Active Learning Approach for Guiding Site-of-Metabolism Measurement and Annotation.

The ability to determine and predict metabolically labile atom positions in a molecule (also called "sites of metabolism" or "SoMs") is of high interest to the design and optimization of bioactive compounds, such as drugs, agrochemicals, and cosmetics. In recent years, several in silico models for SoM prediction have become available, many of which include a machine-learning component. The bottleneck in advancing these approaches is the coverage of distinct atom environments and rare and complex biotransformation events with high-quality experimental data. Pharmaceutical companies typically have measured metabolism data available for several hundred to several thousand compounds. However, even for metabolism experts, interpreting these data and assigning SoMs are challenging and time-consuming. Therefore, a significant proportion of the potential of the existing metabolism data, particularly in machine learning, remains dormant. Here, we report on the development and validation of an active learning approach that identifies the most informative atoms across molecular data sets for SoM annotation. The active learning approach, built on a highly efficient reimplementation of SoM predictor FAME 3, enables experts to prioritize their SoM experimental measurements and annotation efforts on the most rewarding atom environments. We show that this active learning approach yields competitive SoM predictors while requiring the annotation of only 20% of the atom positions required by FAME 3. The source code of the approach presented in this work is publicly available.

High-Quality Conformer Generation with CONFORGE: Algorithm and Performance Assessment.

Knowledge of the putative bound-state conformation of a molecule is an essential prerequisite for the successful application of many computer-aided drug design methods that aim to assess or predict its capability to bind to a particular target receptor. An established approach to predict bioactive conformers in the absence of receptor structure information is to sample the low-energy conformational space of the investigated molecules and derive representative conformer ensembles that can be expected to comprise members closely resembling possible bound-state ligand conformations. The high relevance of such conformer generation functionality led to the development of a wide panel of dedicated commercial and open-source software tools throughout the last decades. Several published benchmarking studies have shown that open-source tools usually lag behind their commercial competitors in many key aspects. In this work, we introduce the open-source conformer ensemble generator CONFORGE, which aims at delivering state-of-the-art performance for all types of organic molecules in drug-like chemical space. The ability of CONFORGE and several well-known commercial and open-source conformer ensemble generators to reproduce experimental 3D structures as well as their computational efficiency and robustness has been assessed thoroughly for both typical drug-like molecules and macrocyclic structures. For small molecules, CONFORGE clearly outperformed all other tested open-source conformer generators and performed at least equally well as the evaluated commercial generators in terms of both processing speed and accuracy. In the case of macrocyclic structures, CONFORGE achieved the best average accuracy among all benchmarked generators, with RDKit's generator coming close in second place.

Apo2ph4: A Versatile Workflow for the Generation of Receptor-based Pharmacophore Models for Virtual Screening.

Pharmacophore models are widely used as efficient virtual screening (VS) filters for the target-directed enrichment of large compound libraries. However, the generation of pharmacophore models that have the power to discriminate between active and inactive molecules traditionally requires structural information about ligand-target complexes or at the very least knowledge of one active ligand. The fact that the discovery of the first known active ligand of a newly investigated target represents a major hurdle at the beginning of every drug discovery project underscores the need for methods that are able to derive high-quality pharmacophore models even without the prior knowledge of any active ligand structures. In this work, we introduce a novel workflow, called apo2ph4, that enables the rapid derivation of pharmacophore models solely from the three-dimensional structure of the target receptor. The utility of this workflow is demonstrated retrospectively for the generation of a pharmacophore model for the M2 muscarinic acetylcholine receptor. Furthermore, in order to show the general applicability of apo2ph4, the workflow was employed for all 15 targets of the recently published LIT-PCBA dataset. Pharmacophore-based VS runs using the apo2ph4-derived models achieved a significant enrichment of actives for 13 targets. In the last presented example, a pharmacophore model derived from the etomidate site of the α1ß2γ2 GABAA receptor was used in VS campaigns. Subsequent in vitro testing of selected hits revealed that 19 out of 20 (95%) tested compounds were able to significantly enhance GABA currents, which impressively demonstrates the applicability of apo2ph4 for real-world drug design projects.

GABAA receptor-mediated seizure liabilities: a mixed-methods screening approach.

