Quantifying the deformability of malaria-infected red blood cells using deep learning trained on synthetic cells.

Several hematologic diseases, including malaria, diabetes, and sickle cell anemia, result in a reduced red blood cell deformability. This deformability can be measured using a microfluidic device with channels of varying width. Nevertheless, it is challenging to algorithmically recognize large numbers of red blood cells and quantify their deformability from image data. Deep learning has become the method of choice to handle noisy and complex image data. However, it requires a significant amount of labeled data to train the neural networks. By creating images of cells and mimicking noise and plasticity in those images, we generate synthetic data to train a network to detect and segment red blood cells from video-recordings, without the need for manually annotated labels. Using this new method, we uncover significant differences between the deformability of RBCs infected with different strains of Plasmodium falciparum, providing clues to the variation in virulence of these strains.

Entropy and Variability: A Second Opinion by Deep Learning.

BACKGROUND: Analysis of the distribution of amino acid types found at equivalent positions in multiple sequence alignments has found applications in human genetics, protein engineering, drug design, protein structure prediction, and many other fields. These analyses tend to revolve around measures of the distribution of the twenty amino acid types found at evolutionary equivalent positions: the columns in multiple sequence alignments. Commonly used measures are variability, average hydrophobicity, or Shannon entropy. One of these techniques, called entropy-variability analysis, as the name already suggests, reduces the distribution of observed residue types in one column to two numbers: the Shannon entropy and the variability as defined by the number of residue types observed. RESULTS: We applied a deep learning, unsupervised feature extraction method to analyse the multiple sequence alignments of all human proteins. An auto-encoder neural architecture was trained on 27,835 multiple sequence alignments for human proteins to obtain the two features that best describe the seven million variability patterns. These two unsupervised learned features strongly resemble entropy and variability, indicating that these are the projections that retain most information when reducing the dimensionality of the information hidden in columns in multiple sequence alignments.

Structural anomalies in a published NMR-derived structure of IRAK-M.

Signaling by Toll-Like Receptors and the Interleukin-1 Receptor (IL1-R) involves intracellular binding of MyD88, followed by assembly of IL1-R Associated Kinases (IRAKs) into the so-called Myddosome. Using NMR, Nechama et al. determined the structure of the IRAK-M death domain monomer (PDBid: 5UKE). With this structure, they performed a docking study to model the location of IRAK-M in the Myddosome. Based on this, they present a molecular basis for selectivity of IRAK-M towards IRAK1 over IRAK2 binding. When we attempted to use 5UKE as a homology modeling template, we noticed that our 5UKE-based models had structural issues, such as disallowed torsion angles and solvent exposed tryptophans. We therefore analyzed the NMR ensemble of 5UKE using structure validation tools and we compared 5UKE with homologous high-resolution X-ray structures. We identified several structural anomalies in 5UKE, including packing issues, frayed helices and improbable side chain conformations. We used Yasara to build a homology model, based on two high resolution death domain crystal structures, as an alternative model for the IRAK-M death domain (atomic coordinates, modeling details and validation are available at https://swift.cmbi.umcn.nl/gv/service/5uke/). Our model agrees better with known death domain structure information than 5UKE and also with the chemical shift data that was deposited for 5UKE.

Motions around conserved helical weak spots facilitate GPCR activation.

G protein-coupled receptors (GPCRs) participate in most physiological processes and are important drug targets in many therapeutic areas. Recently, many GPCR X-ray structures became available, facilitating detailed studies of their sequence-structure-mobility-function relations. We show that the functional role of many conserved GPCR sequence motifs is to create weak spots in the transmembrane helices that provide the structural plasticity necessary for ligand binding and signaling. Different receptor families use different conserved sequence motifs to obtain similar helix irregularities that allow for the same motions upon GPCR activation. These conserved motions come together to facilitate the timely release of the conserved sodium ion to the cytosol. Most GPCR crystal structures could be determined only after stabilization of the transmembrane helices by mutations that remove weak spots. These mutations often lead to diminished binding of agonists, but not antagonists, which logically agrees with the fact that large helix rearrangements occur only upon agonist binding. Upon activation, six of the seven TM helices in GPCRs undergo helix motions and/or deformations facilitated by weak spots in these helices. The location of these weak spots is much more conserved than the sequence motifs that cause them. Knowledge about these weak spots helps understand the activation process of GPCRs and thus helps design medicines.

The Future of Protein Secondary Structure Prediction Was Invented by Oleg Ptitsyn.

When Oleg Ptitsyn and his group published the first secondary structure prediction for a protein sequence, they started a research field that is still active today. Oleg Ptitsyn combined fundamental rules of physics with human understanding of protein structures. Most followers in this field, however, use machine learning methods and aim at the highest (average) percentage correctly predicted residues in a set of proteins that were not used to train the prediction method. We show that one single method is unlikely to predict the secondary structure of all protein sequences, with the exception, perhaps, of future deep learning methods based on very large neural networks, and we suggest that some concepts pioneered by Oleg Ptitsyn and his group in the 70s of the previous century likely are today's best way forward in the protein secondary structure prediction field.

