T Helper Cell Cytokines Modulate Intestinal Stem Cell Renewal and Differentiation.

In the small intestine, a niche of accessory cell types supports the generation of mature epithelial cell types from intestinal stem cells (ISCs). It is unclear, however, if and how immune cells in the niche affect ISC fate or the balance between self-renewal and differentiation. Here, we use single-cell RNA sequencing (scRNA-seq) to identify MHC class II (MHCII) machinery enrichment in two subsets of Lgr5+ ISCs. We show that MHCII+ Lgr5+ ISCs are non-conventional antigen-presenting cells in co-cultures with CD4+ T helper (Th) cells. Stimulation of intestinal organoids with key Th cytokines affects Lgr5+ ISC renewal and differentiation in opposing ways: pro-inflammatory signals promote differentiation, while regulatory cells and cytokines reduce it. In vivo genetic perturbation of Th cells or MHCII expression on Lgr5+ ISCs impacts epithelial cell differentiation and IEC fate during infection. These interactions between Th cells and Lgr5+ ISCs, thus, orchestrate tissue-wide responses to external signals.

Organism-Level Analysis of Vaccination Reveals Networks of Protection across Tissues.

A fundamental challenge in immunology is to decipher the principles governing immune responses at the whole-organism scale. Here, using a comparative infection model, we observe immune signal propagation within and between organs to obtain a dynamic map of immune processes at the organism level. We uncover two inter-organ mechanisms of protective immunity mediated by soluble and cellular factors. First, analyzing ligand-receptor connectivity across tissues reveals that type I IFNs trigger a whole-body antiviral state, protecting the host within hours after skin vaccination. Second, combining parabiosis, single-cell analyses, and gene knockouts, we uncover a multi-organ web of tissue-resident memory T cells that functionally adapt to their environment to stop viral spread across the organism. These results have implications for manipulating tissue-resident memory T cells through vaccination and open up new lines of inquiry for the analysis of immune responses at the organism level.

Chemokine guidance of central memory T cells is critical for antiviral recall responses in lymph nodes.

A defining feature of vertebrate immunity is the acquisition of immunological memory, which confers enhanced protection against pathogens by mechanisms that are incompletely understood. Here, we compared responses by virus-specific naive T cells (T(N)) and central memory T cells (T(CM)) to viral antigen challenge in lymph nodes (LNs). In steady-state LNs, both T cell subsets localized in the deep T cell area and interacted similarly with antigen-presenting dendritic cells. However, upon entry of lymph-borne virus, only T(CM) relocalized rapidly and efficiently toward the outermost LN regions in the medullary, interfollicular, and subcapsular areas where viral infection was initially confined. This rapid peripheralization was coordinated by a cascade of cytokines and chemokines, particularly ligands for T(CM)-expressed CXCR3. Consequently, in vivo recall responses to viral infection by CXCR3-deficient T(CM) were markedly compromised, indicating that early antigen detection afforded by intranodal chemokine guidance of T(CM) is essential for efficient antiviral memory.

Adipose Type One Innate Lymphoid Cells Regulate Macrophage Homeostasis through Targeted Cytotoxicity.

Adipose tissue has a dynamic immune system that adapts to changes in diet and maintains homeostatic tissue remodeling. Adipose type 1 innate lymphoid cells (AT1-ILCs) promote pro-inflammatory macrophages in obesity, but little is known about their functions at steady state. Here we found that human and murine adipose tissue harbor heterogeneous populations of AT1-ILCs. Experiments using parabiotic mice fed a high-fat diet (HFD) showed differential trafficking of AT1-ILCs, particularly in response to short- and long-term HFD and diet restriction. At steady state, AT1-ILCs displayed cytotoxic activity toward adipose tissue macrophages (ATMs). Depletion of AT1-ILCs and perforin deficiency resulted in alterations in the ratio of inflammatory to anti-inflammatory ATMs, and adoptive transfer of AT1-ILCs exacerbated metabolic disorder. Diet-induced obesity impaired AT1-ILC killing ability. Our findings reveal a role for AT1-ILCs in regulating ATM homeostasis through cytotoxicity and suggest that this function is relevant in both homeostasis and metabolic disease.

