Predicting hospitalisations related to ambulatory care sensitive conditions with machine learning for population health planning: derivation and validation cohort study.

OBJECTIVE: To predict older adults' risk of avoidable hospitalisation related to ambulatory care sensitive conditions (ACSC) using machine learning applied to administrative health data of Ontario, Canada. DESIGN, SETTING AND PARTICIPANTS: A retrospective cohort study was conducted on a large cohort of all residents covered under a single-payer system in Ontario, Canada over the period of 10 years (2008-2017). The study included 1.85 million Ontario residents between 65 and 74 years old at any time throughout the study period. DATA SOURCES: Administrative health data from Ontario, Canada obtained from the (ICES formely known as the Institute for Clinical Evaluative Sciences Data Repository. MAIN OUTCOME MEASURES: Risk of hospitalisations due to ACSCs 1 year after the observation period. RESULTS: The study used a total of 1 854 116 patients, split into train, validation and test sets. The ACSC incidence rates among the data points were 1.1% for all sets. The final XGBoost model achieved an area under the receiver operating curve of 80.5% and an area under precision-recall curve of 0.093 on the test set, and the predictions were well calibrated, including in key subgroups. When ranking the model predictions, those at the top 5% of risk as predicted by the model captured 37.4% of those presented with an ACSC-related hospitalisation. A variety of features such as the previous number of ambulatory care visits, presence of ACSC-related hospitalisations during the observation window, age, rural residence and prescription of certain medications were contributors to the prediction. Our model was also able to capture the geospatial heterogeneity of ACSC risk in Ontario, and especially the elevated risk in rural and marginalised regions. CONCLUSIONS: This study aimed to predict the 1-year risk of hospitalisation from ambulatory-care sensitive conditions in seniors aged 65-74 years old with a single, large-scale machine learning model. The model shows the potential to inform population health planning and interventions to reduce the burden of ACSC-related hospitalisations.

Risk stratification for COVID-19 hospitalization: a multivariable model based on gradient-boosting decision trees.

BACKGROUND: The COVID-19 pandemic has led to an increased demand for health care resources and, in some cases, shortage of medical equipment and staff. Our objective was to develop and validate a multivariable model to predict risk of hospitalization for patients infected with SARS-CoV-2. METHODS: We used routinely collected health records in a patient cohort to develop and validate our prediction model. This cohort included adult patients (age ≥ 18 yr) from Ontario, Canada, who tested positive for SARS-CoV-2 ribonucleic acid by polymerase chain reaction between Feb. 2 and Oct. 5, 2020, and were followed up through Nov. 5, 2020. Patients living in long-term care facilities were excluded, as they were all assumed to be at high risk of hospitalization for COVID-19. Risk of hospitalization within 30 days of diagnosis of SARS-CoV-2 infection was estimated via gradient-boosting decision trees, and variable importance examined via Shapley values. We built a gradient-boosting model using the Extreme Gradient Boosting (XGBoost) algorithm and compared its performance against 4 empirical rules commonly used for risk stratifications based on age and number of comorbidities. RESULTS: The cohort included 36 323 patients with 2583 hospitalizations (7.1%). Hospitalized patients had a higher median age (64 yr v. 43 yr), were more likely to be male (56.3% v. 47.3%) and had a higher median number of comorbidities (3, interquartile range [IQR] 2-6 v. 1, IQR 0-3) than nonhospitalized patients. Patients were split into development (n = 29 058, 80.0%) and held-out validation (n = 7265, 20.0%) cohorts. The gradient-boosting model achieved high discrimination (development cohort: area under the receiver operating characteristic curve across the 5 folds of 0.852; validation cohort: 0.8475) and strong calibration (slope = 1.01, intercept = -0.01). The patients who scored at the top 10% captured 47.4% of hospitalizations, and those who scored at the top 30% captured 80.6%. INTERPRETATION: We developed and validated an accurate risk stratification model using routinely collected health administrative data. We envision that modelling such risk stratification based on routinely collected health data could support management of COVID-19 on a population health level.

Model-based assessment of mammalian cell metabolic functionalities using omics data.