GABAA receptors, members of the pentameric ligand-gated ion channel superfamily, are widely expressed in the central nervous system and mediate a broad range of pharmaco-toxicological effects including bidirectional changes to seizure threshold. Thus, detection of GABAA receptor-mediated seizure liabilities is a big, partly unmet need in early preclinical drug development. This is in part due to the plethora of allosteric binding sites that are present on different subtypes of GABAA receptors and the critical lack of screening methods that detect interactions with any of these sites. To improve in silico screening methods, we assembled an inventory of allosteric binding sites based on structural data. Pharmacophore models representing several of the binding sites were constructed. These models from the NeuroDeRisk IL Profiler were used for in silico screening of a compiled collection of drugs with known GABAA receptor interactions to generate testable hypotheses. Amoxapine was one of the hits identified and subjected to an array of in vitro assays to examine molecular and cellular effects on neuronal excitability and in vivo locomotor pattern changes in zebrafish larvae. An additional level of analysis for our compound collection is provided by pharmacovigilance alerts using FAERS data. Inspired by the Adverse Outcome Pathway framework, we postulate several candidate pathways leading from specific binding sites to acute seizure induction. The whole workflow can be utilized for any compound collection and should inform about GABAA receptor-mediated seizure risks more comprehensively compared to standard displacement screens, as it rests chiefly on functional data.

Temporal localized states and square-waves in semiconductor micro-resonators with strong time-delayed feedback.

In this paper, we study the dynamics of a vertically emitting micro-cavity operated in the Gires-Tournois regime that contains a semiconductor quantum-well and that is subjected to strong time-delayed optical feedback and detuned optical injection. Using a first principle time-delay model for the optical response, we disclose sets of multistable dark and bright temporal localized states coexisting on their respective bistable homogeneous backgrounds. In the case of anti-resonant optical feedback, we identify square-waves with a period of twice the round-trip in the external cavity. Finally, we perform a multiple time scale analysis in the good cavity limit. The resulting normal form is in good agreement with the original time-delayed model.

Ryanodine Receptor Staining Identifies Viable Cardiomyocytes in Human and Rabbit Cardiac Tissue Slices.

In terms of preserving multicellularity and myocardial function in vitro, the cultivation of beating myocardial slices is an emerging technique in basic and translational cardiac research. It can be used, for example, for drug screening or to study pathomechanisms. Here, we describe staining for viable cardiomyocytes based on the immunofluorescence of ryanodine receptors (RyRs) in human and rabbit myocardial slices. Biomimetic chambers were used for culture and measurements of contractile force. Fixable fluorophore-conjugated dextran, entering cells with a permeable membrane, was used for death staining. RyRs, nuclei and the extracellular matrix, including the t-system, were additionally stained and analyzed by confocal microscopy and image processing. We found the mutual exclusion of the RyR and dextran signals in cultivated slices. T-System density and nucleus size were reduced in RyR-negative/dextran-positive myocytes. The fraction of RyR-positive myocytes and pixels correlated with the contractile force. In RyR-positive/dextran-positive myocytes, we found irregular RyR clusters and SERCA distribution patterns, confirmed by an altered power spectrum. We conclude that RyR immunofluorescence indicates viable cardiomyocytes in vibratome-cut myocardial slices, facilitating the detection and differential structural analysis of living vs. dead or dying myocytes. We suggest the loss of sarcoplasmic reticulum integrity as an early event during cardiomyocyte death.

A normal form for frequency combs and localized states in Kerr-Gires-Tournois interferometers.

We elucidate the mechanisms that underlay the formation of temporal localized states and frequency combs in vertical external-cavity Kerr-Gires-Tournois interferometers. We reduce our first-principles model based upon delay algebraic equations to a minimal pattern formation scenario. It consists in a real cubic Ginzburg-Landau equation modified by high-order effects such as third-order dispersion and nonlinear drift, which are responsible for generating localized states via the locking of domain walls connecting the high and low intensity levels of the injected micro-cavity. We interpret the effective parameters of the normal form in relation with the configuration of the optical setup. Comparing the two models, we observe an excellent agreement close to the onset of bistability.