The Chemokine Receptor CX3CR1 Defines Three Antigen-Experienced CD8 T Cell Subsets with Distinct Roles in Immune Surveillance and Homeostasis.

Infections induce pathogen-specific T cell differentiation into diverse effectors (Teff) that give rise to memory (Tmem) subsets. The cell-fate decisions and lineage relationships that underlie these transitions are poorly understood. Here, we found that the chemokine receptor CX3CR1 identifies three distinct CD8+ Teff and Tmem subsets. Classical central (Tcm) and effector memory (Tem) cells and their corresponding Teff precursors were CX3CR1- and CX3CR1high, respectively. Viral infection also induced a numerically stable CX3CR1int subset that represented ∼15% of blood-borne Tmem cells. CX3CR1int Tmem cells underwent more frequent homeostatic divisions than other Tmem subsets and not only self-renewed, but also contributed to the expanding CX3CR1- Tcm pool. Both Tcm and CX3CR1int cells homed to lymph nodes, but CX3CR1int cells, and not Tem cells, predominantly surveyed peripheral tissues. As CX3CR1int Tmem cells present unique phenotypic, homeostatic, and migratory properties, we designate this subset peripheral memory (tpm) cells and propose that tpm cells are chiefly responsible for the global surveillance of non-lymphoid tissues.

A RSC/nucleosome complex determines chromatin architecture and facilitates activator binding.

How is chromatin architecture established and what role does it play in transcription? We show that the yeast regulatory locus UASg bears, in addition to binding sites for the activator Gal4, sites bound by the RSC complex. RSC positions a nucleosome, evidently partially unwound, in a structure that facilitates Gal4 binding to its sites. The complex comprises a barrier that imposes characteristic features of chromatin architecture. In the absence of RSC, ordinary nucleosomes encroach over the UASg and compete with Gal4 for binding. Taken with our previous work, the results show that both prior to and following induction, specific DNA-binding proteins are the predominant determinants of chromatin architecture at the GAL1/10 genes. RSC/nucleosome complexes are also found scattered around the yeast genome. Higher eukaryotic RSC lacks the specific DNA-binding determinants found on yeast RSC, and evidently Gal4 works in those organisms despite whatever obstacle broadly positioned nucleosomes present.

SOCS3 regulates pathological retinal angiogenesis through modulating SPP1 expression in microglia and macrophages.

Pathological ocular angiogenesis has long been associated with myeloid cell activation. However, the precise cellular and molecular mechanisms governing the intricate crosstalk between the immune system and vascular changes during ocular neovascularization formation remain elusive. In this study, we demonstrated that the absence of the suppressor of cytokine signaling 3 (SOCS3) in myeloid cells led to a substantial accumulation of microglia and macrophage subsets during the neovascularization process. Our single-cell RNA sequencing data analysis revealed a remarkable increase in the expression of the secreted phosphoprotein 1 (Spp1) gene within these microglia and macrophages, identifying subsets of Spp1-expressing microglia and macrophages during neovascularization formation in angiogenesis mouse models. Notably, the number of Spp1-expressing microglia and macrophages exhibited further elevation during neovascularization in mice lacking myeloid SOCS3. Moreover, our investigation unveiled the Spp1 gene as a direct transcriptional target gene of signal transducer and activator of transcription 3. Importantly, pharmaceutical activation of SOCS3 or blocking of SPP1 resulted in a significant reduction in pathological neovascularization. In conclusion, our study highlights the pivotal role of the SOCS3/STAT3/SPP1 axis in the regulation of pathological retinal angiogenesis.

Expression and characterization of a novel ß-1,4-endoglucanase from Bacillus subtilis strain isolated from a pulp and paper mill wastewater.

