Predicting clinical outcomes of SARS-CoV-2 infection during the Omicron wave using machine learning.

The Omicron SARS-CoV-2 variant continues to strain healthcare systems. Developing tools that facilitate the identification of patients at highest risk of adverse outcomes is a priority. The study objectives are to develop population-scale predictive models that: 1) identify predictors of adverse outcomes with Omicron surge SARS-CoV-2 infections, and 2) predict the impact of prioritized vaccination of high-risk groups for said outcome. We prepared a retrospective longitudinal observational study of a national cohort of 172,814 patients in the U.S. Veteran Health Administration who tested positive for SARS-CoV-2 from January 15 to August 15, 2022. We utilized sociodemographic characteristics, comorbidities, and vaccination status, at time of testing positive for SARS-CoV-2 to predict hospitalization, escalation of care (high-flow oxygen, mechanical ventilation, vasopressor use, dialysis, or extracorporeal membrane oxygenation), and death within 30 days. Machine learning models demonstrated that advanced age, high comorbidity burden, lower body mass index, unvaccinated status, and oral anticoagulant use were the important predictors of hospitalization and escalation of care. Similar factors predicted death. However, anticoagulant use did not predict mortality risk. The all-cause death model showed the highest discrimination (Area Under the Curve (AUC) = 0.903, 95% Confidence Interval (CI): 0.895, 0.911) followed by hospitalization (AUC = 0.822, CI: 0.818, 0.826), then escalation of care (AUC = 0.793, CI: 0.784, 0.805). Assuming a vaccine efficacy range of 70.8 to 78.7%, our simulations projected that targeted prevention in the highest risk group may have reduced 30-day hospitalization and death in more than 2 of 5 unvaccinated patients.

Development and external validation of a pretrained deep learning model for the prediction of non-accidental trauma.

Non-accidental trauma (NAT) is deadly and difficult to predict. Transformer models pretrained on large datasets have recently produced state of the art performance on diverse prediction tasks, but the optimal pretraining strategies for diagnostic predictions are not known. Here we report the development and external validation of Pretrained and Adapted BERT for Longitudinal Outcomes (PABLO), a transformer-based deep learning model with multitask clinical pretraining, to identify patients who will receive a diagnosis of NAT in the next year. We develop a clinical interface to visualize patient trajectories, model predictions, and individual risk factors. In two comprehensive statewide databases, approximately 1% of patients experience NAT within one year of prediction. PABLO predicts NAT events with area under the receiver operating characteristic curve (AUROC) of 0.844 (95% CI 0.838-0.851) in the California test set, and 0.849 (95% CI 0.846-0.851) on external validation in Florida, outperforming comparator models. Multitask pretraining significantly improves model performance. Attribution analysis shows substance use, psychiatric, and injury diagnoses, in the context of age and racial demographics, as influential predictors of NAT. As a clinical decision support system, PABLO can identify high-risk patients and patient-specific risk factors, which can be used to target secondary screening and preventive interventions at the point-of-care.

Profiling chromatin accessibility responses in human neutrophils with sensitive pathogen detection.

Sepsis, sequela of bloodstream infections and dysregulated host responses, is a leading cause of death globally. Neutrophils tightly regulate responses to pathogens to prevent organ damage. Profiling early host epigenetic responses in neutrophils may aid in disease recognition. We performed assay for transposase-accessible chromatin (ATAC)-seq of human neutrophils challenged with six toll-like receptor ligands and two organisms; and RNA-seq after Escherichia coli exposure for 1 and 4 h along with ATAC-seq. ATAC-seq of neutrophils facilitates detection of pathogen DNA. In addition, despite similarities in genomic distribution of differential chromatin changes across challenges, only a fraction overlaps between the challenges. Ligands depict shared signatures, but majority are unique in position, function, and challenge. Epigenomic changes are plastic, only ∼120 are shared by E coli challenges over time, resulting in varied differential genes and associated processes. We identify three classes of gene regulation, chromatin access changes in the promoter; changes in the promoter and distal enhancers; and controlling expression through changes solely in distal enhancers. These and transcription factor footprinting reveal timely and challenge specific mechanisms of transcriptional regulation in neutrophils.

Comparative transcriptome profiling provides insights into plant salt tolerance in seashore paspalum (Paspalum vaginatum).

