Target-agnostic drug prediction integrated with medical record analysis uncovers differential associations of statins with increased survival in COVID-19 patients.

Drug repurposing requires distinguishing established drug class targets from novel molecule-specific mechanisms and rapidly derisking their therapeutic potential in a time-critical manner, particularly in a pandemic scenario. In response to the challenge to rapidly identify treatment options for COVID-19, several studies reported that statins, as a drug class, reduce mortality in these patients. However, it is unknown if different statins exhibit consistent function or may have varying therapeutic benefit. A Bayesian network tool was used to predict drugs that shift the host transcriptomic response to SARS-CoV-2 infection towards a healthy state. Drugs were predicted using 14 RNA-sequencing datasets from 72 autopsy tissues and 465 COVID-19 patient samples or from cultured human cells and organoids infected with SARS-CoV-2. Top drug predictions included statins, which were then assessed using electronic medical records containing over 4,000 COVID-19 patients on statins to determine mortality risk in patients prescribed specific statins versus untreated matched controls. The same drugs were tested in Vero E6 cells infected with SARS-CoV-2 and human endothelial cells infected with a related OC43 coronavirus. Simvastatin was among the most highly predicted compounds (14/14 datasets) and five other statins, including atorvastatin, were predicted to be active in > 50% of analyses. Analysis of the clinical database revealed that reduced mortality risk was only observed in COVID-19 patients prescribed a subset of statins, including simvastatin and atorvastatin. In vitro testing of SARS-CoV-2 infected cells revealed simvastatin to be a potent direct inhibitor whereas most other statins were less effective. Simvastatin also inhibited OC43 infection and reduced cytokine production in endothelial cells. Statins may differ in their ability to sustain the lives of COVID-19 patients despite having a shared drug target and lipid-modifying mechanism of action. These findings highlight the value of target-agnostic drug prediction coupled with patient databases to identify and clinically evaluate non-obvious mechanisms and derisk and accelerate drug repurposing opportunities.

Dense phenotyping from electronic health records enables machine learning-based prediction of preterm birth.

BACKGROUND: Identifying pregnancies at risk for preterm birth, one of the leading causes of worldwide infant mortality, has the potential to improve prenatal care. However, we lack broadly applicable methods to accurately predict preterm birth risk. The dense longitudinal information present in electronic health records (EHRs) is enabling scalable and cost-efficient risk modeling of many diseases, but EHR resources have been largely untapped in the study of pregnancy. METHODS: Here, we apply machine learning to diverse data from EHRs with 35,282 deliveries to predict singleton preterm birth. RESULTS: We find that machine learning models based on billing codes alone can predict preterm birth risk at various gestational ages (e.g., ROC-AUC = 0.75, PR-AUC = 0.40 at 28 weeks of gestation) and outperform comparable models trained using known risk factors (e.g., ROC-AUC = 0.65, PR-AUC = 0.25 at 28 weeks). Examining the patterns learned by the model reveals it stratifies deliveries into interpretable groups, including high-risk preterm birth subtypes enriched for distinct comorbidities. Our machine learning approach also predicts preterm birth subtypes (spontaneous vs. indicated), mode of delivery, and recurrent preterm birth. Finally, we demonstrate the portability of our approach by showing that the prediction models maintain their accuracy on a large, independent cohort (5978 deliveries) from a different healthcare system. CONCLUSIONS: By leveraging rich phenotypic and genetic features derived from EHRs, we suggest that machine learning algorithms have great potential to improve medical care during pregnancy. However, further work is needed before these models can be applied in clinical settings.

Myofibril breakdown during atrophy is a delayed response requiring the transcription factor PAX4 and desmin depolymerization.