Square-wave generation in vertical external-cavity Kerr-Gires-Tournois interferometers.

We study theoretically the mechanisms of square-wave (SW) formation in vertical external-cavity Kerr-Gires-Tournois interferometers in the presence of anti-resonant injection. We provide simple analytical approximations for their plateau intensities and for the conditions of their emergence. We demonstrate that SWs may appear via a homoclinic snaking scenario, leading to the formation of complex-shaped multistable SW solutions. The resulting SWs can host localized structures and robust bound states.

Influence of time-delayed feedback on the dynamics of temporal localized structures in passively mode-locked semiconductor lasers.

In this paper, we analyze the effect of optical feedback on the dynamics of a passively mode-locked ring laser operating in the regime of temporal localized structures. This laser system is modeled by a set of delay differential equations, which include delay terms associated with the laser cavity and the feedback loop. Using a combination of direct numerical simulations and path-continuation techniques, we show that the feedback loop creates echoes of the main pulse whose position and size strongly depend on the feedback parameters. We demonstrate that in the long-cavity regime, these echoes can successively replace the main pulses, which defines their lifetime. This pulse instability mechanism originates from a global bifurcation of the saddle-node infinite-period type. In addition, we show that, under the influence of noise, the stable pulses exhibit forms of a behavior characteristic of excitable systems. Furthermore, for the harmonic solutions consisting of multiple equispaced pulses per round-trip, we show that if the location of the pulses coincides with the echo of another, the range of stability of these solutions is increased. Finally, it is shown that around these resonances, branches of different solutions are connected by period-doubling bifurcations.

In Silico Identification of Potential Druggable Binding Sites on CIN85 SH3 Domain.

Protein-protein interactions (PPIs) play a pivotal role in the regulation of many physiological processes. The dysfunction of some PPIs interactions led to the alteration of different biological pathways causing various diseases including cancer. In this context, the inhibition of PPIs represents an attractive strategy for the design of new antitumoral agents. In recent years, computational approaches were successfully used to study the interactions between proteins, providing useful hints for the design of small molecules able to modulate PPIs. Targeting PPIs presents several challenges mainly due to the large and flat binding surface that lack the typical binding pockets of traditional drug targets. Despite these hurdles, substantial progress has been made in the last decade resulting in the identification of PPI modulators where some of them even found clinical use. This study focuses on MUC1-CIN85 PPI which is involved in the migration and invasion of cancer cells. Particularly, we investigated the presence of druggable binding sites on the CIN85 surface which provided new insights for the structure-based design of novel MUC1-CIN85 PPI inhibitors as anti-metastatic agents.

Greedy 3-Point Search (G3PS)-A Novel Algorithm for Pharmacophore Alignment.

Chemical features of small molecules can be abstracted to 3D pharmacophore models, which are easy to generate, interpret, and adapt by medicinal chemists. Three-dimensional pharmacophores can be used to efficiently match and align molecules according to their chemical feature pattern, which facilitates the virtual screening of even large compound databases. Existing alignment methods, used in computational drug discovery and bio-activity prediction, are often not suitable for finding matches between pharmacophores accurately as they purely aim to minimize RMSD or maximize volume overlap, when the actual goal is to match as many features as possible within the positional tolerances of the pharmacophore features. As a consequence, the obtained alignment results are often suboptimal in terms of the number of geometrically matched feature pairs, which increases the false-negative rate, thus negatively affecting the outcome of virtual screening experiments. We addressed this issue by introducing a new alignment algorithm, Greedy 3-Point Search (G3PS), which aims at finding optimal alignments by using a matching-feature-pair maximizing search strategy while at the same time being faster than competing methods.

Improved Lipophilicity and Aqueous Solubility Prediction with Composite Graph Neural Networks.

The accurate prediction of molecular properties, such as lipophilicity and aqueous solubility, are of great importance and pose challenges in several stages of the drug discovery pipeline. Machine learning methods, such as graph-based neural networks (GNNs), have shown exceptionally good performance in predicting these properties. In this work, we introduce a novel GNN architecture, called directed edge graph isomorphism network (D-GIN). It is composed of two distinct sub-architectures (D-MPNN, GIN) and achieves an improvement in accuracy over its sub-architectures employing various learning, and featurization strategies. We argue that combining models with different key aspects help make graph neural networks deeper and simultaneously increase their predictive power. Furthermore, we address current limitations in assessment of deep-learning models, namely, comparison of single training run performance metrics, and offer a more robust solution.

