Cross Study Analyses of SEND Data: Analytical and Visual Approaches

<p>The purpose of this collaborative work between BioCelerate and FDA is to develop SEND data transformation strategies and apply analytic techniques to enable an understanding of the differences between several data sets. Example use cases include understanding a single compound’s toxicity profile across all studies performed or evaluating on- versus off-target toxicity for multiple compounds intended for the same pharmacological target. Publicly available SEND data for four studies (rodent and non-rodent) were used to evaluate analytical approaches. The analyses of SEND data variables to integrate numerical data (e.g., body weight, BW). Initial BW analyses involved evaluation of time dependent effects using measured BWs versus time. This visualization enabled the assessment of growth curves and overall consistency of the data. Subsequent analyses involved normalizing the BW data to control and change from baseline. Visualization of BW relative to the control group enables identification of dose groups with the largest decreases in BW while change from baseline visualizations facilitate identification of dose groups that did not gain or lost weight. Integration of change from baseline and change from control data into a heat map enables clustering of dose groups that meet specified changes. Statistical evaluations are planned, and visualizations are being developed to allow for the comparison of the clustered BW and microscopic data across multiple studies. Cross-study comparisons using SEND data are expected to improve the identification of unique findings related to the intended target, species and duration of dosing.</p>