Consensus Pituitary Atlas, a scalable resource for annotation, novel marker discovery, and analyses in mouse pituitary gland research.

Previous single-cell profiling studies of the pituitary gland have yielded minimally reproducible insights due to their low statistical power and methodological inconsistencies. To address this, we generate the uniformly pre-processed Consensus Pituitary Atlas (CPA) using all 283 existing mouse pituitary single-cell datasets (∼1.3 million high-quality cells). The CPA reveals cell typing and lineage markers, including low-expression transcripts that previous analyses could not detect. Leveraging the scale of the CPA, we develop machine learning models to automate and standardize cell type annotation and doublet identification for future studies. Utilizing the curated metadata, we identify sex-biased and age-dependent gene expression patterns at cell type resolution. To uncover drivers of cell fates, first we determine consensus cell communication patterns. Second, we use RNA sequencing and chromatin accessibility data to identify transcription factors associated with cell fates across modalities. The epitome platform provides a user-friendly interface with the CPA and allows streamlined analyses.