Thompson Sampling─An Efficient Method for Searching Ultralarge Synthesis on Demand Databases.

Over the last five years, virtual screening of ultralarge synthesis on-demand libraries has emerged as a powerful tool for hit identification in drug discovery programs. As these libraries have grown to tens of billions of molecules, we have reached a point where it is no longer cost-effective to screen every molecule virtually. To address these challenges, several groups have developed heuristic search methods to rapidly identify the best molecules on a virtual screen. This article describes the application of Thompson sampling (TS), an active learning approach that streamlines the virtual screening of large combinatorial libraries by performing a probabilistic search in the reagent space, thereby never requiring the full enumeration of the library. TS is a general technique that can be applied to various virtual screening modalities, including 2D and 3D similarity search, docking, and application of machine-learning models. In an illustrative example, we show that TS can identify more than half of the top 100 molecules from a docking-based virtual screen of 335 million molecules by evaluating 1% of the data set.

Discovery of complex oxides via automated experiments and data science.

The quest to identify materials with tailored properties is increasingly expanding into high-order composition spaces, with a corresponding combinatorial explosion in the number of candidate materials. A key challenge is to discover regions in composition space where materials have novel properties. Traditional predictive models for material properties are not accurate enough to guide the search. Herein, we use high-throughput measurements of optical properties to identify novel regions in three-cation metal oxide composition spaces by identifying compositions whose optical trends cannot be explained by simple phase mixtures. We screen 376,752 distinct compositions from 108 three-cation oxide systems based on the cation elements Mg, Fe, Co, Ni, Cu, Y, In, Sn, Ce, and Ta. Data models for candidate phase diagrams and three-cation compositions with emergent optical properties guide the discovery of materials with complex phase-dependent properties, as demonstrated by the discovery of a Co-Ta-Sn substitutional alloy oxide with tunable transparency, catalytic activity, and stability in strong acid electrolytes. These results required close coupling of data validation to experiment design to generate a reliable end-to-end high-throughput workflow for accelerating scientific discovery.

Kohn-Sham Equations as Regularizer: Building Prior Knowledge into Machine-Learned Physics.

Including prior knowledge is important for effective machine learning models in physics and is usually achieved by explicitly adding loss terms or constraints on model architectures. Prior knowledge embedded in the physics computation itself rarely draws attention. We show that solving the Kohn-Sham equations when training neural networks for the exchange-correlation functional provides an implicit regularization that greatly improves generalization. Two separations suffice for learning the entire one-dimensional H_{2} dissociation curve within chemical accuracy, including the strongly correlated region. Our models also generalize to unseen types of molecules and overcome self-interaction error.

Catcher's Knee: Posterior Femoral Condyle Juvenile Osteochondritis Dissecans in Children and Adolescents.

BACKGROUND: Juvenile osteochondritis dissecans is an idiopathic condition involving subchondral bone and articular cartilage in skeletally immature patients in whom the growth plates are open, potentially leading to lesion instability. Because of the differing forces experienced by baseball/softball catchers versus position players, the age at which lesions develop and the characteristics of the lesions themselves may differ between these 2 populations. The purpose of the study was to examine relative age and characteristics of osteochondritis dissecans (OCD) knee lesions in catchers compared with position players. METHODS: Using a text-based search tool that queries clinic notes and operative reports, computerized medical records from 1990 to 2014 from the Sports Medicine Program of a tertiary care Children's Hospital were searched to find children and adolescents who had OCD of the knee, played baseball/softball, had a specified field position, and had magnetic resonance imaging of the knee. Ultimately, 98 knees (78 patients) were identified: 33 knees (29 patients) in catchers and 65 knees (49 patients) in noncatchers. Data collected included position played (catcher/noncatcher), demographics (age, unilateral/bilateral, and sex), lesion severity, and sagittal and coronal lesion location. RESULTS: When compared with noncatchers, catchers presented at a younger age (P=0.035) but were similar with respect to bilateral involvement (P=0.115), sex (P=0.457), and lesion severity (P=0.484). Lesions in catchers were more posterior on the femoral condyle in the sagittal plane (P=0.004) but similar in location in the coronal plane (P=0.210). CONCLUSIONS: Catchers developed OCD at a younger age and in a more posterior location on the medial and lateral femoral condyles than noncatchers. These results may represent the effects of repetitive and persistent loading of the knees in the hyperflexed position required of catchers. Increased awareness of this risk may lead to surveillance and prevention programs. LEVEL OF EVIDENCE: Level III-case-control study.

