From tissue to reasoning
This section holds an animation in four chapters. It is stylised throughout: no real patient data, no product interface, no claim about performance.
Observe. A tissue microarray — one small core of tissue per patient — is stained in rounds. The nuclei come up first, then a band of light crosses the slide for each round, and every core it passes blooms into that round’s cells: the stroma, the tumour, the immune cells.
Question. Each patient’s outcome is drawn around their core, a solid ring for a responder and a dashed one for a non-responder, and a question forms over them: why do these patients respond?
Reason. The cores lift off the slide and arrange themselves by their biology. A hypothesis is set over one subgroup — that responders hold more T cells — and the cores sort into what supports it, what is ambiguous, and what contradicts it. One contradiction rises out of its group.
Explain. That core is shown close up beside one that supports the hypothesis, the T cells lit in both. They hold the same number of T cells; what differs is where those cells sit — inside the tumour in the responder, held at its edge in the other. The reasoning is then drawn downwards as a graph: the question to the hypothesis, the hypothesis to the evidence on either side, each piece of evidence to the tissue it came from, and both tissues to a revised hypothesis. Every piece keeps its origin and its type, and the contradiction is not discarded — it is what revises the hypothesis.
The section closes on the words: AI reasoning for experimental biology.
© 2026 Quaint Technologies, Inc.