The production of fermentable sugars from lignocellulosic biomass is achieved by the synergistic action of a group of enzymes called cellulases. Cellulose is a long chain of chemically linked glucoses by ß-1,4 bonds. The enzyme ß-1,4-endoglucanase is the first cellulase involved in the degradation, breaking the bond of the amorphous regions. A ß-1,4-endoglucanase enzyme with high activity was obtained from a Bacillus subtilis strain isolated from wastewater of a pulp and paper mill. Sequencing and bioinformatic analysis showed that the gene amplified by PCR consisting of 1407 nucleotides and coding for a ß-1,4-endoglucanase enzyme of approximately 55 kDa. The open reading frame (ORF) encoding the mature endoglucanase (eglS) was successfully inserted in a modified cloning plasmid (pITD03) and into the pYD1 plasmid used for its expression in yeast. Carboxymethylcellulose (CMC) plate assay, SDS-PAGE, and zymogram confirmed the production and secretion by the transformed E. coli BL21-SI strain of a 39 kDa ß-1,4-endoglucanase consistent with the catalytic domain without the cellulose-binding module (CBM). The results showed that the truncated ß-1,4-endoglucanase had higher activity and stability.

De facto Water Reuse: Investigating the Fate and Transport of Chemicals of Emerging Concern from Wastewater Discharge through Drinking Water Treatment Using Non-targeted Analysis and Suspect Screening.

Wastewater is a source for many contaminants of emerging concern (CECs), and surface waters receiving wastewater discharge often serve as source water for downstream drinking water treatment plants. Nontargeted analysis and suspect screening methods were used to characterize chemicals in residence-time-weighted grab samples and companion polar organic chemical integrative samplers (POCIS) collected on three separate hydrologic sampling events along a surface water flow path representative of de facto water reuse. The goal of this work was to examine the fate of CECs along the study flow path as water is transported from wastewater effluent through drinking water treatment. Grab and POCIS samples provided a comparison between residence-time-weighted single-point and integrative sample results. This unique and rigorous study design, coupled with advanced analytical chemistry tools, provided important insights into chemicals found in drinking water and their potential sources, which can be used to help prioritize chemicals for further study. K-means clustering analysis was used to identify patterns in chemical occurrences across both sampling sites and sampling events. Chemical features that occurred frequently or survived drinking water treatment were prioritized for identification, resulting in the probable identification of over 100 CECs in the watershed and 28 CECs in treated drinking water.

An Evolved 5' Untranslated Region of Alfalfa Mosaic Virus Allows the RNA Transport of Movement-Defective Variants.

Although the coat protein (CP) has a relevant role in the long-distance movement of alfalfa mosaic virus (AMV) and brome mosaic virus (BMV), its precise function is not fully understood. Previous results showed that a specific interaction between the C termini of the movement protein (MP) and the cognate CP is required for systemic transport. Thus, we have performed a compensatory evolution experiment using an AMV RNA3 derivative defective in long-distance transport that carries a BMV MP lacking the C-terminal 48 residues and unable to interact with the AMV CP. After several passages, five independent evolution lineages were able to move long distance. The analysis of the viral RNA of these lineages showed the presence of three different modifications located exclusively at the 5' untranslated region (5' UTR). The three evolved 5' UTR variants accumulated comparable levels of viral RNA and CP but reduced the accumulation of virus particles and the affinity between the 5' UTR and the AMV CP. In addition, the evolved 5' UTR increased cell-to-cell transport for both the AMV RNA3 carrying the BMV MP and that carrying the AMV MP. Finally, the evolved 5' UTRs allowed the systemic transport of an AMV RNA3 carrying a CP mutant defective in virus particles and increased the systemic transport of several AMV RNA3 derivatives carrying different viral MPs associated with the 30K superfamily. Altogether, our findings indicate that virus particles are not required for the systemic transport of AMV but also that BMV MP is competent for the short- and long-distance transport without the interaction with the CP. IMPORTANCE The results obtained in the present work could challenge the view of the role of the virus particle in the systemic transport of plant viruses. In this sense, we show that two different MPs are competent to systemically transport the AMV genome without the requirement of the virus particles, as reported for viruses lacking a CP (e.g., Umbravirus). The incapability of the viral MP to interact with the CP triggered virus variants that evolved to reduce the formation of virus particles, probably to increase the accessibility of the MP to the viral progeny. Our results point to the idea that virus particles would not be necessary for the viral systemic transport but would be necessary for vector virus transmission. This idea is reinforced by the observation that heterologous MPs also increased the systemic transport of the AMV constructs that have reduced encapsidation capabilities.