Molecular graph convolutions: moving beyond fingerprints.

Molecular "fingerprints" encoding structural information are the workhorse of cheminformatics and machine learning in drug discovery applications. However, fingerprint representations necessarily emphasize particular aspects of the molecular structure while ignoring others, rather than allowing the model to make data-driven decisions. We describe molecular graph convolutions, a machine learning architecture for learning from undirected graphs, specifically small molecules. Graph convolutions use a simple encoding of the molecular graph-atoms, bonds, distances, etc.-which allows the model to take greater advantage of information in the graph structure. Although graph convolutions do not outperform all fingerprint-based methods, they (along with other graph-based methods) represent a new paradigm in ligand-based virtual screening with exciting opportunities for future improvement.

Matrix Rigidity Mediates TGFß1-Induced Epithelial-Myofibroblast Transition by Controlling Cytoskeletal Organization and MRTF-A Localization.

Myofibroblasts mediate normal wound healing and upon chronic activation can contribute to the development of pathological conditions including organ fibrosis and cancer. Myofibroblasts can develop from epithelial cells through an epithelial-mesenchymal transition (EMT) during which epithelial cells exhibit drastic morphological changes and upregulate cytoskeletal associated proteins that enable exertion of large contractile forces and remodeling of the surrounding microenvironment. Increased matrix rigidity is a hallmark of fibrosis and tumor progression and mechanical tension has been identified as a regulator of EMT; however, the mechanisms governing the mechanical regulation of EMT are not completely understood. Here, we find that matrix rigidity regulates transforming growth factor (TGF)-ß1-induced EMT, with rigid substrata enabling increased myofibroblast marker expression, cell morphology changes, and cytoskeletal reorganization while soft matrices block these changes. Furthermore, we find that matrix rigidity controls the subcellular localization of myocardin related transcription factor (MRTF)-A, a regulator of cytoskeletal protein expression that contributes to the acquisition of myogenic features during EMT. Results from these studies provide insight into how biophysical cues contribute to myofibroblast development from epithelial cells and may suggest ways to enhance wound healing or to engineer therapeutic solutions for fibrosis and cancer.

Oxidative calcium release from catechol.

Oxidation of 4-methylcatechol previously exposed to aqueous calcium chloride was shown by ion chromatography to be associated with release of calcium ions. The catechol was oxidised to the corresponding orthoquinone by the use of tyrosinase from Agaricus bisporus. The oxidative release of calcium from the catechol is ascribed to the diminution of the available hydroxyl functions able to act as chelating groups. Our results suggest that the redox status of melanin may regulate calcium binding and influence calcium levels in pigmented cells.

Mechanistic aspects of the tyrosinase oxidation of hydroquinone.

Contradictory reports on the behaviour of hydroquinone as a tyrosinase substrate are reconciled in terms of the ability of the initially formed ortho-quinone to tautomerise to the thermodynamically more stable para-quinone isomer. Oxidation of phenols by native tyrosinase requires activation by in situ formation of a catechol formed via an enzyme generated ortho-quinone. In the special case of hydroquinone, catechol formation is precluded by rapid tautomerisation of the ortho-quinone precursor to catechol formation.

Tyrosinase: the four oxidation states of the active site and their relevance to enzymatic activation, oxidation and inactivation.