Facilities that make the PDB data collection more powerful.

We describe a series of databases and tools that directly or indirectly support biomedical research on macromolecules, with focus on their applicability in protein structure bioinformatics research. DSSP, that determines secondary structures of proteins, has been updated to work well with extremely large structures in multiple formats. The PDBREPORT database that lists anomalies in protein structures has been remade to remove many small problems. These reports are now available as PDF-formatted files with a computer-readable summary. The VASE software has been added to analyze and visualize HSSP multiple sequence alignments for protein structures. The Lists collection of databases has been extended with a series of databases, most noticeably with a database that gives each protein structure a grade for usefulness in protein structure bioinformatics projects. The PDB-REDO collection of reanalyzed and re-refined protein structures that were solved by X-ray crystallography has been improved by dealing better with sugar residues and with hydrogen bonds, and adding many missing surface loops. All academic software underlying these protein structure bioinformatics applications and databases are now publicly accessible, either directly from the authors or from the GitHub software repository.

G protein-coupled receptors of class A harness the energy of membrane potential to increase their sensitivity and selectivity.

The human genome contains about 700 genes of G protein-coupled receptors (GPCRs) of class A; these seven-helical membrane proteins are the targets of almost half of all known drugs. In the middle of the helix bundle, crystal structures reveal a highly conserved sodium-binding site, which is connected with the extracellular side by a water-filled tunnel. This binding site contains a sodium ion in those GPCRs that are crystallized in their inactive conformations but does not in those GPCRs that are trapped in agonist-bound active conformations. The escape route of the sodium ion upon the inactive-to-active transition and its very direction have until now remained obscure. Here, by modeling the available experimental data, we show that the sodium gradient over the cell membrane increases the sensitivity of GPCRs if their activation is thermodynamically coupled to the sodium ion translocation into the cytoplasm but decreases it if the sodium ion retreats into the extracellular space upon receptor activation. The model quantitatively describes the available data on both activation and suppression of distinct GPCRs by membrane voltage. The model also predicts selective amplification of the signal from (endogenous) agonists if only they, but not their (partial) analogs, induce sodium translocation. Comparative structure and sequence analyses of sodium-binding GPCRs indicate a key role for the conserved leucine residue in the second transmembrane helix (Leu2.46) in coupling sodium translocation to receptor activation. Hence, class A GPCRs appear to harness the energy of the transmembrane sodium potential to increase their sensitivity and selectivity.

The bio.tools registry of software tools and data resources for the life sciences.

Bioinformaticians and biologists rely increasingly upon workflows for the flexible utilization of the many life science tools that are needed to optimally convert data into knowledge. We outline a pan-European enterprise to provide a catalogue ( https://bio.tools ) of tools and databases that can be used in these workflows. bio.tools not only lists where to find resources, but also provides a wide variety of practical information.

Loop modelling 1.0.

Engineering surface loops is a sub-topic of protein engineering that is used routinely in many research fields in academia and industry alike. We provide some tools that search in the PDB for loops satisfying a wide variety of constraints. We illustrate the usefulness of these tools by applying them to a series of recently published studies that included loop engineering or loop modelling. LoopFinder finds loops that fit between two anchor stretches of typically 2, 3, or 4 amino acids each. ProDA find loops of a given length with predefined secondary structure, residue types, hydrophobicity, etc. WHAT IF has gotten a series of new options to scan the whole PDB for loops combining the LoopFinder and ProDA techniques. The open nature of these tools will allow bioinformaticians in this field to easily design their own loop modelling software around our tools. AVAILABILITY AND IMPLEMENTATION: LoopFinder is a stand-alone Fortran program that is likely to compile and run on every computer. The LoopFinder source code, data files, and documentation are freely available from swift.cmbi.ru.nl/gv/loops/. ProDA is free to all users. There is no login requirement. It is available at: http://bioinf.modares.ac.ir/software/linda/. WHAT IF is shareware that is available from https://swift.cmbi.ru.nl/whatif/.

Targeting CD40-Induced TRAF6 Signaling in Macrophages Reduces Atherosclerosis.