A hallmark of muscle atrophy is the excessive degradation of myofibrillar proteins primarily by the ubiquitin proteasome system. In mice, during the rapid muscle atrophy induced by fasting, the desmin cytoskeleton and the attached Z-band-bound thin filaments are degraded after ubiquitination by the ubiquitin ligase tripartite motif-containing protein 32 (Trim32). To study the order of events leading to myofibril destruction, we investigated the slower atrophy induced by denervation (disuse). We show that myofibril breakdown is a two-phase process involving the initial disassembly of desmin filaments by Trim32, which leads to the later myofibril breakdown by enzymes, whose expression is increased by the paired box 4 (PAX4) transcription factor. After denervation of mouse tibialis anterior muscles, phosphorylation and Trim32-dependent ubiquitination of desmin filaments increased rapidly and stimulated their gradual depolymerization (unlike their rapid degradation during fasting). Trim32 down-regulation attenuated the loss of desmin and myofibrillar proteins and reduced atrophy. Although myofibrils and desmin filaments were intact at 7 d after denervation, inducing the dissociation of desmin filaments caused an accumulation of ubiquitinated proteins and rapid destruction of myofibrils. The myofibril breakdown normally observed at 14 d after denervation required not only dissociation of desmin filaments, but also gene induction by PAX4. Down-regulation of PAX4 or its target gene encoding the p97/VCP ATPase reduced myofibril disassembly and degradation on denervation or fasting. Thus, during atrophy, the initial loss of desmin is critical for the subsequent myofibril destruction, and over time, myofibrillar proteins become more susceptible to PAX4-induced enzymes that promote proteolysis.

A diverse epigenetic landscape at human exons with implication for expression.

DNA methylation is an important epigenetic marker associated with gene expression regulation in eukaryotes. While promoter methylation is relatively well characterized, the role of intragenic DNA methylation remains unclear. Here, we investigated the relationship of DNA methylation at exons and flanking introns with gene expression and histone modifications generated from a human fibroblast cell-line and primary B cells. Consistent with previous work we found that intragenic methylation is positively correlated with gene expression and that exons are more highly methylated than their neighboring intronic environment. Intriguingly, in this study we identified a unique subset of hypomethylated exons that demonstrate significantly lower methylation levels than their surrounding introns. Furthermore, we observed a negative correlation between exon methylation and the density of the majority of histone modifications. Specifically, we demonstrate that hypo-methylated exons at highly expressed genes are associated with open chromatin and have a characteristic histone code comprised of significantly high levels of histone markings. Overall, our comprehensive analysis of the human exome supports the presence of regulatory hypomethylated exons in protein coding genes. In particular our results reveal a previously unrecognized diverse and complex role of the epigenetic landscape within the gene body.

RBPmap: a web server for mapping binding sites of RNA-binding proteins.

Regulation of gene expression is executed in many cases by RNA-binding proteins (RBPs) that bind to mRNAs as well as to non-coding RNAs. RBPs recognize their RNA target via specific binding sites on the RNA. Predicting the binding sites of RBPs is known to be a major challenge. We present a new webserver, RBPmap, freely accessible through the website http://rbpmap.technion.ac.il/ for accurate prediction and mapping of RBP binding sites. RBPmap has been developed specifically for mapping RBPs in human, mouse and Drosophila melanogaster genomes, though it supports other organisms too. RBPmap enables the users to select motifs from a large database of experimentally defined motifs. In addition, users can provide any motif of interest, given as either a consensus or a PSSM. The algorithm for mapping the motifs is based on a Weighted-Rank approach, which considers the clustering propensity of the binding sites and the overall tendency of regulatory regions to be conserved. In addition, RBPmap incorporates a position-specific background model, designed uniquely for different genomic regions, such as splice sites, 5' and 3' UTRs, non-coding RNA and intergenic regions. RBPmap was tested on high-throughput RNA-binding experiments and was proved to be highly accurate.

The hnRNP F/H homologue of Trypanosoma brucei is differentially expressed in the two life cycle stages of the parasite and regulates splicing and mRNA stability.

Trypanosomes are protozoan parasites that cycle between a mammalian host (bloodstream form) and an insect host, the Tsetse fly (procyclic stage). In trypanosomes, all mRNAs are trans-spliced as part of their maturation. Genome-wide analysis of trans-splicing indicates the existence of alternative trans-splicing, but little is known regarding RNA-binding proteins that participate in such regulation. In this study, we performed functional analysis of the Trypanosoma brucei heterogeneous nuclear ribonucleoproteins (hnRNP) F/H homologue, a protein known to regulate alternative splicing in metazoa. The hnRNP F/H is highly expressed in the bloodstream form of the parasite, but is also functional in the procyclic form. Transcriptome analyses of RNAi-silenced cells were used to deduce the RNA motif recognized by this protein. A purine rich motif, AAGAA, was enriched in both the regulatory regions flanking the 3' splice site and poly (A) sites of the regulated genes. The motif was further validated using mini-genes carrying wild-type and mutated sequences in the 3' and 5' UTRs, demonstrating the role of hnRNP F/H in mRNA stability and splicing. Biochemical studies confirmed the binding of the protein to this proposed site. The differential expression of the protein and its inverse effects on mRNA level in the two lifecycle stages demonstrate the role of hnRNP F/H in developmental regulation.

