Enhancing reinforcement learning for de novo molecular design applying self-attention mechanisms.

The drug discovery process can be significantly improved by applying deep reinforcement learning (RL) methods that learn to generate compounds with desired pharmacological properties. Nevertheless, RL-based methods typically condense the evaluation of sampled compounds into a single scalar value, making it difficult for the generative agent to learn the optimal policy. This work combines self-attention mechanisms and RL to generate promising molecules. The idea is to evaluate the relative significance of each atom and functional group in their interaction with the target, and to utilize this information for optimizing the Generator. Therefore, the framework for de novo drug design is composed of a Generator that samples new compounds combined with a Transformer-encoder and a biological affinity Predictor that evaluate the generated structures. Moreover, it takes the advantage of the knowledge encapsulated in the Transformer's attention weights to evaluate each token individually. We compared the performance of two output prediction strategies for the Transformer: standard and masked language model (MLM). The results show that the MLM Transformer is more effective in optimizing the Generator compared with the state-of-the-art works. Additionally, the evaluation models identified the most important regions of each molecule for the biological interaction with the target. As a case study, we generated synthesizable hit compounds that can be putative inhibitors of the enzyme ubiquitin-specific protein 7 (USP7).

Effectiveness of Fludrocortisone Plus Hydrocortisone versus Hydrocortisone Alone in Septic Shock: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials.

Rationale: The use of hydrocortisone in adult patients with septic shock is controversial, and the effectiveness of adding fludrocortisone to hydrocortisone remains uncertain. Objectives: To assess the comparative effectiveness and safety of fludrocortisone plus hydrocortisone, hydrocortisone alone, and placebo or usual care in adults with septic shock. Methods: A systematic review and a Bayesian network meta-analysis of peer-reviewed randomized trials were conducted. The primary outcome was all-cause mortality at last follow-up. Treatment effects are presented as relative risks (RRs) with 95% credible intervals (CrIs). Placebo or usual care was the reference treatment. Measurements and Main Results: Among 7,553 references, we included 17 trials (7,688 patients). All-cause mortality at last follow-up was lowest with fludrocortisone plus hydrocortisone (RR, 0.85; 95% CrI, 0.72-0.99; 98.3% probability of superiority, moderate-certainty evidence), followed by hydrocortisone alone (RR, 0.97; 95% CrI, 0.87-1.07; 73.1% probability of superiority, low-certainty evidence). The comparison of fludrocortisone plus hydrocortisone versus hydrocortisone alone was based primarily on indirect evidence (only two trials with direct evidence). Fludrocortisone plus hydrocortisone was associated with a 12% lower risk of all-cause mortality compared with hydrocortisone alone (RR, 0.88; 95% CrI, 0.74-1.03; 94.2% probability of superiority, moderate-certainty evidence). Conclusions: In adult patients with septic shock, fludrocortisone plus hydrocortisone was associated with lower risk of all-cause mortality at last follow-up than placebo and hydrocortisone alone. The scarcity of head-to-head trials comparing fludrocortisone plus hydrocortisone versus hydrocortisone alone led our network meta-analysis to rely primarily on indirect evidence for this comparison. Although we undertook several sensitivity analyses and assessments, these findings should be considered while also acknowledging the heterogeneity of included trials.

Deep generative model for therapeutic targets using transcriptomic disease-associated data-USP7 case study.

The generation of candidate hit molecules with the potential to be used in cancer treatment is a challenging task. In this context, computational methods based on deep learning have been employed to improve in silico drug design methodologies. Nonetheless, the applied strategies have focused solely on the chemical aspect of the generation of compounds, disregarding the likely biological consequences for the organism's dynamics. Herein, we propose a method to implement targeted molecular generation that employs biological information, namely, disease-associated gene expression data, to conduct the process of identifying interesting hits. When applied to the generation of USP7 putative inhibitors, the framework managed to generate promising compounds, with more than 90% of them containing drug-like properties and essential active groups for the interaction with the target. Hence, this work provides a novel and reliable method for generating new promising compounds focused on the biological context of the disease.