Transformer Performance for Chemical Reactions: Analysis of Different Predictive and Evaluation Scenarios.

The prediction of chemical reaction pathways has been accelerated by the development of novel machine learning architectures based on the deep learning paradigm. In this context, deep neural networks initially designed for language translation have been used to accurately predict a wide range of chemical reactions. Among models suited for the task of language translation, the recently introduced molecular transformer reached impressive performance in terms of forward-synthesis and retrosynthesis predictions. In this study, we first present an analysis of the performance of transformer models for product, reactant, and reagent prediction tasks under different scenarios of data availability and data augmentation. We find that the impact of data augmentation depends on the prediction task and on the metric used to evaluate the model performance. Second, we probe the contribution of different combinations of input formats, tokenization schemes, and embedding strategies to model performance. We find that less stable input settings generally lead to better performance. Lastly, we validate the superiority of round-trip accuracy over simpler evaluation metrics, such as top-k accuracy, using a committee of human experts and show a strong agreement for predictions that pass the round-trip test. This demonstrates the usefulness of more elaborate metrics in complex predictive scenarios and highlights the limitations of direct comparisons to a predefined database, which may include a limited number of chemical reaction pathways.

Uptake of Per- and Polyfluoroalkyl Substances by Fish, Mussel, and Passive Samplers in Mobile-Laboratory Exposures Using Groundwater from a Contamination Plume at a Historical Fire Training Area, Cape Cod, Massachusetts.

Aqueous film-forming foams historically were used during fire training activities on Joint Base Cape Cod, Massachusetts, and created an extensive per- and polyfluoroalkyl substances (PFAS) groundwater contamination plume. The potential for PFAS bioconcentration from exposure to the contaminated groundwater, which discharges to surface water bodies, was assessed with mobile-laboratory experiments using groundwater from the contamination plume and a nearby reference location. The on-site continuous-flow 21-day exposures used male and female fathead minnows, freshwater mussels, polar organic chemical integrative samplers (POCIS), and polyethylene tube samplers (PETS) to evaluate biotic and abiotic uptake. The composition of the PFAS-contaminated groundwater was complex and 9 PFAS were detected in the reference groundwater and 17 PFAS were detected in the contaminated groundwater. The summed PFAS concentrations ranged from 120 to 140 ng L-1 in reference groundwater and 6100 to 15,000 ng L-1 in contaminated groundwater. Biotic concentration factors (CFb) for individual PFAS were species, sex, source, and compound-specific and ranged from 2.9 to 1000 L kg-1 in whole-body male fish exposed to contaminated groundwater for 21 days. The fish and mussel CFb generally increased with increasing fluorocarbon chain length and were greater for sulfonates than for carboxylates. The exception was perfluorohexane sulfonate, which deviated from the linear trend and had a 10-fold difference in CFb between sites, possibly because of biotransformation of precursors such as perfluorohexane sulfonamide. Uptake for most PFAS in male fish was linear over time, whereas female fish had bilinear uptake indicated by an initial increase in tissue concentrations followed by a decrease. Uptake of PFAS was less for mussels (maximum CFb = 200) than for fish, and mussel uptake of most PFAS also was bilinear. Although abiotic concentration factors were greater than CFb, and values for POCIS were greater than for PETS, passive samplers were useful for assessing PFAS that potentially bioconcentrate in fish but are present at concentrations below method quantitation limits in water. Passive samplers also accumulate short-chain PFAS that are not bioconcentrated.