Diverse lifestyles and strategies of plant pathogenesis encoded in the genomes of eighteen Dothideomycetes fungi.

The class Dothideomycetes is one of the largest groups of fungi with a high level of ecological diversity including many plant pathogens infecting a broad range of hosts. Here, we compare genome features of 18 members of this class, including 6 necrotrophs, 9 (hemi)biotrophs and 3 saprotrophs, to analyze genome structure, evolution, and the diverse strategies of pathogenesis. The Dothideomycetes most likely evolved from a common ancestor more than 280 million years ago. The 18 genome sequences differ dramatically in size due to variation in repetitive content, but show much less variation in number of (core) genes. Gene order appears to have been rearranged mostly within chromosomal boundaries by multiple inversions, in extant genomes frequently demarcated by adjacent simple repeats. Several Dothideomycetes contain one or more gene-poor, transposable element (TE)-rich putatively dispensable chromosomes of unknown function. The 18 Dothideomycetes offer an extensive catalogue of genes involved in cellulose degradation, proteolysis, secondary metabolism, and cysteine-rich small secreted proteins. Ancestors of the two major orders of plant pathogens in the Dothideomycetes, the Capnodiales and Pleosporales, may have had different modes of pathogenesis, with the former having fewer of these genes than the latter. Many of these genes are enriched in proximity to transposable elements, suggesting faster evolution because of the effects of repeat induced point (RIP) mutations. A syntenic block of genes, including oxidoreductases, is conserved in most Dothideomycetes and upregulated during infection in L. maculans, suggesting a possible function in response to oxidative stress.

Proteomic profiles of cytokines and chemokines in moderate to severe depression: Implications for comorbidities and biomarker discovery.

Objective: This study assessed the proteomic profiles of cytokines and chemokines in individuals with moderate to severe depression, with or without comorbid medical disorders, compared to healthy controls. Two proteomic multiplex platforms were employed for this purpose. Metods: An immunofluorescent multiplex platform and an aptamer-based method were used to evaluate 32 protein analytes from 153 individuals with moderate to severe major depressive disorder (MDD) and healthy controls (HCs). The study focused on determining the level of agreement between the two platforms and evaluating the ability of individual analytes and principal components (PCs) to differentiate between the MDD and HC groups. Additionally, the study investigated the relationship between PCs consisting of chemokines and cytokines and comorbid inflammatory and cardiometabolic diseases. Findings: Analysis revealed a small or moderate correlation between 47% of the analytes measured by the two platforms. Two proteomic profiles were identified that differentiated individuals with moderate to severe MDD from HCs. High eotaxin, age, BMI, IP-10, or IL-10 characterized profile 1. This profile was associated with several cardiometabolic risk factors, including hypertension, hyperlipidemia, and type 2 diabetes. Profile 2 is characterized by higher age, BMI, interleukins, and a strong negative loading for eotaxin. This profile was associated with inflammation but not cardiometabolic risk factors. Conclusion: This study provides further evidence that proteomic profiles can be used to identify potential biomarkers and pathways associated with MDD and comorbidities. Our findings suggest that MDD is associated with distinct profiles of proteins that are also associated with cardiometabolic risk factors, inflammation, and obesity. In particular, the chemokines eotaxin and IP-10 appear to play a role in the relationship between MDD and cardiometabolic risk factors. These findings suggest that a focus on the interplay between MDD and comorbidities may be useful in identifying potential targets for intervention and improving overall health outcomes.

Identifying therapeutic candidates for endometriosis through a transcriptomics-based drug repositioning approach.

