Disease models drift. Single-cell analysis sees what bulk sequencing hides.

Disease models drift over time — through passaging, engraftment, and culture — and a population-average readout can miss it. This guide is for scientists relying on patient-derived xenografts, organoids, cell lines, or engineered lines to make drug-program decisions, and it lays out the difference between two kinds of questions: catalog questions (which variants are present, at what frequency), which bulk DNA sequencing answers efficiently, and structure questions (which cells, which rare subclones, what changes over time), which require single-cell resolution.

The guide covers model fidelity confirmation, lead-candidate confirmation (which clones a candidate clears versus which survive it), and how a single clonal profile can carry forward from bench to trial design. It includes a scored worksheet readers can run against their own models today, answers to practical questions on sample input, panel coverage, cost, and bioinformatics support, and a walkthrough of a published AML xenograft study showing two outcomes — resistance emergence and full eradication — that look identical on a bulk average but diverge completely at single-cell resolution.