Performance Analysis of Software-Defined Networks to Mitigate Private VLAN Attacks.

The defence-in-depth (DiD) methodology is a defensive approach usually performed by network administrators to implement secure networks by layering and segmenting them. Typically, segmentation is implemented in the second layer using the standard virtual local area networks (VLANs) or private virtual local area networks (PVLANs). Although defence in depth is usually manageable in small networks, it is not easily scalable to larger environments. Software-defined networks (SDNs) are emerging technologies that can be very helpful when performing network segmentation in such environments. In this work, a corporate networking scenario using PVLANs is emulated in order to carry out a comparative performance analysis on defensive strategies regarding CPU and memory usage, communications delay, packet loss, and power consumption. To do so, a well-known PVLAN attack is executed using simulated attackers located within the corporate network. Then, two mitigation strategies are analysed and compared using the traditional approach involving access control lists (ACLs) and SDNs. The results show the operation of the two mitigation strategies under different network scenarios and demonstrate the better performance of the SDN approach in oversubscribed network designs.

Ocular Vascular Diseases: From Retinal Immune Privilege to Inflammation.

The eye is an immune privileged tissue that insulates the visual system from local and systemic immune provocation to preserve homeostatic functions of highly specialized retinal neural cells. If immune privilege is breached, immune stimuli will invade the eye and subsequently trigger acute inflammatory responses. Local resident microglia become active and release numerous immunological factors to protect the integrity of retinal neural cells. Although acute inflammatory responses are necessary to control and eradicate insults to the eye, chronic inflammation can cause retinal tissue damage and cell dysfunction, leading to ocular disease and vision loss. In this review, we summarized features of immune privilege in the retina and the key inflammatory responses, factors, and intracellular pathways activated when retinal immune privilege fails, as well as a highlight of the recent clinical and research advances in ocular immunity and ocular vascular diseases including retinopathy of prematurity, age-related macular degeneration, and diabetic retinopathy.

Variation and correlation in the timing of breeding of North Atlantic seabirds across multiple scales.

Timing of breeding, an important driver of fitness in many populations, is widely studied in the context of global change, yet despite considerable efforts to identify environmental drivers of seabird nesting phenology, for most populations we lack evidence of strong drivers. Here we adopt an alternative approach, examining the degree to which different populations positively covary in their annual phenology to infer whether phenological responses to environmental drivers are likely to be (a) shared across species at a range of spatial scales, (b) shared across populations of a species or (c) idiosyncratic to populations. We combined 51 long-term datasets on breeding phenology spanning 50 years from nine seabird species across 29 North Atlantic sites and examined the extent to which different populations share early versus late breeding seasons depending on a hierarchy of spatial scales comprising breeding site, small-scale region, large-scale region and the whole North Atlantic. In about a third of cases, we found laying dates of populations of different species sharing the same breeding site or small-scale breeding region were positively correlated, which is consistent with the hypothesis that they share phenological responses to the same environmental conditions. In comparison, we found no evidence for positive phenological covariation among populations across species aggregated at larger spatial scales. In general, we found little evidence for positive phenological covariation between populations of a single species, and in many instances the inter-year variation specific to a population was substantial, consistent with each population responding idiosyncratically to local environmental conditions. Black-legged kittiwake Rissa tridactyla was the exception, with populations exhibiting positive covariation in laying dates that decayed with the distance between breeding sites, suggesting that populations may be responding to a similar driver. Our approach sheds light on the potential factors that may drive phenology in our study species, thus furthering our understanding of the scales at which different seabirds interact with interannual variation in their environment. We also identify additional systems and phenological questions to which our inferential approach could be applied.

Percutaneous Versus Transcutaneous Electrical Nerve Stimulation for the Treatment of Musculoskeletal Pain. A Systematic Review and Meta-Analysis.