Location and function of transient receptor potential canonical channel 1 in ventricular myocytes.

Transient receptor potential canonical 1 (TRPC1) protein is abundantly expressed in cardiomyocytes. While TRPC1 is supposed to be critically involved in cardiac hypertrophy, its physiological role in cardiomyocytes is poorly understood. We investigated the subcellular location of TRPC1 and its contribution to Ca2+ signaling in mammalian ventricular myocytes. Immunolabeling, three-dimensional scanning confocal microscopy and quantitative colocalization analysis revealed an abundant intracellular location of TRPC1 in neonatal rat ventricular myocytes (NRVMs) and adult rabbit ventricular myocytes. TRPC1 was colocalized with intracellular proteins including sarco/endoplasmic reticulum Ca2+ ATPase 2 in the sarcoplasmic reticulum (SR). Colocalization with wheat germ agglutinin, which labels the glycocalyx and thus marks the sarcolemma including the transverse tubular system, was low. Super-resolution and immunoelectron microscopy supported the intracellular location of TRPC1. We investigated Ca2+ signaling in NRVMs after adenoviral TRPC1 overexpression or silencing. In NRVMs bathed in Na+ and Ca2+ free solution, TRPC1 overexpression and silencing was associated with a decreased and increased SR Ca2+ content, respectively. In isolated rabbit cardiomyocytes bathed in Na+ and Ca2+ free solution, we found an increased decay of the cytosolic Ca2+ concentration [Ca2+]i and increased SR Ca2+ content in the presence of the TRPC channel blocker SKF-96365. In a computational model of rabbit ventricular myocytes at physiological pacing rates, Ca2+ leak through SR TRPC channels increased the systolic and diastolic [Ca2+]i with only minor effects on the action potential and SR Ca2+ content. Our studies suggest that TRPC1 channels are localized in the SR, and not present in the sarcolemma of ventricular myocytes. The studies provide evidence for a role of TRPC1 as a contributor to SR Ca2+ leak in cardiomyocytes, which was previously explained by ryanodine receptors only. We propose that the findings will guide us to an understanding of TRPC1 channels as modulators of [Ca2+]i and contractility in cardiomyocytes.

A compact review of molecular property prediction with graph neural networks.

As graph neural networks are becoming more and more powerful and useful in the field of drug discovery, many pharmaceutical companies are getting interested in utilizing these methods for their own in-house frameworks. This is especially compelling for tasks such as the prediction of molecular properties which is often one of the most crucial tasks in computer-aided drug discovery workflows. The immense hype surrounding these kinds of algorithms has led to the development of many different types of promising architectures and in this review we try to structure this highly dynamic field of AI-research by collecting and classifying 80 GNNs that have been used to predict more than 20 molecular properties using 48 different datasets.

Discrete light bullets in passively mode-locked semiconductor lasers.

In this paper, we analyze the formation and dynamical properties of discrete light bullets in an array of passively mode-locked lasers coupled via evanescent fields in a ring geometry. Using a generic model based upon a system of nearest-neighbor coupled Haus master equations, we show numerically the existence of discrete light bullets for different coupling strengths. In order to reduce the complexity of the analysis, we approximate the full problem by a reduced set of discrete equations governing the dynamics of the transverse profile of the discrete light bullets. This effective theory allows us to perform a detailed bifurcation analysis via path-continuation methods. In particular, we show the existence of multistable branches of discrete localized states, corresponding to different number of active elements in the array. These branches are either independent of each other or are organized into a snaking bifurcation diagram where the width of the discrete localized states grows via a process of successive increase and decrease of the gain. Mechanisms are revealed by which the snaking branches can be created and destroyed as a second parameter, e.g., the linewidth enhancement factor or the coupling strength is varied. For increasing couplings, the existence of moving bright and dark discrete localized states is also demonstrated.

Glucocorticoids preserve the t-tubular system in ventricular cardiomyocytes by upregulation of autophagic flux.