Omics experiments are ubiquitous in biological studies, leading to a deluge of data. However, it is still challenging to connect changes in these data to changes in cell functions because of complex interdependencies between genes, proteins, and metabolites. Here, we present a framework allowing researchers to infer how metabolic functions change on the basis of omics data. To enable this, we curated and standardized lists of metabolic tasks that mammalian cells can accomplish. Genome-scale metabolic networks were used to define gene sets associated with each metabolic task. We further developed a framework to overlay omics data on these sets and predict pathway usage for each metabolic task. We demonstrated how this approach can be used to quantify metabolic functions of diverse biological samples from the single cell to whole tissues and organs by using multiple transcriptomic datasets. To facilitate its adoption, we integrated the approach into GenePattern (www.genepattern.org-CellFie).

Cardiovascular adverse events associated with BRAF versus BRAF/MEK inhibitor: Cross-sectional and longitudinal analysis using two large national registries.

BACKGROUND: Cardiovascular adverse events (CVAEs) associated with BRAF inhibitors alone versus combination BRAF/MEK inhibitors are not fully understood. METHODS: This study included all adult patients who received BRAF inhibitors (vemurafenib, dabrafenib, encorafenib) or combinations BRAF/MEK inhibitors (vemurafenib/cobimetinib; dabrafenib/trametinib; encorafenib/binimetinib). We utilized the cross-sectional FDA's Adverse Events Reporting System (FAERS) and longitudinal Truven Health Analytics/IBM MarketScan database from 2011 to 2018. Various CVAEs, including arterial hypertension, heart failure (HF), and venous thromboembolism (VTE), were studied using adjusted regression techniques. RESULTS: In FAERS, 7752 AEs were reported (40% BRAF and 60% BRAF/MEK). Median age was 60 (IQR 49-69) years with 45% females and 97% with melanoma. Among these, 567 (7.4%) were cardiovascular adverse events (mortality rate 19%). Compared with monotherapy, combination therapy was associated with increased risk for HF (reporting odds ratio [ROR] = 1.62 (CI = 1.14-2.30); p = 0.007), arterial hypertension (ROR = 1.75 (CI = 1.12-2.89); p = 0.02) and VTE (ROR = 1.80 (CI = 1.12-2.89); p = 0.02). Marketscan had 657 patients with median age of 53 years (IQR 46-60), 39.3% female, and 88.7% with melanoma. There were 26.2% CVAEs (CI: 14.8%-36%) within 6 months of medication start in those receiving combination therapy versus 16.7% CVAEs (CI: 13.1%-20.2%) among those receiving monotherapy. Combination therapy was associated with CVAEs compared to monotherapy (adjusted HR: 1.56 (CI: 1.01-2.42); p = 0.045). CONCLUSIONS AND RELEVANCE: In two independent real-world cohorts, combination BRAF/MEK inhibitors were associated with increased CVAEs compared to monotherapy, especially HF, and hypertension.

Coordinate regulation of systemic and kidney tryptophan metabolism by the drug transporters OAT1 and OAT3.

How organs sense circulating metabolites is a key question. Here, we show that the multispecific organic anion transporters of drugs, OAT1 (SLC22A6 or NKT) and OAT3 (SLC22A8), play a role in organ sensing. Metabolomics analyses of the serum of Oat1 and Oat3 knockout mice revealed changes in tryptophan derivatives involved in metabolism and signaling. Several of these metabolites are derived from the gut microbiome and are implicated as uremic toxins in chronic kidney disease. Direct interaction with the transporters was supported with cell-based transport assays. To assess the impact of the loss of OAT1 or OAT3 function on the kidney, an organ where these uptake transporters are highly expressed, knockout transcriptomic data were mapped onto a "metabolic task"-based computational model that evaluates over 150 cellular functions. Despite the changes of tryptophan metabolites in both knockouts, only in the Oat1 knockout were multiple tryptophan-related cellular functions increased. Thus, deprived of the ability to take up kynurenine, kynurenate, anthranilate, and N-formylanthranilate through OAT1, the kidney responds by activating its own tryptophan-related biosynthetic pathways. The results support the Remote Sensing and Signaling Theory, which describes how "drug" transporters help optimize levels of metabolites and signaling molecules by facilitating organ cross talk. Since OAT1 and OAT3 are inhibited by many drugs, the data implies potential for drug-metabolite interactions. Indeed, treatment of humans with probenecid, an OAT-inhibitor used to treat gout, elevated circulating tryptophan metabolites. Furthermore, given that regulatory agencies have recommended drugs be tested for OAT1 and OAT3 binding or transport, it follows that these metabolites can be used as endogenous biomarkers to determine if drug candidates interact with OAT1 and/or OAT3.