Existing medical treatments for endometriosis-related pain are often ineffective, underscoring the need for new therapeutic strategies. In this study, we applied a computational drug repurposing pipeline to stratified and unstratified disease signatures based on endometrial gene expression data to identify potential therapeutics from existing drugs, based on expression reversal. Of 3,131 unique genes differentially expressed by at least one of six endometriosis signatures, only 308 (9.8%) were in common; however, 221 out of 299 drugs identified, (73.9%) were shared. We selected fenoprofen, an uncommonly prescribed NSAID that was the top therapeutic candidate for further investigation. When testing fenoprofen in an established rat model of endometriosis, fenoprofen successfully alleviated endometriosis-associated vaginal hyperalgesia, a surrogate marker for endometriosis-related pain. These findings validate fenoprofen as a therapeutic that could be utilized more frequently for endometriosis and suggest the utility of the aforementioned computational drug repurposing approach for endometriosis.

Microbiome preterm birth DREAM challenge: Crowdsourcing machine learning approaches to advance preterm birth research.

Every year, 11% of infants are born preterm with significant health consequences, with the vaginal microbiome a risk factor for preterm birth. We crowdsource models to predict (1) preterm birth (PTB; <37 weeks) or (2) early preterm birth (ePTB; <32 weeks) from 9 vaginal microbiome studies representing 3,578 samples from 1,268 pregnant individuals, aggregated from public raw data via phylogenetic harmonization. The predictive models are validated on two independent unpublished datasets representing 331 samples from 148 pregnant individuals. The top-performing models (among 148 and 121 submissions from 318 teams) achieve area under the receiver operator characteristic (AUROC) curve scores of 0.69 and 0.87 predicting PTB and ePTB, respectively. Alpha diversity, VALENCIA community state types, and composition are important features in the top-performing models, most of which are tree-based methods. This work is a model for translation of microbiome data into clinically relevant predictive models and to better understand preterm birth.

An integrated regulatory network reveals pervasive cross-regulation among transcription and splicing factors.

Traditionally the gene expression pathway has been regarded as being comprised of independent steps, from RNA transcription to protein translation. To date there is increasing evidence of coupling between the different processes of the pathway, specifically between transcription and splicing. To study the interplay between these processes we derived a transcription-splicing integrated network. The nodes of the network included experimentally verified human proteins belonging to three groups of regulators: transcription factors, splicing factors and kinases. The nodes were wired by instances of predicted transcriptional and alternative splicing regulation. Analysis of the network indicated a pervasive cross-regulation among the nodes; specifically, splicing factors are significantly more connected by alternative splicing regulatory edges relative to the two other subgroups, while transcription factors are more extensively controlled by transcriptional regulation. Furthermore, we found that splicing factors are the most regulated of the three regulatory groups and are subject to extensive combinatorial control by alternative splicing and transcriptional regulation. Consistent with the network results, our bioinformatics analyses showed that the subgroup of kinases have the highest density of predicted phosphorylation sites. Overall, our systematic study reveals that an organizing principle in the logic of integrated networks favor the regulation of regulatory proteins by the specific regulation they conduct. Based on these results, we propose a new regulatory paradigm postulating that gene expression regulation of the master regulators in the cell is predominantly achieved by cross-regulation.

Leveraging electronic health records to identify risk factors for recurrent pregnancy loss across two medical centers: a case-control study.

Recurrent pregnancy loss (RPL), defined as 2 or more pregnancy losses, affects 5-6% of ever-pregnant individuals. Approximately half of these cases have no identifiable explanation. To generate hypotheses about RPL etiologies, we implemented a case-control study comparing the history of over 1,600 diagnoses between RPL and live-birth patients, leveraging the University of California San Francisco (UCSF) and Stanford University electronic health record databases. In total, our study included 8,496 RPL (UCSF: 3,840, Stanford: 4,656) and 53,278 Control (UCSF: 17,259, Stanford: 36,019) patients. Menstrual abnormalities and infertility-associated diagnoses were significantly positively associated with RPL in both medical centers. Age-stratified analysis revealed that the majority of RPL-associated diagnoses had higher odds ratios for patients <35 compared with 35+ patients. While Stanford results were sensitive to control for healthcare utilization, UCSF results were stable across analyses with and without utilization. Intersecting significant results between medical centers was an effective filter to identify associations that are robust across center-specific utilization patterns.