BACKGROUND: Seashore paspalum (Paspalum vaginatum), a halophytic warm-seasoned perennial grass, is tolerant of many environmental stresses, especially salt stress. To investigate molecular mechanisms underlying salinity tolerance in seashore paspalum, physiological characteristics and global transcription profiles of highly (Supreme) and moderately (Parish) salinity-tolerant cultivars under normal and salt stressed conditions were analyzed. RESULTS: Physiological characterization comparing highly (Supreme) and moderately (Parish) salinity-tolerant cultivars revealed that Supreme's higher salinity tolerance is associated with higher Na+ and Ca2+ accumulation under normal conditions and further increase of Na+ under salt-treated conditions (400 mM NaCl), possibly by vacuolar sequestration. Moreover, K+ retention under salt treatment occurs in both cultivars, suggesting that it may be a conserved mechanism for prevention of Na+ toxicity. We sequenced the transcriptome of the two cultivars under both normal and salt-treated conditions (400 mM NaCl) using RNA-seq. De novo assembly of about 153 million high-quality reads and identification of Open Reading Frames (ORFs) uncovered a total of 82,608 non-redundant unigenes, of which 3250 genes were identified as transcription factors (TFs). Gene Ontology (GO) annotation revealed the presence of genes involved in diverse cellular processes in seashore paspalum's transcriptome. Differential expression analysis identified a total of 828 and 2222 genes that are responsive to high salinity for Supreme and Parish, respectively. "Oxidation-reduction process" and "nucleic acid binding" are significantly enriched GOs among differentially expressed genes in both cultivars under salt treatment. Interestingly, compared to Parish, a number of salt stress induced transcription factors are enriched and show higher abundance in Supreme under normal conditions, possibly due to enhanced Ca2+ signaling transduction out of Na+ accumulation, which may be another contributor to Supreme's higher salinity tolerance. CONCLUSION: Physiological and transcriptome analyses of seashore paspalum reveal major molecular underpinnings contributing to plant response to salt stress in this halophytic warm-seasoned perennial grass. The data obtained provide valuable molecular resources for functional studies and developing strategies to engineer plant salinity tolerance.

Genomic data mining for functional annotation of human long noncoding RNAs.

Life may have begun in an RNA world, which is supported by increasing evidence of the vital role that RNAs perform in biological systems. In the human genome, most genes actually do not encode proteins; they are noncoding RNA genes. The largest class of noncoding genes is known as long noncoding RNAs (lncRNAs), which are transcripts greater in length than 200 nucleotides, but with no protein-coding capacity. While some lncRNAs have been demonstrated to be key regulators of gene expression and 3D genome organization, most lncRNAs are still uncharacterized. We thus propose several data mining and machine learning approaches for the functional annotation of human lncRNAs by leveraging the vast amount of data from genetic and genomic studies. Recent results from our studies and those of other groups indicate that genomic data mining can give insights into lncRNA functions and provide valuable information for experimental studies of candidate lncRNAs associated with human disease.

Co-expression of long non-coding RNAs and autism risk genes in the developing human brain.

BACKGROUND: Autism Spectrum Disorder (ASD) is the umbrella term for a group of neurodevelopmental disorders convergent on behavioral phenotypes. While many genes have been implicated in the disorder, the predominant focus of previous research has been on protein coding genes. This leaves a vast number of long non-coding RNAs (lncRNAs) not characterized for their role in the disorder although lncRNAs have been shown to play important roles in development and are highly represented in the brain. Studies have also shown lncRNAs to be differentially expressed in ASD affected brains. However, there has yet to be an enrichment analysis of the shared ontologies and pathways of known ASD genes and lncRNAs in normal brain development. RESULTS: In this study, we performed co-expression network analysis on the developing brain transcriptome to identify potential lncRNAs associated with ASD and possible annotations for functional role of lncRNAs in brain development. We found co-enrichment of lncRNA genes and ASD risk genes in two distinct groups of modules showing elevated prenatal and postnatal expression patterns, respectively. Further enrichment analysis of the module groups indicated that the early expression modules were comprised mainly of transcriptional regulators while the later expression modules were associated with synapse formation. Finally, lncRNAs were prioritized for their connectivity with the known ASD risk genes through analysis of an adjacency matrix. Collectively, the results imply early developmental repression of synaptic genes through lncRNAs and ASD transcriptional regulators. CONCLUSION: Here we demonstrate the utility of mining the publically available brain gene expression data to further functionally annotate the role of lncRNAs in ASD. Our analysis indicates that lncRNAs potentially have a key role in ASD due to their convergence on shared pathways, and we identify lncRNAs of interest that may lead to further avenues of study.

A "Culture" Shift: Broad Bacterial Detection, Identification, and Antimicrobial Susceptibility Testing Directly from Whole Blood.