BACKGROUND: The justification for this review is the need for high-quality evidence to assist in the decision-making process when applying percutaneous electrical nerve stimulation (PENS) or transcutaneous electrical nerve stimulation (TENS) in a clinical setting. The main aim was to determine if the use of PENS is more effective and should be recommended when compared to TENS for the reduction of musculoskeletal pain intensity. METHODS: A search for randomized controlled trials (RCTs) was performed. Studies published until 31/12/2020, comparing the effectiveness of PENS and TENS, were considered. The main outcome was pain assessed with a visual analog scale or numerical pain rating scale. RESULTS: Nine RCTs were included in the qualitative analysis, with seven of them in the quantitative analysis (n = 527). The overall effect of PENS on pain was statistically but not clinically superior to TENS (mean difference [MD]=-1.0 cm; 95% confidence interval [CI]: -1.5 to -0.4) with a high level of heterogeneity (I2=76%, P > .01). When only studies with a lower risk of bias (n = 3) were analyzed, the heterogeneity decreased to I = 0% (P = .06) and no difference was observed between TENS and PENS (MD=-0.81 cm; 95% CI:-1.6 to 0.02) with a moderate recommendation level according to GRADE. There were no data concerning adverse effects. CONCLUSIONS: There is low-quality of evidence for more pain intensity reduction with PENS, but the difference was not clinically significant. However, when only studies with low risk of bias are meta-analyzed, there is a moderate quality of evidence that there is no difference when TENS or PENS is applied for pain intensity.

Improvement in the Thermostability of a Recombinant ß-Glucosidase Immobilized in Zeolite under Different Conditions.

ß-Glucosidase is part of the cellulases and is responsible for degrading cellobiose into glucose, a compound that can be used to produce biofuels. However, the use of the free enzyme makes the process more expensive. Enzyme immobilization improves catalytic characteristics and supports, such as zeolites, which have physical-chemical characteristics and ion exchange capacity that have a promising application in the biotechnological industry. This research aimed to immobilize by adsorption a recombinant ß-glucosidase from Trichoderma reesei, obtained in Escherichia coli BL21 (DE3), in a commercial zeolite. A Box Behnken statistical design was applied to find the optimal immobilization parameters, the stability against pH and temperature was determined, and the immobilized enzyme was characterized by SEM. The highest enzymatic activity was determined with 100 mg of zeolite at 35 °C and 175 min. Compared to the free enzyme, the immobilized recombinant ß-glucosidase presented greater activity from pH 2 to 4 and greater thermostability. The kinetic parameters were calculated, and a lower KM value was obtained for the immobilized enzyme compared to the free enzyme. The obtained immobilization parameters by a simple adsorption method and the significant operational stability indicate promising applications in different fields.

Defining the clinical, molecular and imaging spectrum of adaptor protein complex 4-associated hereditary spastic paraplegia.

