Synthetic Health Data Can Augment Community Research Efforts to Better Inform the Public During Emerging Pandemics.

The COVID-19 pandemic had disproportionate effects on the Veteran population due to the increased prevalence of medical and environmental risk factors. Synthetic electronic health record (EHR) data can help meet the acute need for Veteran population-specific predictive modeling efforts by avoiding the strict barriers to access, currently present within Veteran Health Administration (VHA) datasets. The U.S. Food and Drug Administration (FDA) and the VHA launched the precisionFDA COVID-19 Risk Factor Modeling Challenge to develop COVID-19 diagnostic and prognostic models; identify Veteran population-specific risk factors; and test the usefulness of synthetic data as a substitute for real data. The use of synthetic data boosted challenge participation by providing a dataset that was accessible to all competitors. Models trained on synthetic data showed similar but systematically inflated model performance metrics to those trained on real data. The important risk factors identified in the synthetic data largely overlapped with those identified from the real data, and both sets of risk factors were validated in the literature. Tradeoffs exist between synthetic data generation approaches based on whether a real EHR dataset is required as input. Synthetic data generated directly from real EHR input will more closely align with the characteristics of the relevant cohort. This work shows that synthetic EHR data will have practical value to the Veterans' health research community for the foreseeable future.

A Machine Learning Pipeline for Accurate COVID-19 Health Outcome Prediction using Longitudinal Electronic Health Records.

Current COVID-19 predictive models primarily focus on predicting the risk of mortality, and rely on COVID-19 specific medical data such as chest imaging after COVID-19 diagnosis. In this project, we developed an innovative supervised machine learning pipeline using longitudinal Electronic Health Records (EHR) to accurately predict COVID-19 related health outcomes including mortality, ventilation, days in hospital or ICU. In particular, we developed unique and effective data processing algorithms, including data cleaning, initial feature screening, vector representation. Then we trained models using state-of-the-art machine learning strategies combined with different parameter settings. Based on routinely collected EHR, our machine learning pipeline not only consistently outperformed those developed by other research groups using the same set of data, but also achieved similar accuracy as those trained on medical data that were only available after COVID-19 diagnosis. In addition, top risk factors for COVID-19 were identified, and are consistent with epidemiologic findings.

A mixed methods approach to exploring the relationship between Norway rat (Rattus norvegicus) abundance and features of the urban environment in an inner-city neighborhood of Vancouver, Canada.

Urban rats (Rattus spp.) are among the most ubiquitous pest species in the world. Previous research has shown that rat abundance is largely determined by features of the environment; however, the specific urban environmental factors that influence rat population density within cities have yet to be clearly identified. Additionally, there are no well described tools or methodologies for conducting an in-depth evaluation of the relationship between urban rat abundance and the environment. In this study, we developed a systematic environmental observation tool using methods borrowed from the field of systematic social observation. This tool, which employed a combination of quantitative and qualitative methodologies, was then used to identify environmental factors associated with the relative abundance of Norway rats (Rattus norvegicus) in an inner-city neighborhood of Vancouver, Canada. Using a multivariate zero-inflated negative binomial model, we found that a variety of factors, including specific land use, building condition, and amount of refuse, were related to rat presence and abundance. Qualitative data largely supported and further clarified observed statistical relationships, but also identified conflicting and unique situations not easily captured through quantitative methods. Overall, the tool helped us to better understand the relationship between features of the urban environment and relative rat abundance within our study area and may useful for studying environmental determinants of zoonotic disease prevalence/distribution among urban rat populations in the future.

The characteristics of wild rat (Rattus spp.) populations from an inner-city neighborhood with a focus on factors critical to the understanding of rat-associated zoonoses.

Norway and black rats (Rattus norvegicus and Rattus rattus) are among the most ubiquitous urban wildlife species and are the source of a number of zoonotic diseases responsible for significant human morbidity and mortality in cities around the world. Rodent ecology is a primary determinant of the dynamics of zoonotic pathogens in rodent populations and the risk of pathogen transmission to people, yet many studies of rat-associated zoonoses do not account for the ecological characteristics of urban rat populations. This hinders the development of an in-depth understanding of the ecology of rat-associated zoonoses, limits comparability among studies, and can lead to erroneous conclusions. We conducted a year-long trapping-removal study to describe the ecological characteristics of urban rat populations in an inner-city neighborhood of Vancouver, Canada. The study focused on factors that might influence the ecology of zoonotic pathogens in these populations and/or our understanding of that ecology. We found that rat population density varied remarkably over short geographical distances, which could explain observed spatial distributions of rat-associated zoonoses and have implications for sampling and data analysis during research and surveillance. Season appeared to influence rat population composition even within the urban environment, which could cause temporal variation in pathogen prevalence. Body mass and bite wounds, which are often used in epidemiologic analyses as simple proxies for age and aggression, were shown to be more complex than previously thought. Finally, we found that factors associated with trapping can determine the size and composition of sampled rat population, and thus influence inferences made about the source population. These findings may help guide future studies of rats and rat-associated zoonoses.