BACKGROUND: Disrupting the costimulatory CD40-CD40L dyad reduces atherosclerosis, but can result in immune suppression. The authors recently identified small molecule inhibitors that block the interaction between CD40 and tumor necrosis factor receptor-associated factor (TRAF) 6 (TRAF-STOPs), while leaving CD40-TRAF2/3/5 interactions intact, thereby preserving CD40-mediated immunity. OBJECTIVES: This study evaluates the potential of TRAF-STOP treatment in atherosclerosis. METHODS: The effects of TRAF-STOPs on atherosclerosis were investigated in apolipoprotein E deficient (Apoe-/-) mice. Recombinant high-density lipoprotein (rHDL) nanoparticles were used to target TRAF-STOPs to macrophages. RESULTS: TRAF-STOP treatment of young Apoe-/- mice reduced atherosclerosis by reducing CD40 and integrin expression in classical monocytes, thereby hampering monocyte recruitment. When Apoe-/- mice with established atherosclerosis were treated with TRAF-STOPs, plaque progression was halted, and plaques contained an increase in collagen, developed small necrotic cores, and contained only a few immune cells. TRAF-STOP treatment did not impair "classical" immune pathways of CD40, including T-cell proliferation and costimulation, Ig isotype switching, or germinal center formation, but reduced CD40 and ß2-integrin expression in inflammatory monocytes. In vitro testing and transcriptional profiling showed that TRAF-STOPs are effective in reducing macrophage migration and activation, which could be attributed to reduced phosphorylation of signaling intermediates of the canonical NF-κB pathway. To target TRAF-STOPs specifically to macrophages, TRAF-STOP 6877002 was incorporated into rHDL nanoparticles. Six weeks of rHDL-6877002 treatment attenuated the initiation of atherosclerosis in Apoe-/- mice. CONCLUSIONS: TRAF-STOPs can overcome the current limitations of long-term CD40 inhibition in atherosclerosis and have the potential to become a future therapeutic for atherosclerosis.

3D-e-Chem: Structural Cheminformatics Workflows for Computer-Aided Drug Discovery.

eScience technologies are needed to process the information available in many heterogeneous types of protein-ligand interaction data and to capture these data into models that enable the design of efficacious and safe medicines. Here we present scientific KNIME tools and workflows that enable the integration of chemical, pharmacological, and structural information for: i) structure-based bioactivity data mapping, ii) structure-based identification of scaffold replacement strategies for ligand design, iii) ligand-based target prediction, iv) protein sequence-based binding site identification and ligand repurposing, and v) structure-based pharmacophore comparison for ligand repurposing across protein families. The modular setup of the workflows and the use of well-established standards allows the re-use of these protocols and facilitates the design of customized computer-aided drug discovery workflows.

Aggregation of population-based genetic variation over protein domain homologues and its potential use in genetic diagnostics.

Whole exomes of patients with a genetic disorder are nowadays routinely sequenced but interpretation of the identified genetic variants remains a major challenge. The increased availability of population-based human genetic variation has given rise to measures of genetic tolerance that have been used, for example, to predict disease-causing genes in neurodevelopmental disorders. Here, we investigated whether combining variant information from homologous protein domains can improve variant interpretation. For this purpose, we developed a framework that maps population variation and known pathogenic mutations onto 2,750 "meta-domains." These meta-domains consist of 30,853 homologous Pfam protein domain instances that cover 36% of all human protein coding sequences. We find that genetic tolerance is consistent across protein domain homologues, and that patterns of genetic tolerance faithfully mimic patterns of evolutionary conservation. Furthermore, for a significant fraction (68%) of the meta-domains high-frequency population variation re-occurs at the same positions across domain homologues more often than expected. In addition, we observe that the presence of pathogenic missense variants at an aligned homologous domain position is often paired with the absence of population variation and vice versa. The use of these meta-domains can improve the interpretation of genetic variation.

Molden 2.0: quantum chemistry meets proteins.

Since the first distribution of Molden in 1995 and the publication of the first article about this software in 2000 work on Molden has continued relentlessly. A few of the many improved or fully novel features such as improved and broadened support for quantum chemistry calculations, preparation of ligands for use in drug design related softwares, and working with proteins for the purpose of ligand docking.

Outer membrane phospholipase A's roles in Helicobacter pylori acid adaptation.

BACKGROUND: The pH of the human gastric mucosa varies around 2.5 so that only bacteria with strong acidic stress tolerance can colonize it. The ulcer causing Helicobacter pylori thrives in the gastric mucosa. We analyse the roles of the key outer membrane protein OMPLA in its roles in acid tolerance. RESULTS: The homology model of Helicobacter pylori outer membrane phospholipase A (OMPLA) reveals a twelve stranded ß-barrel with a pore that allows molecules to pass with a diameter up to 4 Å. Structure based multiple sequence alignments revealed the functional roles of many amino acids, and led to the suggestion that OMPLA has multiple functions. Besides its role as phospholipase it lets urea enter and ammonium exit the periplasm. Combined with an extensive literature study, our work leads to a comprehensive model for H. pylori's acid tolerance. This model is based on the conversion of urea into ammonium, and it includes multiple roles for OMPLA and involves two hitherto little studied membrane channels in the OMPLA operon. CONCLUSION: The three-dimensional model of OMPLA predicts a transmembrane pore that can aid H. pylori's acid tolerance through urea influx and ammonium efflux. After urea passes through OMPLA into the periplasm, it passes through the pH-gated inner membrane channel UreI into the cytoplasm where urease hydrolyses it into NH3 and CO2. Most of the NH3 becomes NH4+ that is likely to need an inner membrane channel to reach the periplasm. Two genes that are co-regulated with OMPLA in gastric Helicobacter operons could aid this transport. The NH4+ that might leave the cell through the OMPLA pore has been implicated in H. pylor's pathogenesis.