Bi-allelic loss-of-function variants in genes that encode subunits of the adaptor protein complex 4 (AP-4) lead to prototypical yet poorly understood forms of childhood-onset and complex hereditary spastic paraplegia: SPG47 (AP4B1), SPG50 (AP4M1), SPG51 (AP4E1) and SPG52 (AP4S1). Here, we report a detailed cross-sectional analysis of clinical, imaging and molecular data of 156 patients from 101 families. Enrolled patients were of diverse ethnic backgrounds and covered a wide age range (1.0-49.3 years). While the mean age at symptom onset was 0.8 ± 0.6 years [standard deviation (SD), range 0.2-5.0], the mean age at diagnosis was 10.2 ± 8.5 years (SD, range 0.1-46.3). We define a set of core features: early-onset developmental delay with delayed motor milestones and significant speech delay (50% non-verbal); intellectual disability in the moderate to severe range; mild hypotonia in infancy followed by spastic diplegia (mean age: 8.4 ± 5.1 years, SD) and later tetraplegia (mean age: 16.1 ± 9.8 years, SD); postnatal microcephaly (83%); foot deformities (69%); and epilepsy (66%) that is intractable in a subset. At last follow-up, 36% ambulated with assistance (mean age: 8.9 ± 6.4 years, SD) and 54% were wheelchair-dependent (mean age: 13.4 ± 9.8 years, SD). Episodes of stereotypic laughing, possibly consistent with a pseudobulbar affect, were found in 56% of patients. Key features on neuroimaging include a thin corpus callosum (90%), ventriculomegaly (65%) often with colpocephaly, and periventricular white-matter signal abnormalities (68%). Iron deposition and polymicrogyria were found in a subset of patients. AP4B1-associated SPG47 and AP4M1-associated SPG50 accounted for the majority of cases. About two-thirds of patients were born to consanguineous parents, and 82% carried homozygous variants. Over 70 unique variants were present, the majority of which are frameshift or nonsense mutations. To track disease progression across the age spectrum, we defined the relationship between disease severity as measured by several rating scales and disease duration. We found that the presence of epilepsy, which manifested before the age of 3 years in the majority of patients, was associated with worse motor outcomes. Exploring genotype-phenotype correlations, we found that disease severity and major phenotypes were equally distributed among the four subtypes, establishing that SPG47, SPG50, SPG51 and SPG52 share a common phenotype, an 'AP-4 deficiency syndrome'. By delineating the core clinical, imaging, and molecular features of AP-4-associated hereditary spastic paraplegia across the age spectrum our results will facilitate early diagnosis, enable counselling and anticipatory guidance of affected families and help define endpoints for future interventional trials.

Utilizing Artificial Intelligence to Manage COVID-19 Scientific Evidence Torrent with Risklick AI: A Critical Tool for Pharmacology and Therapy Development.

INTRODUCTION: The SARS-CoV-2 pandemic has led to one of the most critical and boundless waves of publications in the history of modern science. The necessity to find and pursue relevant information and quantify its quality is broadly acknowledged. Modern information retrieval techniques combined with artificial intelligence (AI) appear as one of the key strategies for COVID-19 living evidence management. Nevertheless, most AI projects that retrieve COVID-19 literature still require manual tasks. METHODS: In this context, we pre-sent a novel, automated search platform, called Risklick AI, which aims to automatically gather COVID-19 scientific evidence and enables scientists, policy makers, and healthcare professionals to find the most relevant information tailored to their question of interest in real time. RESULTS: Here, we compare the capacity of Risklick AI to find COVID-19-related clinical trials and scientific publications in comparison with clinicaltrials.gov and PubMed in the field of pharmacology and clinical intervention. DISCUSSION: The results demonstrate that Risklick AI is able to find COVID-19 references more effectively, both in terms of precision and recall, compared to the baseline platforms. Hence, Risklick AI could become a useful alternative assistant to scientists fighting the COVID-19 pandemic.

Information Retrieval in an Infodemic: The Case of COVID-19 Publications.

BACKGROUND: The COVID-19 global health crisis has led to an exponential surge in published scientific literature. In an attempt to tackle the pandemic, extremely large COVID-19-related corpora are being created, sometimes with inaccurate information, which is no longer at scale of human analyses. OBJECTIVE: In the context of searching for scientific evidence in the deluge of COVID-19-related literature, we present an information retrieval methodology for effective identification of relevant sources to answer biomedical queries posed using natural language. METHODS: Our multistage retrieval methodology combines probabilistic weighting models and reranking algorithms based on deep neural architectures to boost the ranking of relevant documents. Similarity of COVID-19 queries is compared to documents, and a series of postprocessing methods is applied to the initial ranking list to improve the match between the query and the biomedical information source and boost the position of relevant documents. RESULTS: The methodology was evaluated in the context of the TREC-COVID challenge, achieving competitive results with the top-ranking teams participating in the competition. Particularly, the combination of bag-of-words and deep neural language models significantly outperformed an Okapi Best Match 25-based baseline, retrieving on average, 83% of relevant documents in the top 20. CONCLUSIONS: These results indicate that multistage retrieval supported by deep learning could enhance identification of literature for COVID-19-related questions posed using natural language.