VMAP: Vaginal Microbiome Atlas During Pregnancy.

The vaginal microbiome has been shown to be associated with pregnancy outcomes including preterm birth (PTB) risk. Here we present VMAP: Vaginal Microbiome Atlas during Pregnancy (http://vmapapp.org), an application to visualize features of 3,909 vaginal microbiome samples of 1,416 pregnant individuals from 11 studies, aggregated from raw public and newly generated sequences via an open-source tool, MaLiAmPi. Our visualization tool (http://vmapapp.org) includes microbial features such as various measures of diversity, VALENCIA community state types (CST), and composition (via phylotypes and taxonomy). This work serves as a resource for the research community to further analyze and visualize vaginal microbiome data in order to better understand both healthy term pregnancies and those associated with adverse outcomes.

Global endometrial DNA methylation analysis reveals insights into mQTL regulation and associated endometriosis disease risk and endometrial function.

Endometriosis is a leading cause of pain and infertility affecting millions of women globally. Herein, we characterize variation in DNA methylation (DNAm) and its association with menstrual cycle phase, endometriosis, and genetic variants through analysis of genotype data and methylation in endometrial samples from 984 deeply-phenotyped participants. We estimate that 15.4% of the variation in endometriosis is captured by DNAm and identify significant differences in DNAm profiles associated with stage III/IV endometriosis, endometriosis sub-phenotypes and menstrual cycle phase, including opening of the window for embryo implantation. Menstrual cycle phase was a major source of DNAm variation suggesting cellular and hormonally-driven changes across the cycle can regulate genes and pathways responsible for endometrial physiology and function. DNAm quantitative trait locus (mQTL) analysis identified 118,185 independent cis-mQTLs including 51 associated with risk of endometriosis, highlighting candidate genes contributing to disease risk. Our work provides functional evidence for epigenetic targets contributing to endometriosis risk and pathogenesis. Data generated serve as a valuable resource for understanding tissue-specific effects of methylation on endometrial biology in health and disease.

Microbiome Preterm Birth DREAM Challenge: Crowdsourcing Machine Learning Approaches to Advance Preterm Birth Research.

Globally, every year about 11% of infants are born preterm, defined as a birth prior to 37 weeks of gestation, with significant and lingering health consequences. Multiple studies have related the vaginal microbiome to preterm birth. We present a crowdsourcing approach to predict: (a) preterm or (b) early preterm birth from 9 publicly available vaginal microbiome studies representing 3,578 samples from 1,268 pregnant individuals, aggregated from raw sequences via an open-source tool, MaLiAmPi. We validated the crowdsourced models on novel datasets representing 331 samples from 148 pregnant individuals. From 318 DREAM challenge participants we received 148 and 121 submissions for our two separate prediction sub-challenges with top-ranking submissions achieving bootstrapped AUROC scores of 0.69 and 0.87, respectively. Alpha diversity, VALENCIA community state types, and composition (via phylotype relative abundance) were important features in the top performing models, most of which were tree based methods. This work serves as the foundation for subsequent efforts to translate predictive tests into clinical practice, and to better understand and prevent preterm birth.

SFmap: a web server for motif analysis and prediction of splicing factor binding sites.

Alternative splicing (AS) is a post-transcriptional process considered to be responsible for the huge diversity of proteins in higher eukaryotes. AS events are regulated by different splicing factors (SFs) that bind to sequence elements on the RNA. SFmap is a web server for predicting putative SF binding sites in genomic data (http://sfmap.technion.ac.il). SFmap implements the COS(WR) algorithm, which computes similarity scores for a given regulatory motif based on information derived from its sequence environment and its evolutionary conservation. Input for SFmap is a human genomic sequence or a list of sequences in FASTA format that can either be uploaded from a file or pasted into a window. SFmap searches within a given sequence for significant hits of binding motifs that are either stored in our database or defined by the user. SFmap results are provided both as a text file and as a graphical web interface.

Target-agnostic drug prediction integrated with medical record analysis uncovers differential associations of statins with increased survival in COVID-19 patients.

