Scaling Phenotypic Toxicology: How We Use iPSC Hepatocytes and Foundation Models to Decode DILI
Drug-induced liver injury (DILI) remains a major cause of clinical failure because traditional assays often miss the early, subtle, and heterogeneous cellular responses that precede overt toxicity. At pixlbio, we’re building a new data foundation for liver safety by combining our human iPSC-derived hepatocytes (pixHep), large-scale Cell Painting, and advanced computer vision to quantify these early-stage, sublethal phenotypes with single-cell precision. In this post, we walk through the data we presented at Cytodata Berlin—how we generate and process high-resolution phenomics datasets and what the emerging DILI landscape looks like when viewed through morphology. Ultimately, this work represents a growing foundation for predictive biology spanning mechanism, safety, and toxicology.