Tyrosinase is an enzyme widely distributed in the biosphere. It is one of a group of proteins with a strongly conserved bicopper active centre able to bind molecular oxygen. Tyrosinase manifests two catalytic properties; monooxygenase and oxidase activity. These actions reflect the oxidation states of the active centre. Tyrosinase has four possible oxidation states and the details of their interaction are shown to give rise to the unusual kinetic behaviour of the enzyme. The resting state of the enzyme is met-tyrosinase [Cu(II)2] and activation, associated with a 'lag period', involves reduction to deoxy-tyrosinase [Cu(I)2] which is capable of binding dioxygen to form oxy-tyrosinase [Cu(II)2·O2]. Initially the conversion of met- to deoxy-tyrosinase is brought about by a catechol that is indirectly formed from an ortho-quinone product of tyrosinase action. The primary function of the enzyme is monooxygenation of phenols to ortho-quinones by oxy-tyrosinase. Inactivation of the enzyme results from monooxygenase processing of catechols which can lead to reductive elimination of one of the active-site copper ions and conversion of oxy-tyrosinase to the inactive deact-tyrosinase [Cu(II)Cu(0)]. This review describes the tyrosinase pathways and the role of each oxidation state in the enzyme's oxidative transformations of phenols and catechols.

Using machine learning for chemical-free histological tissue staining.

Hematoxylin and eosin staining can be hazardous, expensive, and prone to error and variability. To circumvent these issues, artificial intelligence/machine learning models such as generative adversarial networks (GANs), are being used to 'virtually' stain unstained tissue images indistinguishable from chemically stained tissue. Frameworks such as deep convolutional GANs (DCGAN) and conditional GANs (CGANs) have successfully generated highly reproducible 'stained' images. However, their utility may be limited by requiring registered, paired images which can be difficult to obtain. To avoid these dataset requirements, we attempted to use an unsupervised CycleGAN pix2pix model(5,6) to turn unpaired, unstained bright-field images into pathologist-approved digitally 'stained' images. Using formalin-fixed-paraffin-embedded liver samples, 5µm section images (20x) were obtained before and after staining to create "stained" an "unstained" datasets. Model implementation was conducted using Ubuntu 20.04.4 LTS, 32 GB RAM, Intel Core i7-9750 CPU @2.6 GHz, Nvidia GeForce RTX 2070 Mobile, Python 3.7.11 and Tensorflow 2.9.1. The CycleGAN framework utilized a u-net-based generator and discriminator from pix2pix, a CGAN. The CycleGAN used a modified loss function, cycle consistent loss that assumed unpaired images, so loss was measured twice. To our knowledge, this is the first documented application of this architecture using unpaired bright-field images. Results and suggested improvements are discussed.

Mechanistic studies of the inactivation of tyrosinase by resorcinol.

The inactivation of tyrosinase by resorcinol (1,3-dihydroxybenzene) and seventeen simple derivatives has been investigated using combined spectrophotometry and oximetry together with hplc/ms examination of the oxidation products. The results are consistent with a Quintox mechanism, analogous to that proposed for catechol inactivation of tyrosinase, in which the resorcinol substrate is oxidised via the monooxygenase route leading to a hydroxy intermediate that undergoes deprotonation and results in irreversible elimination of Cu(0) from the active site. Hplc/ms evidence for formation of the resorcinol monooxygenase product (3-hydroxy-ortho-quinone) is presented and the relationship between the ring position of simple resorcinol substituents (H, Me, F, Cl) and tyrosinase inactivation is rationalised.

A prospective comparison of computer-navigated and fluoroscopic-guided in situ fixation of slipped capital femoral epiphysis.