NewProt - a protein engineering portal.

The NewProt protein engineering portal is a one-stop-shop for in silico protein engineering. It gives access to a large number of servers that compute a wide variety of protein structure characteristics supporting work on the modification of proteins through the introduction of (multiple) point mutations. The results can be inspected through multiple visualizers. The HOPE software is included to indicate mutations with possible undesired side effects. The Hotspot Wizard software is embedded for the design of mutations that modify a proteins' activity, specificity, or stability. The NewProt portal is freely accessible at http://newprot.cmbi.umcn.nl/ and http://newprot.fluidops.net/.

CorNet: Assigning function to networks of co-evolving residues by automated literature mining.

CorNet is a web-based tool for the analysis of co-evolving residue positions in protein super-family sequence alignments. CorNet projects external information such as mutation data extracted from literature on interactively displayed groups of co-evolving residue positions to shed light on the functions associated with these groups and the residues in them. We used CorNet to analyse six enzyme super-families and found that groups of strongly co-evolving residues tend to consist of residues involved in a same function such as activity, specificity, co-factor binding, or enantioselectivity. This finding allows to assign a function to residues for which no data is available yet in the literature. A mutant library was designed to mutate residues observed in a group of co-evolving residues predicted to be involved in enantioselectivity, but for which no literature data is available yet. The resulting set of mutations indeed showed many instances of increased enantioselectivity.

Validation and correction of Zn-CysxHisy complexes.

Many crystal structures in the Protein Data Bank contain zinc ions in a geometrically distorted tetrahedral complex with four Cys and/or His ligands. A method is presented to automatically validate and correct these zinc complexes. Analysis of the corrected zinc complexes shows that the average Zn-Cys distances and Cys-Zn-Cys angles are a function of the number of cysteines and histidines involved. The observed trends can be used to develop more context-sensitive targets for model validation and refinement.

The role of the dopamine D1 receptor in social cognition: studies using a novel genetic rat model.

Social cognition is an endophenotype that is impaired in schizophrenia and several other (comorbid) psychiatric disorders. One of the modulators of social cognition is dopamine, but its role is not clear. The effects of dopamine are mediated through dopamine receptors, including the dopamine D1 receptor (Drd1). Because current Drd1 receptor agonists are not Drd1 selective, pharmacological tools are not sufficient to delineate the role of the Drd1. Here, we describe a novel rat model with a genetic mutation in Drd1 in which we measured basic behavioural phenotypes and social cognition. The I116S mutation was predicted to render the receptor less stable. In line with this computational prediction, this Drd1 mutation led to a decreased transmembrane insertion of Drd1, whereas Drd1 expression, as measured by Drd1 mRNA levels, remained unaffected. Owing to decreased transmembrane Drd1 insertion, the mutant rats displayed normal basic motoric and neurological parameters, as well as locomotor activity and anxiety-like behaviour. However, measures of social cognition like social interaction, scent marking, pup ultrasonic vocalizations and sociability, were strongly reduced in the mutant rats. This profile of the Drd1 mutant rat offers the field of neuroscience a novel genetic rat model to study a series of psychiatric disorders including schizophrenia, autism, depression, bipolar disorder and drug addiction.

In Silico Structure and Sequence Analysis of Bacterial Porins and Specific Diffusion Channels for Hydrophilic Molecules: Conservation, Multimericity and Multifunctionality.

Diffusion channels are involved in the selective uptake of nutrients and form the largest outer membrane protein (OMP) family in Gram-negative bacteria. Differences in pore size and amino acid composition contribute to the specificity. Structure-based multiple sequence alignments shed light on the structure-function relations for all eight subclasses. Entropy-variability analysis results are correlated to known structural and functional aspects, such as structural integrity, multimericity, specificity and biological niche adaptation. The high mutation rate in their surface-exposed loops is likely an important mechanism for host immune system evasion. Multiple sequence alignments for each subclass revealed conserved residue positions that are involved in substrate recognition and specificity. An analysis of monomeric protein channels revealed particular sequence patterns of amino acids that were observed in other classes at multimeric interfaces. This adds to the emerging evidence that all members of the family exist in a multimeric state. Our findings are important for understanding the role of members of this family in a wide range of bacterial processes, including bacterial food uptake, survival and adaptation mechanisms.