Cardiovascular Events Associated with Chimeric Antigen Receptor T Cell Therapy: Cross-Sectional FDA Adverse Events Reporting System Analysis.

Chimeric antigen receptor (CAR) T cell therapy is approved in the United States for the treatment of acute lymphocytic leukemia and aggressive B cell lymphomas. Multiple cardiovascular adverse events (CVEs) associated with CAR-Ts have been observed in small studies, but no large-scale studies exist. Leveraging the Food and Drug Administration (FDA) Adverse Events Reporting System (FAERS), we identified all reported adverse events (AEs) associated with CAR-T therapy (tisagenlecleucel and axicabtagene ciloleucel) from 2017 to 2019. Reports with missing age and sex were excluded. CVEs were classified into arrhythmias, heart failure (HF), myocardial infarction (MI), and other CVEs. Logistic regression and hierarchical clustering were used to identify factors associated with CVEs. A total of 996 reported AEs were observed (39.1% associated with tisagenlecleucel and 60% with axicabtagene ciloleucel). Of all patients experiencing AEs, the median age was 54 (interquartile range, 21 to 65) years; 38.9% were females. In total, 19.7% (196) of all AEs reported to the FDA were CVEs. The most common CVEs were arrhythmia (77.6%), followed by HF (14.3%) and MI (0.5%). In adjusted analysis a positive association was observed between those presenting with CVE with neurotoxicity (odds ratio, 1.76; 95% confidence interval, 1.20 to 2.60; P = .004). Additionally, when both CVE and cytokine release syndrome (CRS) are present, neurotoxicity is the most common noncardiac AE, which clusters with them (Jaccard similarity: 73.1). The mortality rate was 21.1% overall but 30.1% for those reporting CVEs. In FAERS, reported CVEs with CAR-T are associated with high reported mortality. The development of either CRS or neurotoxicity should prompt vigilance for cardiovascular events.

Evolutionary and structural analyses of SARS-CoV-2 D614G spike protein mutation now documented worldwide.

The COVID-19 pandemic, caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), was declared on March 11, 2020 by the World Health Organization. As of the 31st of May, 2020, there have been more than 6 million COVID-19 cases diagnosed worldwide and over 370,000 deaths, according to Johns Hopkins. Thousands of SARS-CoV-2 strains have been sequenced to date, providing a valuable opportunity to investigate the evolution of the virus on a global scale. We performed a phylogenetic analysis of over 1,225 SARS-CoV-2 genomes spanning from late December 2019 to mid-March 2020. We identified a missense mutation, D614G, in the spike protein of SARS-CoV-2, which has emerged as a predominant clade in Europe (954 of 1,449 (66%) sequences) and is spreading worldwide (1,237 of 2,795 (44%) sequences). Molecular dating analysis estimated the emergence of this clade around mid-to-late January (10-25 January) 2020. We also applied structural bioinformatics to assess the potential impact of D614G on the virulence and epidemiology of SARS-CoV-2. In silico analyses on the spike protein structure suggests that the mutation is most likely neutral to protein function as it relates to its interaction with the human ACE2 receptor. The lack of clinical metadata available prevented our investigation of association between viral clade and disease severity phenotype. Future work that can leverage clinical outcome data with both viral and human genomic diversity is needed to monitor the pandemic.

Multiplex secretome engineering enhances recombinant protein production and purity.

Host cell proteins (HCPs) are process-related impurities generated during biotherapeutic protein production. HCPs can be problematic if they pose a significant metabolic demand, degrade product quality, or contaminate the final product. Here, we present an effort to create a "clean" Chinese hamster ovary (CHO) cell by disrupting multiple genes to eliminate HCPs. Using a model of CHO cell protein secretion, we predict that the elimination of unnecessary HCPs could have a non-negligible impact on protein production. We analyze the HCP content of 6-protein, 11-protein, and 14-protein knockout clones. These cell lines exhibit a substantial reduction in total HCP content (40%-70%). We also observe higher productivity and improved growth characteristics in specific clones. The reduced HCP content facilitates purification of a monoclonal antibody. Thus, substantial improvements can be made in protein titer and purity through large-scale HCP deletion, providing an avenue to increased quality and affordability of high-value biopharmaceuticals.

