PS: Main figs are only a (thoughtful) selection. There’s so much more (pretty) data, partly in the supp. — worth checking out!

PPS: all data/results are up for grabs, including R-, Py-friendly & language-agnostic versions. — developers, go!

- we observe a bifurcating tumor evolution: concordant with malignancy assessments; one end-point organizes into concentric layers & maps to invasion/metastatic sites

- we introduce “transition crypts” — split histological & transcriptional makeup (from crypt fusion or abrupt transformation?)

- sections that span reference, pre- & malignant regions let us look into what happens along the molecular/spatial neoplastic transformation axis (compositional drifts, cellular density & heterogeneity, pathways & signaling, etc.), from cells/crypts via niches to tissue-wide scale

Thx @hoheyn.bsky.social & Sabine Tejpar for getting/keeping the game afoot;
KU Leuven & @brukerspatial.bsky.social teams for your support.

These data were no breeze yet real gratifying; in brief:

- we survey 3.5M cells (1.6k FOVs = 4.3cm2) via WTx CosMx SMI & snPATHO-seq + digital histopatholgy

Special thx @pascual-reguant.bsky.social & Irene Ruano for your rare combi of standards, humanity, passion & perseverance. It wasn't always fun'n'giggles, but I laughed & learned so much from/with you (e.g., how to say & spell “tubulovillous”). Muchísimas gracias para todo mis ángeles.