Machine Learning Informs RNA-Binding Chemical Space.

Small molecule targeting of RNA has emerged as a new frontier in medicinal chemistry, but compared to the protein targeting literature our understanding of chemical matter that binds to RNA is limited. In this study, we reported Repository Of BInders to Nucleic acids (ROBIN), a new library of nucleic acid binders identified by small molecule microarray (SMM) screening. The complete results of 36 individual nucleic acid SMM screens against a library of 24 572 small molecules were reported (including a total of 1 627 072 interactions assayed). A set of 2 003 RNA-binding small molecules was identified, representing the largest fully public, experimentally derived library of its kind to date. Machine learning was used to develop highly predictive and interpretable models to characterize RNA-binding molecules. This work demonstrates that machine learning algorithms applied to experimentally derived sets of RNA binders are a powerful method to inform RNA-targeted chemical space.

Early risk assessment for COVID-19 patients from emergency department data using machine learning.

Since its emergence in late 2019, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused a pandemic with more than 55 million reported cases and 1.3 million estimated deaths worldwide. While epidemiological and clinical characteristics of COVID-19 have been reported, risk factors underlying the transition from mild to severe disease among patients remain poorly understood. In this retrospective study, we analysed data of 879 confirmed SARS-CoV-2 positive patients admitted to a two-site NHS Trust hospital in London, England, between January 1st and May 26th, 2020, with a majority of cases occurring in March and April. We extracted anonymised demographic data, physiological clinical variables and laboratory results from electronic healthcare records (EHR) and applied multivariate logistic regression, random forest and extreme gradient boosted trees. To evaluate the potential for early risk assessment, we used data available during patients' initial presentation at the emergency department (ED) to predict deterioration to one of three clinical endpoints in the remainder of the hospital stay: admission to intensive care, need for invasive mechanical ventilation and in-hospital mortality. Based on the trained models, we extracted the most informative clinical features in determining these patient trajectories. Considering our inclusion criteria, we have identified 129 of 879 (15%) patients that required intensive care, 62 of 878 (7%) patients needing mechanical ventilation, and 193 of 619 (31%) cases of in-hospital mortality. Our models learned successfully from early clinical data and predicted clinical endpoints with high accuracy, the best model achieving area under the receiver operating characteristic (AUC-ROC) scores of 0.76 to 0.87 (F1 scores of 0.42-0.60). Younger patient age was associated with an increased risk of receiving intensive care and ventilation, but lower risk of mortality. Clinical indicators of a patient's oxygen supply and selected laboratory results, such as blood lactate and creatinine levels, were most predictive of COVID-19 patient trajectories. Among COVID-19 patients machine learning can aid in the early identification of those with a poor prognosis, using EHR data collected during a patient's first presentation at ED. Patient age and measures of oxygenation status during ED stay are primary indicators of poor patient outcomes.

C57BL/6 and 129 inbred mouse strains differ in Gbp2 and Gbp2b expression in response to inflammatory stimuli in vivo.

Background: Infections cause the production of inflammatory cytokines such as Interferon gamma (IFNγ). IFNγ in turn prompts the upregulation of a range of host defence proteins including members of the family of guanylate binding proteins (Gbps). In humans and mice alike, GBPs restrict the intracellular replication of invasive microbes and promote inflammation. To study the physiological functions of Gbp family members, the most commonly chosen in vivo models are mice harbouring loss-of-function mutations in either individual Gbp genes or the entire Gbp gene cluster on mouse chromosome 3. Individual Gbp deletion strains differ in their design, as some strains exist on a pure C57BL/6 genetic background, while other strains contain a 129-derived genetic interval encompassing the Gbp gene cluster on an otherwise C57BL/6 genetic background. Methods: To determine whether the presence of 129 alleles of paralogous Gbps could influence the phenotypes of 129-congenic Gbp-deficient strains, we studied the expression of Gbps in both C57BL/6J and 129/Sv mice following in vivo stimulation with adjuvants and after infection with either Toxoplasma gondii or Shigella flexneri. Results: We show that C57BL/6J relative to 129/Sv mice display moderately elevated expression of Gbp2, but more prominently, are also defective for Gbp2b (formerly Gbp1) mRNA induction upon immune priming. Notably, Toxoplasma infections induce robust Gbp2b protein expression in both strains of mice, suggestive of a Toxoplasma-activated mechanism driving Gbp2b protein translation. We further find that the higher expression of Gbp2b mRNA in 129/Sv mice correlates with a gene duplication event at the Gbp2b locus resulting in two copies of the Gbp2b gene on the haploid genome of the 129/Sv strain. Conclusions: Our findings demonstrate functional differences between 129 and C57BL/6 Gbp alleles which need to be considered in the design and interpretation of studies utilizing mouse models, particularly for phenotypes influenced by Gbp2 or Gbp2b expression.

