Closing the Data Gap in AI Biology Through Cell-Cell Interaction Mapping

Existing biological datasets excel at identifying cell types and locations, yet they frequently fail to capture the dynamic communication—secretion, signaling, and competition—that defines living systems. By treating cells as isolated units, developers of biological foundation models miss the functional ground truth of how cells behave when in contact. The Billion Cell×Cell Project aims to bridge this divide by systematically characterizing cell-cell dyads across diverse biological conditions.

Partillion Bioscience is central to this shift, as their Nanovial technology enables the creation of microscale compartments where researchers can isolate and measure specific cellular pairs. Recent findings, including a preprint on NK cell activation, demonstrate that these interaction-resolved datasets provide insights that standard single-cell sequencing cannot replicate. According to Joe de Rutte, CEO of Partillion, this data is essential for evolving AI from simple snapshot analysis toward predictive models that understand functional outcomes in complex biological environments.