Importance: Drug repurposing requires distinguishing established drug class targets from novel molecule-specific mechanisms and rapidly derisking their therapeutic potential in a time-critical manner, particularly in a pandemic scenario. In response to the challenge to rapidly identify treatment options for COVID-19, several studies reported that statins, as a drug class, reduce mortality in these patients. However, it is unknown if different statins exhibit consistent function or may have varying therapeutic benefit. Objectives: To test if different statins differ in their ability to exert protective effects based on molecular computational predictions and electronic medical record analysis. Main Outcomes and Measures: A Bayesian network tool was used to predict drugs that shift the host transcriptomic response to SARS-CoV-2 infection towards a healthy state. Drugs were predicted using 14 RNA-sequencing datasets from 72 autopsy tissues and 465 COVID-19 patient samples or from cultured human cells and organoids infected with SARS-CoV-2, with a total of 2,436 drugs investigated. Top drug predictions included statins, which were then assessed using electronic medical records containing over 4,000 COVID-19 patients on statins to determine mortality risk in patients prescribed specific statins versus untreated matched controls. The same drugs were tested in Vero E6 cells infected with SARS-CoV-2 and human endothelial cells infected with a related OC43 coronavirus. Results: Simvastatin was among the most highly predicted compounds (14/14 datasets) and five other statins, including atorvastatin, were predicted to be active in > 50% of analyses. Analysis of the clinical database revealed that reduced mortality risk was only observed in COVID-19 patients prescribed a subset of statins, including simvastatin and atorvastatin. In vitro testing of SARS-CoV-2 infected cells revealed simvastatin to be a potent direct inhibitor whereas most other statins were less effective. Simvastatin also inhibited OC43 infection and reduced cytokine production in endothelial cells. Conclusions and Relevance: Different statins may differ in their ability to sustain the lives of COVID-19 patients despite having a shared drug target and lipid-modifying mechanism of action. These findings highlight the value of target-agnostic drug prediction coupled with patient databases to identify and clinically evaluate non-obvious mechanisms and derisk and accelerate drug repurposing opportunities.

ARTS, the unusual septin: structural and functional aspects.

The human Septin 4 gene (Sept4) encodes two major protein isoforms; Sept4_i1 (H5/PNUTL2) and Sept4_i2/ARTS. Septins have been traditionally studied for their role in cytokinesis and their filament-forming abilities, but subsequently have been implicated in diverse functions, including membrane dynamics, cytoskeletal reorganization, vesicle trafficking, and tumorigenesis. ARTS is localized at mitochondria and promotes programmed cell death (apoptosis). These features distinguish ARTS from any other known human septin family member. This review compares the structural and functional properties of ARTS with other septins. In addition, it describes how a combination of two distinct promoters, differential splicing, and intron retention leads to the generation of two different Sept4 variants with diverse biological activity.

Decreased Mortality Rate Among COVID-19 Patients Prescribed Statins: Data From Electronic Health Records in the US.

The severe respiratory illness due to SARS-CoV-2, the virus responsible for coronavirus disease 2019 (COVID-19), is triggered by an intense pro-inflammatory host response. Statins, prescribed primarily for lipid reduction, are known to have anti-inflammatory and immunomodulatory properties and have been associated with a reduced mortality rate among COVID-19 patients taking statins as reported in two recent retrospective studies. However, a meta-analysis that included nine studies showed that statin use did not improve in-hospital outcomes of those with COVID-19. In addition, concerns regarding the use of statins and an increase in COVID-19 infections have been raised, as statins may increase the expression of angiotensin-converting enzyme 2 (ACE2), the primary receptor for the SARS-CoV-2 virus. Our goal was to investigate the effect of statins in COVID-19 patients in a large, diverse patient population across the United States containing nearly 120,000 patients diagnosed with COVID-19. We used propensity score matching of demographics, comorbidities, and medication indication to compare statin-treated patients (N = 2,297) with matched controls (N = 4,594). We observed a small, but statistically significant, decrease in mortality among patients prescribed statins (16.1%) when compared with matched COVID-19-positive controls (18.0 to 20.6%). These results support previous evidence that statins do not increase COVID-19-related mortality and may, in fact, have a mitigating effect on severity of the disease reflected in a slight reduction in mortality. Mixed findings on effects of statins in COVID-19 patients reported in the literature should prompt prospective randomized controlled trials in order to define better who might be advantaged with respect to clinical outcomes.