Natural experiments and long-term monitoring are critical to understand and predict marine host-microbe ecology and evolution.

Marine multicellular organisms host a diverse collection of bacteria, archaea, microbial eukaryotes, and viruses that form their microbiome. Such host-associated microbes can significantly influence the host's physiological capacities; however, the identity and functional role(s) of key members of the microbiome ("core microbiome") in most marine hosts coexisting in natural settings remain obscure. Also unclear is how dynamic interactions between hosts and the immense standing pool of microbial genetic variation will affect marine ecosystems' capacity to adjust to environmental changes. Here, we argue that significantly advancing our understanding of how host-associated microbes shape marine hosts' plastic and adaptive responses to environmental change requires (i) recognizing that individual host-microbe systems do not exist in an ecological or evolutionary vacuum and (ii) expanding the field toward long-term, multidisciplinary research on entire communities of hosts and microbes. Natural experiments, such as time-calibrated geological events associated with well-characterized environmental gradients, provide unique ecological and evolutionary contexts to address this challenge. We focus here particularly on mutualistic interactions between hosts and microbes, but note that many of the same lessons and approaches would apply to other types of interactions.

Gupta Perioperative Risk for Myocardial Infarction or Cardiac Arrest Score is a Long-Term Cardiovascular Risk Predictor After Aortoiliac Revascularization.

BACKGROUND: Gupta Perioperative Risk for Myocardial Infarction or Cardiac Arrest (MICA) is a validated self-explanatory score applied in cardiac or noncardiac surgeries. This study aims to assess the predictive value of the MICA score for cardiovascular events after aortoiliac revascularization. METHODS: This prospective cohort underwent elective aortoiliac revascularization between 2013 and 2021. Patients' demographic, clinical characteristics, and outcomes were registered. The patients were divided into 2 groups according to the MICA score using optimal binning. Survival analysis to test for time-dependent variables and multivariate Cox regression analysis for independent predictors were performed. RESULTS: This study included 130 patients with a median follow-up of 55 months. Preoperative MICA score was ≥6.5 in 41 patients. MICA ≥6.5 presented a statistically significant association, with long-term occurrence of acute heart failure (HR = 1.695, 95% CI 1.208-2.379, P = 0.002), major adverse cardiovascular events (HR = 1.222, 95% CI 1.086-1.376, P < 0.001), and all-cause mortality (HR = 1.256, 95% CI 1.107-1.425, P < 0.001). Multivariable Cox regression confirmed MICA as a significant independent predictor of long-term major adverse cardiovascular events (aHR = 1.145 95% CI 1.010-1.298, P = 0.034) and all-cause mortality (aHR = 1.172 95% CI 1.026-1.339, P = 0.020). CONCLUSIONS: The MICA score is a quick, easy-to-obtain, predictive tool in identifying patients with a higher risk of postaortoiliac revascularization cardiovascular events, such as acute heart failure, major adverse cardiovascular events, and all-cause mortality. Additional research for the validation of the MICA score in the context of aortoiliac revascularization and specific interventions is necessary.

Optimized delay of the second COVID-19 vaccine dose reduces ICU admissions.

Slower than anticipated, COVID-19 vaccine production and distribution have impaired efforts to curtail the current pandemic. The standard administration schedule for most COVID-19 vaccines currently approved is two doses administered 3 to 4 wk apart. To increase the number of individuals with partial protection, some governments are considering delaying the second vaccine dose. However, the delay duration must take into account crucial factors, such as the degree of protection conferred by a single dose, the anticipated vaccine supply pipeline, and the potential emergence of more virulent COVID-19 variants. To help guide decision-making, we propose here an optimization model based on extended susceptible, exposed, infectious, and removed (SEIR) dynamics that determines the optimal delay duration between the first and second COVID-19 vaccine doses. The model assumes lenient social distancing and uses intensive care unit (ICU) admission as a key metric while selecting the optimal duration between doses vs. the standard 4-wk delay. While epistemic uncertainties apply to the interpretation of simulation outputs, we found that the delay is dependent on the vaccine mechanism of action and first-dose efficacy. For infection-blocking vaccines with first-dose efficacy ≥50%, the model predicts that the second dose can be delayed by ≥8 wk (half of the maximal delay), whereas for symptom-alleviating vaccines, the same delay is recommended only if the first-dose efficacy is ≥70%. Our model predicts that a 12-wk second-dose delay of an infection-blocking vaccine with a first-dose efficacy ≥70% could reduce ICU admissions by 400 people per million over 200 d.

