with lateral coil1b tetramers of vimentin and keratin K1/K10. In addition, structural analysis suggests that the A11 interaction in lamins is somewhat weaker than in cytoplasmic IFs, despite a 30% longer overlap. Establishing the structural detail of the A11 interaction across IF types is the first step towards a rational understanding of the IF assembly process which is indispensable for establishing the mechanism of disease-related mutations.

Small angle X-ray scattering-assisted protein structure prediction in CASP13 and emergence of solution structure differences.

Small angle X-ray scattering (SAXS) measures comprehensive distance information on a protein's structure, which can constrain and guide computational structure prediction algorithms. Here, we evaluate structure predictions of 11 monomeric and oligomeric proteins for which SAXS data were collected and provided to predictors in the 13th round of the Critical Assessment of protein Structure Prediction (CASP13). The category for SAXS-assisted predictions made gains in certain areas for CASP13 compared to CASP12. Improvements included higher quality data with size exclusion chromatography-SAXS (SEC-SAXS) and better selection of targets and communication of results by CASP organizers. In several cases, we can track improvements in model accuracy with use of SAXS data. For hard multimeric targets where regular folding algorithms were unsuccessful, SAXS data helped predictors to build models better resembling the global shape of the target. For most models, however, no significant improvement in model accuracy at the domain level was registered from use of SAXS data, when rigorously comparing SAXS-assisted models to the best regular server predictions. To promote future progress in this category, we identify successes, challenges, and opportunities for improved strategies in prediction, assessment, and communication of SAXS data to predictors. An important observation is that, for many targets, SAXS data were inconsistent with crystal structures, suggesting that these proteins adopt different conformation(s) in solution. This CASP13 result, if representative of PDB structures and future CASP targets, may have substantive implications for the structure training databases used for machine learning, CASP, and use of prediction models for biology.

Assessment of protein assembly prediction in CASP13.

We present the assembly category assessment in the 13th edition of the CASP community-wide experiment. For the second time, protein assemblies constitute an independent assessment category. Compared to the last edition we see a clear uptake in participation, more oligomeric targets released, and consistent, albeit modest, improvement of the predictions quality. Looking at the tertiary structure predictions, we observe that ignoring the oligomeric state of the targets hinders modeling success. We also note that some contact prediction groups successfully predicted homomeric interfacial contacts, though it appears that these predictions were not used for assembly modeling. Homology modeling with sizeable human intervention appears to form the basis of the assembly prediction techniques in this round of CASP. Future developments should see more integrated approaches where subunits are modeled in the context of the assemblies they form.

BioJava 5: A community driven open-source bioinformatics library.

BioJava is an open-source project that provides a Java library for processing biological data. The project aims to simplify bioinformatic analyses by implementing parsers, data structures, and algorithms for common tasks in genomics, structural biology, ontologies, phylogenetics, and more. Since 2012, we have released two major versions of the library (4 and 5) that include many new features to tackle challenges with increasingly complex macromolecular structure data. BioJava requires Java 8 or higher and is freely available under the LGPL 2.1 license. The project is hosted on GitHub at https://github.com/biojava/biojava. More information and documentation can be found online on the BioJava website (http://www.biojava.org) and tutorial (https://github.com/biojava/biojava-tutorial). All inquiries should be directed to the GitHub page or the BioJava mailing list (http://lists.open-bio.org/mailman/listinfo/biojava-l).

RCSB Protein Data Bank: biological macromolecular structures enabling research and education in fundamental biology, biomedicine, biotechnology and energy.

The Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB, rcsb.org), the US data center for the global PDB archive, serves thousands of Data Depositors in the Americas and Oceania and makes 3D macromolecular structure data available at no charge and without usage restrictions to more than 1 million rcsb.org Users worldwide and 600 000 pdb101.rcsb.org education-focused Users around the globe. PDB Data Depositors include structural biologists using macromolecular crystallography, nuclear magnetic resonance spectroscopy and 3D electron microscopy. PDB Data Consumers include researchers, educators and students studying Fundamental Biology, Biomedicine, Biotechnology and Energy. Recent reorganization of RCSB PDB activities into four integrated, interdependent services is described in detail, together with tools and resources added over the past 2 years to RCSB PDB web portals in support of a 'Structural View of Biology.'

