<p>Metrics table of scIB pipeline.</p>

To address this, we constructed a comprehensive single-cell atlas of tooth development, integrating data from 261,929 cells across 15 projects and encompassing both odontogenesis and amelogenesis. We developed a novel algorithm, TrajDTW, to identify genes with concordant trajectory dynamics across these large-scale datasets, allowing us to detect robust developmental signals.

To define a shared mineralization trajectory architecture, we then integrated our tooth atlas with a previously constructed bone atlas. Applying TrajDTW to this combined resource revealed common molecular pathways governing the formation of distinct hard tissues, including bone, enamel, and dentin. Furthermore, cross-species analysis between human and mouse data uncovered cross-species conserved mesenchymal/odontoblast programs.