Genome-scale reconstructions of the mammalian secretory pathway predict metabolic costs and limitations of protein secretion.

In mammalian cells, >25% of synthesized proteins are exported through the secretory pathway. The pathway complexity, however, obfuscates its impact on the secretion of different proteins. Unraveling its impact on diverse proteins is particularly important for biopharmaceutical production. Here we delineate the core secretory pathway functions and integrate them with genome-scale metabolic reconstructions of human, mouse, and Chinese hamster ovary cells. The resulting reconstructions enable the computation of energetic costs and machinery demands of each secreted protein. By integrating additional omics data, we find that highly secretory cells have adapted to reduce expression and secretion of other expensive host cell proteins. Furthermore, we predict metabolic costs and maximum productivities of biotherapeutic proteins and identify protein features that most significantly impact protein secretion. Finally, the model successfully predicts the increase in secretion of a monoclonal antibody after silencing a highly expressed selection marker. This work represents a knowledgebase of the mammalian secretory pathway that serves as a novel tool for systems biotechnology.

Combating viral contaminants in CHO cells by engineering innate immunity.

Viral contamination in biopharmaceutical manufacturing can lead to shortages in the supply of critical therapeutics. To facilitate the protection of bioprocesses, we explored the basis for the susceptibility of CHO cells to RNA virus infection. Upon infection with certain ssRNA and dsRNA viruses, CHO cells fail to generate a significant interferon (IFN) response. Nonetheless, the downstream machinery for generating IFN responses and its antiviral activity is intact in these cells: treatment of cells with exogenously-added type I IFN or poly I:C prior to infection limited the cytopathic effect from Vesicular stomatitis virus (VSV), Encephalomyocarditis virus (EMCV), and Reovirus-3 virus (Reo-3) in a STAT1-dependent manner. To harness the intrinsic antiviral mechanism, we used RNA-Seq to identify two upstream repressors of STAT1: Gfi1 and Trim24. By knocking out these genes, the engineered CHO cells exhibited activation of cellular immune responses and increased resistance to the RNA viruses tested. Thus, omics-guided engineering of mammalian cell culture can be deployed to increase safety in biotherapeutic protein production among many other biomedical applications.

The emerging role of systems biology for engineering protein production in CHO cells.

To meet the ever-growing demand for effective, safe, and affordable protein therapeutics, decades of intense efforts have aimed to maximize the quantity and quality of recombinant proteins produced in CHO cells. Bioprocessing innovations and cell engineering efforts have improved product titer; however, uncharacterized cellular processes and gene regulatory mechanisms still hinder cell growth, specific productivity, and protein quality. Herein, we summarize recent advances in systems biology and data-driven approaches aiming to unravel how molecular pathways, cellular processes, and extrinsic factors (e.g. media supplementation) influence recombinant protein production. In particular, as the available omics data for CHO cells continue to grow, predictive models and screens will be increasingly used to unravel the biological drivers of protein production, which can be used with emerging genome editing technologies to rationally engineer cells to further control the quantity, quality and affordability of many biologic drugs.

A Consensus Genome-scale Reconstruction of Chinese Hamster Ovary Cell Metabolism.

Chinese hamster ovary (CHO) cells dominate biotherapeutic protein production and are widely used in mammalian cell line engineering research. To elucidate metabolic bottlenecks in protein production and to guide cell engineering and bioprocess optimization, we reconstructed the metabolic pathways in CHO and associated them with >1,700 genes in the Cricetulus griseus genome. The genome-scale metabolic model based on this reconstruction, iCHO1766, and cell-line-specific models for CHO-K1, CHO-S, and CHO-DG44 cells provide the biochemical basis of growth and recombinant protein production. The models accurately predict growth phenotypes and known auxotrophies in CHO cells. With the models, we quantify the protein synthesis capacity of CHO cells and demonstrate that common bioprocess treatments, such as histone deacetylase inhibitors, inefficiently increase product yield. However, our simulations show that the metabolic resources in CHO are more than three times more efficiently utilized for growth or recombinant protein synthesis following targeted efforts to engineer the CHO secretory pathway. This model will further accelerate CHO cell engineering and help optimize bioprocesses.