BACKGROUND: Slipped capital femoral epiphysis (SCFE) is usually treated with percutaneous in situ screw fixation to prevent further progression of deformity. The purpose of this investigation is to compare computer navigation (CN) techniques with traditional fluoroscopic (fluoro) techniques for in situ fixation of SCFE. METHODS: This study was an IRB-approved prospective study of 39 hip pinnings in 33 children. CN techniques were used in 22 cases and fluoro in 17. The CN and fluoro groups were statistically similar in terms of grade and acuity of the slip. Children were assigned to the groups based on the intraoperative imaging technique used by the attending on call, with 3 surgeons in each group taking equal amounts of call. The "approach-withdraw" technique was used in all cases. Postoperative limited-cut, reduced-dose computed tomography (CT) scans were obtained to evaluate screw placement. This included blinded analysis for screw penetration of the joint, screw tip-to-apex distance, the distance the screw passed to the center of the physis, and attainment of center-center position. The number of pin passes, intraoperative radiation exposure, and operating room (OR) time were also analyzed. Statistics used included ANOVA, the χ and median tests. RESULTS: Compared with the fluoro group, CN resulted in more accurate screw placement. There was 1 case of joint penetration in the fluoro group not appreciated intraoperatively but detected on postoperative CT. CN also resulted in statistically significant (P < 0.05) reduced screw tip-to-apex distance and distance to the center of the physis. There was no statistically significant difference between the 2 groups in attainment of the center-center position, number of pin passes, or intraoperative radiation exposure. OR time averaged 19 minutes longer in the CN group. There was no case of avascular necrosis or chondrolysis in either of the groups. CONCLUSIONS: Compared with traditional fluoro techniques, CN in situ fixation of SCFE results in more accurate screw placement, comparable number of pin passes and intraoperative radiation exposure, and increased OR time. The cost-benefit ratio of this technology requires careful consideration at each individual institution. LEVEL OF EVIDENCE: II.

Colorimetric Gas Detection Tubes: Limits of Detection and Evaluation Using Active Chemical Warfare Agents.

Chemical weapons continue to be an ongoing threat that necessitates the improvement of existing detection technologies where new technologies are absent. Lower limits of detection will facilitate early warning of exposure to chemical weapons and enable more rapid deployment of countermeasures. Here, we evaluate two colorimetric gas detection tubes, developed by Draeger Inc., for sarin and sulfur mustard chemical warfare agents and determine their limits of detection using active chemical agent. Being that commercial companies are only able to use chemical agent simulants during sensor development, it is imperative to determine limits of detection using active agent. The limit of detection was determined based on the absence of a reasonably perceptible color response at incrementally lower concentrations. A chemical vapor generator was constructed to produce stable and quantifiable concentrations of chemical agent vapor, with the presence of chemical agent verified and monitored by a secondary detector. The limits of detection of the colorimetric gas detection tubes were determined to be 0.0046 ± 0.0002 and 2.1 ± 0.3 mg/m3 for sarin and sulfur mustard, respectively. The response of the sarin detection tube was readily observable with little issue. The sulfur mustard detection tube exhibited a weaker response to active agent compared to the simulant that was used during development, which will affect their concept of operations in real-world detection scenarios.

The influence of hydroquinone on tyrosinase kinetics.

In vitro studies, using combined spectrophotometry and oximetry together with hplc/ms examination of the products of tyrosinase action demonstrate that hydroquinone is not a primary substrate for the enzyme but is vicariously oxidised by a redox exchange mechanism in the presence of either catechol, L-3,4-dihydroxyphenylalanine or 4-ethylphenol. Secondary addition products formed in the presence of hydroquinone are shown to stimulate, rather than inhibit, the kinetics of substrate oxidation.

Evolving symbolic density functionals.

Systematic development of accurate density functionals has been a decades-long challenge for scientists. Despite emerging applications of machine learning (ML) in approximating functionals, the resulting ML functionals usually contain more than tens of thousands of parameters, leading to a huge gap in the formulation with the conventional human-designed symbolic functionals. We propose a new framework, Symbolic Functional Evolutionary Search (SyFES), that automatically constructs accurate functionals in the symbolic form, which is more explainable to humans, cheaper to evaluate, and easier to integrate to existing codes than other ML functionals. We first show that, without prior knowledge, SyFES reconstructed a known functional from scratch. We then demonstrate that evolving from an existing functional ωB97M-V, SyFES found a new functional, GAS22 (Google Accelerated Science 22), that performs better for most of the molecular types in the test set of Main Group Chemistry Database (MGCDB84). Our framework opens a new direction in leveraging computing power for the systematic development of symbolic density functionals.

Defining Levels of Automated Chemical Design.

One application area of computational methods in drug discovery is the automated design of small molecules. Despite the large number of publications describing methods and their application in both retrospective and prospective studies, there is a lack of agreement on terminology and key attributes to distinguish these various systems. We introduce Automated Chemical Design (ACD) Levels to clearly define the level of autonomy along the axes of ideation and decision making. To fully illustrate this framework, we provide literature exemplars and place some notable methods and applications into the levels. The ACD framework provides a common language for describing automated small molecule design systems and enables medicinal chemists to better understand and evaluate such systems.