Exploring Biosensors' Scientific Production and Research Patterns: A Bibliometric Analysis.

More sustainable biosensor production is growing in importance, allowing for the development of technological solutions for several industries, such as those in the health, chemical, and food sectors. Tracking the latest advancements in biosensors' scientific production is fundamental to determining the opportunities for the future of the biosensing field. This article aims to map scientific production in the biosensors field by running a bibliometric analysis of journal articles registered in the Web of Science database under biosensor-related vital concepts. The key concepts were selected by researchers and biosensor technology developers working on the BioAssembler Horizon project. The findings lead to identifying the scientific and technological knowledge base on biosensing devices and tracking the main scientific organisations developing this technology throughout the COVID-19 period (2019-2023). The institutional origin of the publications characterised the global distribution of related knowledge competencies and research partnerships. These results are discussed, shedding light on the scientific, economic, political, and structural factors that contribute to the formation of a scientific knowledge-based focus on the performance and design of these sensors. Moreover, the lack of scientific ties between the three axes of organisations producing expertise in this area (China, USA, and Russia) points towards the need to find synergies through new mechanisms of co-authorship and collaboration.

Vision System for a Forestry Navigation Machine.

This article presents the development of a vision system designed to enhance the autonomous navigation capabilities of robots in complex forest environments. Leveraging RGBD and thermic cameras, specifically the Intel RealSense 435i and FLIR ADK, the system integrates diverse visual sensors with advanced image processing algorithms. This integration enables robots to make real-time decisions, recognize obstacles, and dynamically adjust their trajectories during operation. The article focuses on the architectural aspects of the system, emphasizing the role of sensors and the formulation of algorithms crucial for ensuring safety during robot navigation in challenging forest terrains. Additionally, the article discusses the training of two datasets specifically tailored to forest environments, aiming to evaluate their impact on autonomous navigation. Tests conducted in real forest conditions affirm the effectiveness of the developed vision system. The results underscore the system's pivotal contribution to the autonomous navigation of robots in forest environments.

The evidence base of US Food and Drug Administration approvals of novel cancer therapies from 2000 to 2020.

Concerns have been raised that regulatory programs to accelerate approval of cancer drugs in cancer may increase uncertainty about benefits and harms for survival and quality of life (QoL). We analyzed all pivotal clinical trials and all non-pivotal randomized controlled trials (RCTs) for all cancer drugs approved for the first time by the FDA between 2000 and 2020. We report regulatory and trial characteristics. Effects on overall survival (OS), progression-free survival and tumor response were summarized in meta-analyses. Effects on QoL were qualitatively summarized. Between 2000 and 2020, the FDA approved 145 novel cancer drugs for 156 indications based on 190 clinical trials. Half of indications (49%) were approved without RCT evidence; 82% had a single clinical trial only. OS was primary endpoint in 14% of trials and QoL data were available from 25%. The median OS benefit was 2.55 months (IQR, 1.33-4.28) with a mean hazard ratio for OS of 0.75 (95%CI, 0.72-0.79, I2  = 42). Improvement for QoL was reported for 7 (4%) of 156 indications. Over time, priority review was used increasingly and the mean number of trials per indication decreased from 1.45 to 1.12. More trials reported results on QoL (19% in 2000-2005; 41% in 2016-2020). For 21 years, novel cancer drugs have typically been approved based on one single, often uncontrolled, clinical trial, measuring surrogate endpoints. This leaves cancer patients without solid evidence that novel drugs improve their survival or QoL and there is no indication towards improvement.

The microbiome of the pelagic tunicate Dolioletta gegenbauri: A potential link between the grazing and microbial food web.