Optimal data-driven parameterization of coiled coils.

α-Helical coiled coils (CCs) represent an important, highly regular protein folding motif. To date, many thousands of CC structures have been determined experimentally. Their geometry is usually modelled by theoretical equations introduced by F. Crick that involve a predefined set of parameters. Here we have addressed the problem of efficient CC parameterization from scratch by performing a statistical evaluation of all available CC structures. The procedure is based on the principal component analysis and yields a minimal set of independent parameters that provide for the reconstruction of the complete CC structure at a required precision. The approach is successfully validated on a set of canonical parallel CC dimers. Its applications include all cases where an efficient sampling of the CC geometry is important, such as for solving the phase problem in crystallography.

CCFold: rapid and accurate prediction of coiled-coil structures and application to modelling intermediate filaments.

MOTIVATION: Accurate molecular structure of the protein dimer representing the elementary building block of intermediate filaments (IFs) is essential towards the understanding of the filament assembly, rationalizing their mechanical properties and explaining the effect of disease-related IF mutations. The dimer contains a ∼300-residue long α-helical coiled coil which cannot be assessed by either direct experimental structure determination or modelling using standard approaches. At the same time, coiled coils are well-represented in structural databases. RESULTS: Here we present CCFold, a generally applicable threading-based algorithm which produces coiled-coil models from protein sequence only. The algorithm is based on a statistical analysis of experimentally determined structures and can handle any hydrophobic repeat patterns in addition to the most common heptads. We demonstrate that CCFold outperforms general-purpose computational folding in terms of accuracy, while being faster by orders of magnitude. By combining the CCFold algorithm and Rosetta folding we generate representative dimer models for all IF protein classes. AVAILABILITY AND IMPLEMENTATION: The source code is freely available at https://github.com/biocryst/IF; a web server to run the program is at http://pharm.kuleuven.be/Biocrystallography/cc. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Granular clustering of de novo protein models.

Motivation: Modern algorithms for de novo prediction of protein structures typically output multiple full-length models (decoys) rather than a single solution. Subsequent clustering of such decoys is used both to gauge the success of the modelling and to decide on the most native-like conformation. At the same time, partial protein models are sufficient for some applications such as crystallographic phasing by molecular replacement (MR) in particular, provided these models represent a certain part of the target structure with reasonable accuracy. Results: Here we propose a novel clustering algorithm that natively operates in the space of partial models through an approach known as granular clustering (GC). The algorithm is based on growing local similarities found in a pool of initial decoys. We demonstrate that the resulting clusters of partial models provide a substantially more accurate structural detail on the target protein than those obtained upon a global alignment of decoys. As the result, the partial models output by our GC algorithm are also much more effective towards the MR procedure, compared to the models produced by existing software. Availability and Implementation: The source code is freely available at https://github.com/biocryst/gc Contact: sergei.strelkov@kuleuven.be Suplementary Information: Supplementary data are available at Bioinformatics online.

Crystallographic Studies of Intermediate Filament Proteins.