Fine-mapping and phenotypic analysis of the Ity3 Salmonella susceptibility locus identify a complex genetic structure.

Experimental animal models of Salmonella infections have been widely used to identify genes important in the host immune response to infection. Using an F2 cross between the classical inbred strain C57BL/6J and the wild derived strain MOLF/Ei, we have previously identified Ity3 (Immunity to Typhimurium locus 3) as a locus contributing to the early susceptibility of MOLF/Ei mice to infection with Salmonella Typhimurium. We have also established a congenic strain (B6.MOLF-Ity/Ity3) with the MOLF/Ei Ity3 donor segment on a C57BL/6J background. The current study was designed to fine map and characterize functionally the Ity3 locus. We generated 12 recombinant sub-congenic strains that were characterized for susceptibility to infection, bacterial load in target organs, cytokine profile and anti-microbial mechanisms. These analyses showed that the impact of the Ity3 locus on survival and bacterial burden was stronger in male mice compared to female mice. Fine mapping of Ity3 indicated that two subloci contribute collectively to the susceptibility of B6.MOLF-Ity/Ity3 congenic mice to Salmonella infection. The Ity3.1 sublocus controls NADPH oxidase activity and is characterized by decreased ROS production, reduced inflammatory cytokine response and increased bacterial burden, thereby supporting a role for Ncf2 (neutrophil cytosolic factor 2 a subunit of NADPH oxidase) as the gene underlying this sublocus. The Ity3.2 sub-locus is characterized by a hyperresponsive inflammatory cytokine phenotype after exposure to Salmonella. Overall, this research provides support to the combined action of hormonal influences and complex genetic factors within the Ity3 locus in the innate immune response to Salmonella infection in wild-derived MOLF/Ei mice.

Functional validation of the genetic architecture of Salmonella Enteritidis persistence in 129S6 mice.

The Gram-negative bacteria, Salmonella, cause a broad spectrum of clinical diseases in humans, ranging from asymptomatic carriage to life-threatening sepsis. We have designed an experimental model to study the contribution of genetic factors to the persistence of Salmonella Enteritidis during the late phase of infection in 129S6/SvEvTac and C57BL/6J mice. C57BL/6J mice cleared the bacteria from their reticuloendothelial system within a period of 42 days, whereas the 129S6 mice still presented a high bacterial load. Using this model, we have identified ten Salmonella Enteritidis susceptibility loci (Ses1, Ses1.1, and Ses3-Ses10) associated with bacterial persistence in target organs of 129S6/SvEvTac mice using a two-locus epistasis QTL linkage mapping approach. Significant statistical interactions were detected between Ses1 on chromosome 1 and Ses5 on chromosome 7 and between Ses1 and Ses4 on chromosome X. In this study, we functionally validated the genetic architecture of Salmonella persistence in 129S6 mice using single- (129S6.B6-Ses1.2 that combines Ses1 and Ses1.1 loci, 129S6.B6-Ses4, and 129S6.B6-Ses5) and double-congenic mice (129S6.B6-Ses1.2/Ses4 and 129S6.B6-Ses1.2/Ses5). These experiments demonstrate functional interactions between Ses1.2 and Ses4 or Ses5 that improve Salmonella Enteritidis clearance, validating the critical role played by gene-gene interactions in the contribution to bacterial clearance heritability. Improved bacterial clearance in double-congenic mice could be explained by the impact of Ses4 and Ses5 in combination with Ses1.2 on TH polarization since a TH2 bias (decreased Ifng and increased Il4 mRNA levels and reduced IgG2a immunoglobulins in the serum) was observed in 129S6.B6-Ses1.2/Ses5 mice and a TH17 (high Il17 expression) bias in 129S6.B6-Ses1.2/Ses4.