Bloom-forming gelatinous zooplankton occur circumglobally and significantly influence the structure of pelagic marine food webs and biogeochemical cycling through interactions with microbial communities. During bloom conditions especially, gelatinous zooplankton are keystone taxa that help determine the fate of primary production, nutrient remineralization, and carbon export. Using the pelagic tunicate Dolioletta gegenbauri as a model system for gelatinous zooplankton, we carried out a laboratory-based feeding experiment to investigate the potential ecosystem impacts of doliolid gut microbiomes and microbial communities associated with doliolid faecal pellets and the surrounding seawater. Metabarcoding targeting Bacteria and Archaea 16S rRNA genes/Archaea) and qPCR approaches were used to characterize microbiome assemblages. Comparison between sample types revealed distinct patterns in microbial diversity and biomass that were replicable across experiments. These observations support the hypothesis that through their presence and trophic activity, doliolids influence the structure of pelagic food webs and biogeochemical cycling in subtropical continental shelf systems where tunicate blooms are common. Bacteria associated with starved doliolids (representative of the resident gut microbiome) possessed distinct low-biomass and low-diversity microbial assemblages, suggesting that the doliolid microbiome is optimized to support a detrital trophic mode. Bacterial genera Pseudoalteromomas and Shimia were the most abundant potential core microbiome taxa, similar to patterns observed in other marine invertebrates. Exploratory bioinformatic analyses of predicted functional genes suggest that doliolids, via their interactions with bacterial communities, may affect important biogeochemical processes including nitrogen, sulphur, and organic matter cycling.

Artificial intelligence for prediction of biological activities and generation of molecular hits using stereochemical information.

In this work, we develop a method for generating targeted hit compounds by applying deep reinforcement learning and attention mechanisms to predict binding affinity against a biological target while considering stereochemical information. The novelty of this work is a deep model Predictor that can establish the relationship between chemical structures and their corresponding [Formula: see text] values. We thoroughly study the effect of different molecular descriptors such as ECFP4, ECFP6, SMILES and RDKFingerprint. Also, we demonstrated the importance of attention mechanisms to capture long-range dependencies in molecular sequences. Due to the importance of stereochemical information for the binding mechanism, this information was employed both in the prediction and generation processes. To identify the most promising hits, we apply the self-adaptive multi-objective optimization strategy. Moreover, to ensure the existence of stereochemical information, we consider all the possible enumerated stereoisomers to provide the most appropriate 3D structures. We evaluated this approach against the Ubiquitin-Specific Protease 7 (USP7) by generating putative inhibitors for this target. The predictor with SMILES notations as descriptor plus bidirectional recurrent neural network using attention mechanism has the best performance. Additionally, our methodology identify the regions of the generated molecules that are important for the interaction with the receptor's active site. Also, the obtained results demonstrate that it is possible to discover synthesizable molecules with high biological affinity for the target, containing the indication of their optimal stereochemical conformation.

What Is the CRISPR System and How It Is Used?

CRISPR is a revolutionary gene editing technology that has enabled scientists worldwide to explore the cell's genetic blueprint in an unprecedented easy way. In this chapter, we will briefly present the history behind the development of this innovative tool, how it emerged from a natural bacterial mechanism for antiviral defense, its key components (Cas9 endonuclease and single guide RNA), mode of action (DNA cleavage and repair via NHEJ or HDR), and versatility (acting on single- or double-stranded DNA or RNA) for diverse purposes beyond gene editing such as stochastic marking, digital encoding, high-fidelity SNP genotyping, programmed chromosome fission/fusion, gene mapping, nucleic acid detection, regulation of gene expression, DNA/RNA labeling or tracking, and more.

Statistical analysis of blast-induced vibration near an open pit mine.