Breast cancer patient characterisation and visualisation using deep learning and fisher information networks.

Breast cancer is the most commonly diagnosed female malignancy globally, with better survival rates if diagnosed early. Mammography is the gold standard in screening programmes for breast cancer, but despite technological advances, high error rates are still reported. Machine learning techniques, and in particular deep learning (DL), have been successfully used for breast cancer detection and classification. However, the added complexity that makes DL models so successful reduces their ability to explain which features are relevant to the model, or whether the model is biased. The main aim of this study is to propose a novel visualisation to help characterise breast cancer patients using Fisher Information Networks on features extracted from mammograms using a DL model. In the proposed visualisation, patients are mapped out according to their similarities and can be used to study new patients as a 'patient-like-me' approach. When applied to the CBIS-DDSM dataset, it was shown that it is a competitive methodology that can (i) facilitate the analysis and decision-making process in breast cancer diagnosis with the assistance of the FIN visualisations and 'patient-like-me' analysis, and (ii) help improve diagnostic accuracy and reduce overdiagnosis by identifying the most likely diagnosis based on clinical similarities with neighbouring patients.

Non-ablative doses of focal ionizing radiation alters function of central neural circuits.

BACKGROUND: Modulation of pathological neural circuit activity in the brain with a minimum of complications is an area of intense interest. OBJECTIVE: The goal of the study was to alter neurons' physiological states without apparent damage of cellular integrity using stereotactic radiosurgery (SRS). METHODS: We treated a 7.5 mm-diameter target on the visual cortex of Göttingen minipigs with doses of 40, 60, 80, and 100 Gy. Six months post-irradiation, the pigs were implanted with a 9 mm-wide, eight-shank multi-electrode probe, which spanned the radiation focus as well as the low-exposure neighboring areas. RESULTS: Doses of 40 Gy led to an increase of spontaneous firing rate, six months post-irradiation, while doses of 60 Gy and greater were associated with a decrease. Subjecting the animals to visual stimuli resulted in typical visual evoked potentials (VEP). At 40 Gy, a significant reduction of the P1 peak time, indicative of higher network excitability was observed. At 80 Gy, P1 peak time was not affected, while a minor reduction at 60 Gy was seen. No distance-dependent effects on spontaneous firing rate, or on VEP were observed. Post-mortem histology revealed no evidence of necrosis at doses below 60 Gy. In an in vitro assay comprising of iPS-derived human neuron-astrocyte co-cultures, we found a higher vulnerability of inhibitory neurons than excitatory neurons with respect to radiation, which might provide the cellular mechanism of the disinhibitory effect observed in vivo. CONCLUSION: We provide initial evidence for a rather circuit-wide, long-lasting disinhibitory effect of low sub-ablative doses of SRS.

CACHE (Critical Assessment of Computational Hit-finding Experiments): A public-private partnership benchmarking initiative to enable the development of computational methods for hit-finding.

One aspirational goal of computational chemistry is to predict potent and drug-like binders for any protein, such that only those that bind are synthesized. In this Roadmap, we describe the launch of Critical Assessment of Computational Hit-finding Experiments (CACHE), a public benchmarking project to compare and improve small molecule hit-finding algorithms through cycles of prediction and experimental testing. Participants will predict small molecule binders for new and biologically relevant protein targets representing different prediction scenarios. Predicted compounds will be tested rigorously in an experimental hub, and all predicted binders as well as all experimental screening data, including the chemical structures of experimentally tested compounds, will be made publicly available, and not subject to any intellectual property restrictions. The ability of a range of computational approaches to find novel binders will be evaluated, compared, and openly published. CACHE will launch 3 new benchmarking exercises every year. The outcomes will be better prediction methods, new small molecule binders for target proteins of importance for fundamental biology or drug discovery, and a major technological step towards achieving the goal of Target 2035, a global initiative to identify pharmacological probes for all human proteins.