BACKGROUND: The time required for bloodstream pathogen detection, identification (ID), and antimicrobial susceptibility testing (AST) does not satisfy the acute needs of disease management. Conventional methods take up to 3 days for ID and AST. Molecular diagnostics have reduced times for ID, but their promise to supplant culture is unmet because AST times remain slow. We developed a combined quantitative PCR (qPCR)-based ID+AST assay with sequential detection, ID, and AST of leading nosocomial bacterial pathogens. METHODS: ID+AST was performed on whole blood samples by (a) removing blood cells, (b) brief bacterial enrichment, (c) bacterial detection and ID, and (d) species-specific antimicrobial treatment. Broad-spectrum qPCR of the internal transcribed spacer between the 16S and 23S was amplified for detection. High-resolution melting identified the species with a curve classifier. AST was enabled by Ct differences between treated and untreated samples. RESULTS: A detection limit of 1 CFU/mL was achieved for Acinetobacter baumannii, Escherichia coli, Klebsiella pneumoniae, and Staphylococcus aureus. All species were accurately identified by unique melting curves. Antimicrobial minimum inhibitory concentrations were identified with Ct differences of ≥1 cycle. Using an RNA target allowed reduction of AST incubation time from 60 min to 5 min. Rapid-cycle amplification reduced qPCR times by 83% to 30 min. CONCLUSIONS: Combined, sequential ID+AST protocols allow rapid and reliable detection, ID, and AST for the diagnosis of bloodstream infections, enabling conversion of empiric to targeted therapy by the second dose of antimicrobials.

A review on cluster estimation methods and their application to neural spike data.

The extracellular action potentials recorded on an electrode result from the collective simultaneous electrophysiological activity of an unknown number of neurons. Identifying and assigning these action potentials to their firing neurons-'spike sorting'-is an indispensable step in studying the function and the response of an individual or ensemble of neurons to certain stimuli. Given the task of neural spike sorting, the determination of the number of clusters (neurons) is arguably the most difficult and challenging issue, due to the existence of background noise and the overlap and interactions among neurons in neighbouring regions. It is not surprising that some researchers still rely on visual inspection by experts to estimate the number of clusters in neural spike sorting. Manual inspection, however, is not suitable to processing the vast, ever-growing amount of neural data. To address this pressing need, in this paper, thirty-three clustering validity indices have been comprehensively reviewed and implemented to determine the number of clusters in neural datasets. To gauge the suitability of the indices to neural spike data, and inform the selection process, we then calculated the indices by applying k-means clustering to twenty widely used synthetic neural datasets and one empirical dataset, and compared the performance of these indices against pre-existing ground truth labels. The results showed that the top five validity indices work consistently well across variations in noise level, both for the synthetic datasets and the real dataset. Using these top performing indices provides strong support for the determination of the number of neural clusters, which is essential in the spike sorting process.

Microbial Typing by Machine Learned DNA Melt Signatures.

There is still an ongoing demand for a simple broad-spectrum molecular diagnostic assay for pathogenic bacteria. For this purpose, we developed a single-plex High Resolution Melt (HRM) assay that generates complex melt curves for bacterial identification. Using internal transcribed spacer (ITS) region as the phylogenetic marker for HRM, we observed complex melt curve signatures as compared to 16S rDNA amplicons with enhanced interspecies discrimination. We also developed a novel Naïve Bayes curve classification algorithm with statistical interpretation and achieved 95% accuracy in differentiating 89 bacterial species in our library using leave-one-out cross-validation. Pilot clinical validation of our method correctly identified the etiologic organisms at the species-level in 59 culture-positive mono-bacterial blood culture samples with 90% accuracy. Our findings suggest that broad bacterial sequences may be simply, reliably and automatically profiled by ITS HRM assay for clinical adoption.

Co-expression Network Analysis of Human lncRNAs and Cancer Genes.

We used gene co-expression network analysis to functionally annotate long noncoding RNAs (lncRNAs) and identify their potential cancer associations. The integrated microarray data set from our previous study was used to extract the expression profiles of 1,865 lncRNAs. Known cancer genes were compiled from the Catalogue of Somatic Mutations in Cancer and UniProt databases. Co-expression analysis identified a list of previously uncharacterized lncRNAs that showed significant correlation in expression with core cancer genes. To further annotate the lncRNAs, we performed a weighted gene co-expression network analysis, which resulted in 37 co-expression modules. Three biologically interesting modules were analyzed in depth. Two of the modules showed relatively high expression in blood and brain tissues, whereas the third module was found to be downregulated in blood cells. Hub lncRNA genes and enriched functional annotation terms were identified within the modules. The results suggest the utility of this approach as well as potential roles of uncharacterized lncRNAs in leukemia and neuroblastoma.