A major contributor to contractile dysfunction in heart failure is remodelling and loss of the cardiomyocyte transverse tubular system (t-system), but underlying mechanisms and signalling pathways remain elusive. It has been shown that dexamethasone promotes t-tubule development in stem cell-derived cardiomyocytes and that cardiomyocyte-specific glucocorticoid receptor (GR) knockout (GRKO) leads to heart failure. Here, we studied if the t-system is altered in GRKO hearts and if GR signalling is required for t-system preservation in adult cardiomyocytes. Confocal and 3D STED microscopy of myocardium from cardiomyocyte-specific GRKO mice revealed decreased t-system density and increased distances between ryanodine receptors (RyR) and L-type Ca2+ channels (LTCC). Because t-system remodelling and heart failure are intertwined, we investigated the underlying mechanisms in vitro. Ventricular cardiomyocytes from failing human and healthy adult rat hearts cultured in the absence of glucocorticoids (CTRL) showed distinctively lower t-system density than cells treated with dexamethasone (EC50 1.1 nM) or corticosterone. The GR antagonist mifepristone abrogated the effect of dexamethasone. Dexamethasone improved RyR-LTCC coupling and synchrony of intracellular Ca2+ release, but did not alter expression levels of t-system-associated proteins junctophilin-2 (JPH2), bridging integrator-1 (BIN1) or caveolin-3 (CAV3). Rather, dexamethasone upregulated LC3B and increased autophagic flux. The broad-spectrum protein kinase inhibitor staurosporine prevented dexamethasone-induced upregulation of autophagy and t-system preservation, and autophagy inhibitors bafilomycin A and chloroquine accelerated t-system loss. Conversely, induction of autophagy by rapamycin or amino acid starvation preserved the t-system. These findings suggest that GR signalling and autophagy are critically involved in t-system preservation and remodelling in the heart.

Physiological in vitro sacroiliac joint motion: a study on three-dimensional posterior pelvic ring kinematics.

The sacroiliac joint (SIJ) is a well-known source of low back and pelvic pain, of increasing interest for both conservative and surgical treatment. Alterations in the kinematics of the pelvis have been hypothesized as a major cause of SIJ-related pain. However, definitions of both the range and the extent of physiological movement are controversial, and there are no clear baseline data for pathological alterations. The present study combined a novel biomechanical setup allowing for physiological motion of the lumbosacral transition and pelvis without restricting the SIJ movement in vitro, combined with optical image correlation. Six fresh human pelvises (81 ± 10 years, three females, three males) were tested, with bodyweight-adapted loading applied to the fifth lumbar vertebra and both acetabula. Deformation at the lumbopelvises was determined computationally from three-dimensional image correlation data. Sacroiliac joint motion under the loading of 100% bodyweight primarily consisted of a z-axis rotation (0.16°) and an inferior translation of the sacrum relative to the ilium (0.32 mm). Sacroiliac joint flexion-extension rotations were minute (< 0.02°). Corresponding movements of the SIJ were found at the lumbosacral transition, with an anterior translation of L5 relative to the sacrum of -0.97 mm and an inferior translation of 0.11 mm, respectively. Moreover, a flexion of 1.82° was observed at the lumbosacral transition. Within the innominate bone and at the pubic symphysis, small complementary rotations were seen around a vertical axis, accounting for -0.10° and 0.11°, respectively. Other motions were minute and accompanied by large interindividual variation. The present study provides evidence of different SIJ motions than reported previously when exerted by physiological loading. Sacroiliac joint kinematics were in the sub-degree and sub-millimeter range, in line with previous in vivo and in vitro findings, largely limited to the sagittal rotation and an inferior translation of the sacrum relative to the ilium. This given physiological loading scenario underlines the relevance of the lumbosacral transition when considering the overall motion of the lumbopelvis, and how relatively little the other segments contribute to overall motion.

LigandScout Remote: A New User-Friendly Interface for HPC and Cloud Resources.