Blast-induced vibration may be harmful to facilities in the vicinity of operating mines, mainly causing structural damage and human discomfort. This study presents an application of multivariate statistics to predict vibration levels regarding their potential to cause structural damage and human discomfort. An extensive seismic monitoring campaign was executed in a large open-pit iron ore mine, near a small village, to gather a dataset for a predictive multivariate analysis. Ten blasting events have produced a dataset of 158 valid measurements. Three classes of vibration peak velocity were adopted from legal standards, which later supported a cluster analysis. Then, it was possible to compare how much these two classification modalities respond to discriminant analysis. The next step was to carry out a principal component analysis (PCA) from the original database, and, comparatively, to plot both the scores concerning the classes derived from the vibration standard and those from the groups obtained from cluster analysis. PCA has considerably explained the data variability, while the three classes from cluster analysis resulted very similar to the corresponding ones from the vibration standards. The results have demonstrated that multivariate statistics may be applied to manage blasting-induced vibration and its deleterious effects with few adjustments and automation.

Emergence of chaotic cluster synchronization in heterogeneous networks.

Many real-world complex systems rely on cluster synchronization to function properly. A cluster of nodes exhibits synchronous behavior, while others behave erratically. Predicting the emergence of these clusters and understanding the mechanism behind their structure and variation in response to a parameter change is a daunting task in networks that lack symmetry. We unravel the mechanism for the emergence of cluster synchronization in heterogeneous random networks. We develop heterogeneous mean-field approximation together with a self-consistent theory to determine the onset and stability of the cluster. Our analysis shows that cluster synchronization occurs in a wide variety of heterogeneous networks, node dynamics, and coupling functions. The results could lead to a new understanding of the dynamical behavior of networks ranging from neural to social.

Sensor Integration in a Forestry Machine.

This paper presents the integration of multimodal sensor systems for an autonomous forestry machine. The utilized technology is housed in a single enclosure which consolidates a set of components responsible for executing machine control actions and comprehending its behavior in various scenarios. This sensor box, named Sentry, will subsequently be connected to a forestry machine from MDB, model LV600 PRO. The article outlines previous work in this field and then details the integration and operation of the equipment, integrated into the forest machine, providing descriptions of the adopted architecture at both the hardware and software levels. The gathered data enables the assessment of the forestry machine's orientation and position based on the information collected by the sensors. Finally, practical experiments are presented to demonstrate the system's behavior and to analyze the methods to be employed for autonomous navigation, thereby assessing the performance of the established architecture. The novel aspects of this work include the physical and digital integration of a multimodal sensor system on a forestry machine, its use in a real case scenario, namely, forest vegetation removal, and the strategies adopted to improve the machine localization and navigation performance on unstructured environments.

Exponentially Long Transient Time to Synchronization of Coupled Chaotic Circle Maps in Dense Random Networks.

We study the transition to synchronization in large, dense networks of chaotic circle maps, where an exact solution of the mean-field dynamics in the infinite network and all-to-all coupling limit is known. In dense networks of finite size and link probability of smaller than one, the incoherent state is meta-stable for coupling strengths that are larger than the mean-field critical coupling. We observe chaotic transients with exponentially distributed escape times and study the scaling behavior of the mean time to synchronization.

Optimizing blood-brain barrier permeation through deep reinforcement learning for de novo drug design.

MOTIVATION: The process of placing new drugs into the market is time-consuming, expensive and complex. The application of computational methods for designing molecules with bespoke properties can contribute to saving resources throughout this process. However, the fundamental properties to be optimized are often not considered or conflicting with each other. In this work, we propose a novel approach to consider both the biological property and the bioavailability of compounds through a deep reinforcement learning framework for the targeted generation of compounds. We aim to obtain a promising set of selective compounds for the adenosine A2A receptor and, simultaneously, that have the necessary properties in terms of solubility and permeability across the blood-brain barrier to reach the site of action. The cornerstone of the framework is based on a recurrent neural network architecture, the Generator. It seeks to learn the building rules of valid molecules to sample new compounds further. Also, two Predictors are trained to estimate the properties of interest of the new molecules. Finally, the fine-tuning of the Generator was performed with reinforcement learning, integrated with multi-objective optimization and exploratory techniques to ensure that the Generator is adequately biased. RESULTS: The biased Generator can generate an interesting set of molecules, with approximately 85% having the two fundamental properties biased as desired. Thus, this approach has transformed a general molecule generator into a model focused on optimizing specific objectives. Furthermore, the molecules' synthesizability and drug-likeness demonstrate the potential applicability of the de novo drug design in medicinal chemistry. AVAILABILITY AND IMPLEMENTATION: All code is publicly available in the https://github.com/larngroup/De-Novo-Drug-Design. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Host-associated microbiomes drive structure and function of marine ecosystems.