Intermediate filaments (IFs), together with microtubules and actin microfilaments, are the three main cytoskeletal components in metazoan cells. IFs are formed by a distinct protein family, which is made up of 70 members in humans. Most IF proteins are tissue- or organelle-specific, which includes lamins, the IF proteins of the nucleus. The building block of IFs is an elongated dimer, which consists of a central α-helical 'rod' domain flanked by flexible N- and C-terminal domains. The conserved rod domain is the 'signature feature' of the IF family. Bioinformatics analysis reveals that the rod domain of all IF proteins contains three α-helical segments of largely conserved length, interconnected by linkers. Moreover, there is a conserved pattern of hydrophobic repeats within each segment, which includes heptads and hendecads. This defines the presence of both left-handed and almost parallel coiled-coil regions along the rod length. Using X-ray crystallography on multiple overlapping fragments of IF proteins, the atomic structure of the nearly complete rod domain has been determined. Here, we discuss some specific challenges of this procedure, such as crystallization and diffraction data phasing by molecular replacement. Further insights into the structure of the coiled coil and the terminal domains have been obtained using electron paramagnetic resonance measurements on the full-length protein, with spin labels attached at specific positions. This atomic resolution information, as well as further interesting findings, such as the variation of the coiled-coil stability along the rod length, provide clues towards interpreting the data on IF assembly, collected by a range of methods. However, a full description of this process at the molecular level is not yet at hand.

How to Study Intermediate Filaments in Atomic Detail.

Studies of the intermediate filament (IF) structure are a prerequisite of understanding their function. In addition, the structural information is indispensable if one wishes to gain a mechanistic view on the disease-related mutations in the IFs. Over the years, considerable progress has been made on the atomic structure of the elementary building block of all IFs, the coiled-coil dimer. Here, we discuss the approaches, methods and practices that have contributed to this advance. With abundant genetic information on hand, bioinformatics approaches give important insights into the dimer structure, including the head and tail regions poorly assessable experimentally. At the same time, the most important contribution has been provided by X-ray crystallography. Following the "divide-and-conquer" approach, many fragments from several IF proteins could be crystallized and resolved to atomic resolution. We will systematically cover the main procedures of these crystallographic studies, suggest ways to maximize their efficiency, and also discuss the possible pitfalls and limitations. In addition, electron paramagnetic resonance with site-directed spin labeling was another method providing a major impact toward the understanding of the IF structure. Upon placing the spin labels into specific positions within the full-length protein, one can evaluate the proximity of the labels and their mobility. This makes it possible to make conclusions about the dimer structure in the coiled-coil region and beyond, as well as to explore the dimer-dimer contacts.

Intermediate filament structure: the bottom-up approach.

Intermediate filaments (IFs) result from a key cytoskeletal protein class in metazoan cells, but currently there is no consensus as to their three-dimensional architecture. IF proteins form elongated dimers based on the coiled-coil structure within their central 'rod' domain. Here we focus on the atomic structure of this elementary dimer, elucidated using X-ray crystallography on multiple fragments and electron paramagnetic resonance experiments on spin-labelled vimentin samples. In line with conserved sequence features, the rod of all IF proteins is composed of three coiled-coil segments containing heptad and hendecad repeats and interconnected by linkers. In addition, the next assembly intermediate beyond the dimer, the tetramer, could be modelled. The impact of these structural results towards understanding the assembly mechanism is discussed.

The deubiquitylase USP33 discriminates between RALB functions in autophagy and innate immune response.

The RAS-like GTPase RALB mediates cellular responses to nutrient availability or viral infection by respectively engaging two components of the exocyst complex, EXO84 and SEC5. RALB employs SEC5 to trigger innate immunity signalling, whereas RALB-EXO84 interaction induces autophagocytosis. How this differential interaction is achieved molecularly by the RAL GTPase remains unknown. We found that whereas GTP binding turns on RALB activity, ubiquitylation of RALB at Lys 47 tunes its activity towards a particular effector. Specifically, ubiquitylation at Lys 47 sterically inhibits RALB binding to EXO84, while facilitating its interaction with SEC5. Double-stranded RNA promotes RALB ubiquitylation and SEC5-TBK1 complex formation. In contrast, nutrient starvation induces RALB deubiquitylation by accumulation and relocalization of the deubiquitylase USP33 to RALB-positive vesicles. Deubiquitylated RALB promotes the assembly of the RALB-EXO84-beclin-1 complexes driving autophagosome formation. Thus, ubiquitylation within the effector-binding domain provides the switch for the dual functions of RALB in autophagy and innate immune responses.