High-performance computing (HPC) clusters play a major role in scientific research. However, working with these clusters is often cumbersome, especially for researchers without a formal background in computer science. It requires preparation and transfer of the input data, manual gathering of results, and command-line expertise. Current approaches for improving accessibility to remote HPC clusters are focused on providing web-based graphical front-ends that allow jobs to be submitted to the distributed resource management system running on the cluster. This comes with significant usability benefits over command-line usage but does not circumvent the need for manual handling of the input and output files. With LigandScout Remote, we propose a different solution. Our software enables the seamless integration of HPC resources into the LigandScout desktop application that scientists use also in their day-to-day work. By handling necessary data conversion and network communication transparently to the user, this approach completely evades any HPC usability barriers. We show that the developed software combines the usability of local graphical desktop applications with the performance of HPC clusters.

Sheet-Like Remodeling of the Transverse Tubular System in Human Heart Failure Impairs Excitation-Contraction Coupling and Functional Recovery by Mechanical Unloading.

BACKGROUND: Cardiac recovery in response to mechanical unloading by left ventricular assist devices (LVADs) has been demonstrated in subgroups of patients with chronic heart failure (HF). Hallmarks of HF are depletion and disorganization of the transverse tubular system (t-system) in cardiomyocytes. Here, we investigated remodeling of the t-system in human end-stage HF and its role in cardiac recovery. METHODS: Left ventricular biopsies were obtained from 5 donors and 26 patients with chronic HF undergoing implantation of LVADs. Three-dimensional confocal microscopy and computational image analysis were applied to assess t-system structure, density, and distance of ryanodine receptor clusters to the sarcolemma, including the t-system. Recovery of cardiac function in response to mechanical unloading was assessed by echocardiography during turndown of the LVAD. RESULTS: The majority of HF myocytes showed remarkable t-system remodeling, particularly sheet-like invaginations of the sarcolemma. Circularity of t-system components was decreased in HF versus controls (0.37±0.01 versus 0.46±0.02; P<0.01), and the volume/length ratio was increased in HF (0.36±0.01 versus 0.25±0.02 µm2; P<0.0001). T-system density was reduced in HF, leading to increased ryanodine receptor-sarcolemma distances (0.96±0.05 versus 0.64±0.1 µm; P<0.01). Low ryanodine receptor-sarcolemma distances at the time of LVAD implantation predicted high post-LVAD left ventricular ejection fractions (P<0.01) and ejection fraction increases during unloading (P<0.01). Ejection fraction in patients with pre-LVAD ryanodine receptor-sarcolemma distances >1 µm did not improve after mechanical unloading. In addition, calcium transients were recorded in field-stimulated isolated human cardiomyocytes and analyzed with respect to local t-system density. Calcium release in HF myocytes was restricted to regions proximal to the sarcolemma. Local calcium upstroke was delayed (23.9±4.9 versus 10.3±1.7 milliseconds; P<0.05) and more asynchronous (18.1±1.5 versus 8.9±2.2 milliseconds; P<0.01) in HF cells with low t-system density versus cells with high t-system density. CONCLUSIONS: The t-system in end-stage human HF presents a characteristic novel phenotype consisting of sheet-like invaginations of the sarcolemma. Our results suggest that the remodeled t-system impairs excitation-contraction coupling and functional recovery during chronic LVAD unloading. An intact t-system at the time of LVAD implantation may constitute a precondition and predictor for functional cardiac recovery after mechanical unloading.

Discovery of Potent Inhibitors for the Large Neutral Amino Acid Transporter 1 (LAT1) by Structure-Based Methods.

The large neutral amino acid transporter 1 (LAT1) is a promising anticancer target that is required for the cellular uptake of essential amino acids that serve as building blocks for cancer growth and proliferation. Here, we report a structure-based approach to identify chemically diverse and potent inhibitors of LAT1. First, a homology model of LAT1 that is based on the atomic structures of the prokaryotic homologs was constructed. Molecular docking of nitrogen mustards (NMs) with a wide range of affinity allowed for deriving a common binding mode that could explain the structure-activity relationship pattern in NMs. Subsequently, validated binding hypotheses were subjected to molecular dynamics simulation, which allowed for extracting a set of dynamic pharmacophores. Finally, a library of ~1.1 million molecules was virtually screened against these pharmacophores, followed by docking. Biological testing of the 30 top-ranked hits revealed 13 actives, with the best compound showing an IC50 value in the sub-µM range.