The significance of symbioses between eukaryotic hosts and microbes extends from the organismal to the ecosystem level and underpins the health of Earth's most threatened marine ecosystems. Despite rapid growth in research on host-associated microbes, from individual microbial symbionts to host-associated consortia of significantly relevant taxa, little is known about their interactions with the vast majority of marine host species. We outline research priorities to strengthen our current knowledge of host-microbiome interactions and how they shape marine ecosystems. We argue that such advances in research will help predict responses of species, communities, and ecosystems to stressors driven by human activity and inform future management strategies.

EncephalApp Stroop Test validation for the screening of minimal hepatic encephalopathy in Brazil.

INTRODUCTION AND OBJECTIVES: The EncephalApp Stroop Test was developed to more easily diagnose minimal hepatic encephalopathy (MHE). A cut-off of >274.9sec (ONtime+OFFtime) reached a 78% sensitivity and 90% specificity in the validation study, but it has been poorly studied in Brazil. We aim to analyze the usefulness of this diagnostic method and to describe a cut-off value to screen MHE in Brazil. METHODS: In this cross-sectional and single-center study, three positive psychometric tests defined the diagnosis of MHE as the gold standard. We evaluated gender, age, education, familiarity with smartphones, etiology of cirrhosis, Child-Pugh/MELD scores, and previous hepatic encephalopathy (HE). Healthy controls and patients without HE were compared for the task validation. The Chi-square and Mann-Whitney tests, logistic regression analysis, and ROC curves were used for statistical evaluation. RESULTS: We included 132 patients with cirrhosis (61% male) and 42 controls (62% male) around 51y. Sixty-three were diagnosed with MHE on psychometric tests and 23 had clinical HE. Viral hepatitis (38%) was the major etiology of cirrhosis. The median MELD was 10 and Child-Pugh A was more frequent (70%). There was no significant difference in test results between controls and patients without HE. There was also no influence of gender, age, education, and familiarity with smartphones in the test results. Child-Pugh A was associated with MHE (p=0.0106). A cut-off of >269.8sec (ONtime+OFFtime) had an 87% sensitivity and 77% specificity to detect MHE (p=0.002). CONCLUSION: This is a valid and reliable tool for screening MHE. However, optimal cut-off values need to be validated locally.

User-centered design and spatially-distributed sequential electrical stimulation in cycling for individuals with paraplegia.

BACKGROUND: In this work, we share the enhancements made in our system to take part in the CYBATHLON 2020 Global Edition Functional Electrical Stimulation (FES) Bike Race. Among the main improvements, firstly an overhaul, an overhaul of the system and user interface developed with User-centered design principles with remote access to enable telerehabilitation. Secondly, the implementation and experimental comparison between the traditional single electrode stimulation (SES) and spatially distributed sequential stimulation (SDSS) applied for FES Cycling. METHODS: We report on the main aspects of the developed system. To evaluate the user perception of the system, we applied a System Usability Scale (SUS) questionnaire. In comparing SDSS and SES, we collected data from one subject in four sessions, each simulating one race in the CYBATHLON format. RESULTS: User perception measured with SUS indicates a positive outcome in the developed system. The SDSS trials were superior in absolute and average values to SES regarding total distance covered and velocity. We successfully competed in the CYBATHLON 2020 Global Edition, finishing in 6th position in the FES Bike Race category. CONCLUSIONS: The CYBATHLON format induced us to put the end-user in the center of our system design principle, which was well perceived. However, further improvements are required if the intention is to progress to a commercial product. FES Cycling performance in SDSS trials was superior when compared to SES trials, indicating that this technique may enable faster and possibly longer FES cycling sessions for individuals with paraplegia. More extensive studies are required to assess these aspects.