Ecology of Leptospira interrogans in Norway rats (Rattus norvegicus) in an inner-city neighborhood of Vancouver, Canada.

BACKGROUND: Leptospira interrogans is a bacterial zoonosis with a worldwide distribution for which rats (Rattus spp.) are the primary reservoir in urban settings. In order to assess, monitor, and mitigate the risk to humans, it is important to understand the ecology of this pathogen in rats. The objective of this study was to characterize the ecology of L. interrogans in Norway rats (Rattus norvegicus) in an impoverished inner-city neighborhood of Vancouver, Canada. METHODOLOGY/PRINCIPAL FINDINGS: Trapping was performed in 43 city blocks, and one location within the adjacent port, over a 12 month period. Kidney samples were tested for the presence of L. interrogans using PCR and sequencing. A multivariable model was built to predict L. interrogans infection status in individual rats using season and morphometric data (e.g., weight, sex, maturity, condition, etc.) as independent variables. Spatial analysis was undertaken to identify clusters of high and low L. interrogans prevalence. The prevalence of L. interrogans varied remarkably among blocks (0-66.7%), and spatial clusters of both high and low L. interrogans prevalence were identified. In the final cluster-controlled model, characteristics associated with L. interrogans-infection in rats included weight (OR = 1.14, 95% CI = 1.07-1.20), increased internal fat (OR = 2.12, 95% CI = 1.06-4.25), and number of bite wounds (OR = 1.20, 95% CI = 0.96-1.49). CONCLUSIONS/SIGNIFICANCE: Because L. interrogans prevalence varied with weight, body fat, and bite wounds, this study suggests that social structure and interactions among rats may influence transmission. The prevalence and distribution of L. interrogans in rats was also highly variable even over a short geographic distance. These factors should be considered in future risk management efforts.

Distinct nuclear and cytoplasmic functions of androgen receptor cofactor p44 and association with androgen-independent prostate cancer.

Androgen receptor (AR) mediates transcriptional activation of diverse target genes through interactions with various coactivators that may alter its function and help mediate the switch between prostate cell proliferation and differentiation. We recently identified p44/MEP50 as an AR coactivator and further showed that it is expressed primarily in the nucleus and cytoplasm of benign prostate epithelial and prostate cancer cells, respectively. We also showed that haploinsufficiency in p44(+/-) mice causes prostate epithelial cell proliferation. To establish direct cause-and-effect relationships, we have used p44 fusion proteins that are selectively expressed in the nucleus or cytoplasm of prostate cancer cells (LNCaP), along with RNAi analyses, to examine effects of p44 both in vitro and in vivo (in tumor xenografts). We show that preferential expression of p44 in the nucleus inhibits proliferation of LNCaP cells in an AR-dependent manner, whereas preferential expression of p44 in the cytoplasm enhances cell proliferation. These effects appear to be mediated, at least in part, through the regulation of distinct cell-cycle regulatory genes that include p21 (up-regulated by nuclear p44) and cyclin D2 and CDK6 (up-regulated by cytoplasmic p44). Importantly, we also demonstrate that altered p44 expression is associated with androgen-independent prostate cancer. Our results indicate that nuclear p44 and cytoplasmic p44 have distinct and opposing functions in the regulation of prostate cancer cell proliferation.

Phenotypic manifestations of MECP2 mutations in classical and atypical Rett syndrome.

Since the identification of mutations in MECP2 in girls and women with apparent Rett syndrome, numerous efforts have been made to develop phenotype-genotype correlations. These studies have produced conflicting results in part related to use of different clinical severity scales, different diagnostic criteria, and different stratification by age and mutation group as well as the possible effects of unbalanced X-chromosome inactivation. The present study applied a revised ordinal scoring system that allowed for correction for patient ages. We analyzed 85 patients with mutation in MECP2. Sixty-five (76%) had one of eight common mutations. Patients with missense mutations had lower total severity scores and better language performance than those with nonsense mutations. No difference was noted between severity scores for mutations in the methyl-binding domain (MBD) and the transcriptional repression domain (TRD). However, patients with missense mutations in TRD had the best overall scores and better preservation of head growth and language skills. Analysis of specific mutation groups demonstrated a striking difference for patients with the R306C mutation including better overall score, later regression, and better language with less motoric impairment. Indeed, these patients as a group accounted for the differences in overall scores between the missense and nonsense groups. Thus, the impact of specific mutations coupled with possible variation in X-chromosome inactivation must be considered carefully in the derivation of phenotype-genotype correlations. These results emphasize the limitations of such analyses in larger mutation groups, either by type or position.