Development and implementation of an integrated preclinical atherosclerosis database.

Background: Basic scientists have used preclinical animal models to explore mechanisms driving human diseases for decades, resulting in thousands of publications, each supporting causative inferences. Despite substantial advances in the mechanistic construct of disease, there has been limited translation from individual studies to advances in clinical care. An integrated approach to these individual studies has the potential to improve translational success.

Methods: Using atherosclerosis as a test case, we extracted data from the two most common mouse models of atherosclerosis (ApoE and LDLR knockout). We restricted analyses to manuscripts published in two well-established journals, and , as of query in 2021. Predefined variables including experimental conditions, intervention and outcomes were extracted from each publication to produce a preclinical atherosclerosis database.

Results: Extracted data include animal sex, diet, intervention type and distinct plaque pathologies (size, inflammation, lipid content). Procedures are provided to standardize data extraction, attribute interventions to specific genes and transform the database for use with available transcriptomics software. The database integrates hundreds of genes, each directly tested in vivo for causation in a murine atherosclerosis model. The database is provided to allow the research community to perform integrated analyses that reflect the global impact of decades of atherosclerosis investigation.

Conclusions: Future database uses include interrogation of sub-datasets associated with distinct plaque pathologies, cell-type or sex. We provide the methods and software needed to apply this approach to the extensive repository of peer-reviewed data utilizing preclinical models to interrogate mechanisms of diverse human diseases.