Recon 2.2: from reconstruction to model of human metabolism.

INTRODUCTION: The human genome-scale metabolic reconstruction details all known metabolic reactions occurring in humans, and thereby holds substantial promise for studying complex diseases and phenotypes. Capturing the whole human metabolic reconstruction is an on-going task and since the last community effort generated a consensus reconstruction, several updates have been developed. OBJECTIVES: We report a new consensus version, Recon 2.2, which integrates various alternative versions with significant additional updates. In addition to re-establishing a consensus reconstruction, further key objectives included providing more comprehensive annotation of metabolites and genes, ensuring full mass and charge balance in all reactions, and developing a model that correctly predicts ATP production on a range of carbon sources. METHODS: Recon 2.2 has been developed through a combination of manual curation and automated error checking. Specific and significant manual updates include a respecification of fatty acid metabolism, oxidative phosphorylation and a coupling of the electron transport chain to ATP synthase activity. All metabolites have definitive chemical formulae and charges specified, and these are used to ensure full mass and charge reaction balancing through an automated linear programming approach. Additionally, improved integration with transcriptomics and proteomics data has been facilitated with the updated curation of relationships between genes, proteins and reactions. RESULTS: Recon 2.2 now represents the most predictive model of human metabolism to date as demonstrated here. Extensive manual curation has increased the reconstruction size to 5324 metabolites, 7785 reactions and 1675 associated genes, which now are mapped to a single standard. The focus upon mass and charge balancing of all reactions, along with better representation of energy generation, has produced a flux model that correctly predicts ATP yield on different carbon sources. CONCLUSION: Through these updates we have achieved the most complete and best annotated consensus human metabolic reconstruction available, thereby increasing the ability of this resource to provide novel insights into normal and disease states in human. The model is freely available from the Biomodels database (http://identifiers.org/biomodels.db/MODEL1603150001).

Quantitative feature extraction from the Chinese hamster ovary bioprocess bibliome using a novel meta-analysis workflow.

The scientific literature concerning Chinese hamster ovary (CHO) cells grows annually due to the importance of CHO cells in industrial bioprocessing of therapeutics. In an effort to start to catalogue the breadth of CHO phenotypes, or phenome, we present the CHO bibliome. This bibliographic compilation covers all published CHO cell studies from 1995 to 2015, and each study is classified by the types of phenotypic and bioprocess data contained therein. Using data from selected studies, we also present a quantitative meta-analysis of bioprocess characteristics across diverse culture conditions, yielding novel insights and addressing the validity of long held assumptions. Specifically, we show that bioprocess titers can be predicted using indicator variables derived from viable cell density, viability, and culture duration. We further identified a positive correlation between the cumulative viable cell density (VCD) and final titer, irrespective of cell line, media, and other bioprocess parameters. In addition, growth rate was negatively correlated with performance attributes, such as VCD and titer. In summary, despite assumptions that technical diversity among studies and opaque publication practices can limit research re-use in this field, we show that the statistical analysis of diverse legacy bioprocess data can provide insight into bioprocessing capabilities of CHO cell lines used in industry. The CHO bibliome can be accessed at http://lewislab.ucsd.edu/cho-bibliome/.

Optimizing eukaryotic cell hosts for protein production through systems biotechnology and genome-scale modeling.

Eukaryotic cell lines, including Chinese hamster ovary cells, yeast, and insect cells, are invaluable hosts for the production of many recombinant proteins. With the advent of genomic resources, one can now leverage genome-scale computational modeling of cellular pathways to rationally engineer eukaryotic host cells. Genome-scale models of metabolism include all known biochemical reactions occurring in a specific cell. By describing these mathematically and using tools such as flux balance analysis, the models can simulate cell physiology and provide targets for cell engineering that could lead to enhanced cell viability, titer, and productivity. Here we review examples in which metabolic models in eukaryotic cell cultures have been used to rationally select targets for genetic modification, improve cellular metabolic capabilities, design media supplementation, and interpret high-throughput omics data. As more comprehensive models of metabolism and other cellular processes are developed for eukaryotic cell culture, these will enable further exciting developments in cell line engineering, thus accelerating recombinant protein production and biotechnology in the years to come.