Hourly PhenoMaster Behavioral Phenotyping Dataset for Hormonal-state Classification in Mice
Description
This dataset contains raw and processed PhenoMaster behavioral phenotyping data used for machine-learning-based classification of sex and hormonal state in mice. The raw data include PhenoMaster exports for activity, indirect calorimetry, drinking and feeding, cumulative drinking and feeding, and running wheel behavior across five recording dates in study S003. The accompanying metadata file contains animal-level experimental annotations, including cohort, animal identifier, sex, gonadectomy/intact status, hormonal status, recording start and finish dates, iteration, and PhenoMaster box number. The processed dataset master_df_hourly.csv contains the hourly feature matrix generated from the raw exports by the associated Python analysis pipeline. Each row corresponds to one hourly Box × DateTime observation after timestamp parsing, wide-to-long transformation where required, numeric conversion, cleaning of technical values, hourly aggregation, alignment to valid calorimetry time slots, modality merging, metadata integration, and temporal feature generation. The dataset is intended to support reproducibility of the associated manuscript and GitHub repository. The associated code repository provides scripts for PhenoMaster export parsing, hourly aggregation, feature matrix generation, direct and hierarchical classification, UMAP analysis, SHAP analysis, minimum-days analysis, and export of result tables and figures.
Files
Institutions
- Institute of Biomedical ProblemsMoscow, Moscow