Korean Team Uses Math and Single-Cell Sequencing to Tackle Colorectal Tumors
Tthe Tech Times
A Korean team just showed that, for colorectal tumours, the answer may be yes—and they did it with maths, single‑cell sequencing and a well‑timed molecular “tap” on a protein called USP7.
1. The big idea: crash a tumour’s operating system
Think of every cell state as a ball rolling in a landscape of valleys (biologists call them attractors). A healthy colon cell rests in one valley; a cancer cell, in another. The researchers built REVERT, a Boolean‑network framework that maps that entire landscape from ordinary single‑cell RNA‑seq data, then looks for the steepest lever able to tip the ball back into the healthy valley. Instead of searching gene‑by‑gene, they reconstruct the whole regulatory‑feedback circuit that emerges at the tumour’s critical transition—the shaky middle ground where cells can still swing either way.
Under the bonnet
CopyKAT CNV ordering → locates the sub‑clone straddling normal and malignant genomes.
Pseudotime + DE genes → sets a temporal axis through that sub‑clone.
Boolean logic in strongly‑connected components → captures feedback loops with <30 TF nodes, avoiding combinatorial blow‑up.
Cancer Score (CS) → a quantitative “malignancy volume” that blends attractor depth, width and distance from the tumour apex; CS > 0.5 flags an aggressive landscape.
Knock out or over‑express each node in silico, watch CS plummet, and you have your shortlist of “reversion switches”.
2. Meet the switch: USP7
Among thousands of simulations, silencing the de‑ubiquitinase USP7 collapsed the cancer valley: CS fell from 0.62 to 0.056—an 11‑fold drop that all but erased the malignant attractor.
Why USP7?
It stabilises multiple oncogenic drivers (MYC, β‑catenin, p53 regulators).
It sits at the crossroads of Wnt signalling and DNA‑damage tolerance.
Crucially, its dependency score is high in tumour cells but negligible in normal epithelium, giving it the therapeutic sweet‑spot. (sbie.kaist.ac.kr)
3. From silicon to living tissue
The team dosed patient‑derived colon‑cancer organoids with the USP7 inhibitor P22077. Ten days later:
Organoid expansion shrank by up to ~70 % at 15 µM.
Stem‑cell and MYC/MAPK signatures nosedived, while markers of orderly epithelial architecture resurfaced.
Morphologically the spheres flattened into crypt‑like structures reminiscent of normal colon. (sbie.kaist.ac.kr)
(The chart above visualises the dose‑response curve extracted from their Figure 5g.)
4. Deeper tech nuggets for the curious
Cancer‑score math: CS = Σ_k [(1−S_k/S_max) × B_k/(B_N + B_x + B_C) × d_k] where S is attractor entropy, B basin size and d effective distance along the normal‑cancer axis.
Landscape smoothing: discrete attractors are interpolated with a Gaussian kernel f(x) ∼ -S e^(−x²/B²) to render the 3‑D plot you see in their Fig. 4d.
Double‑hit logic: knocking down YY1 together with CDX2 was second‑best, because it widens the normal basin rather than merely shallowing the cancer one—subtle but important when you worry about relapse dynamics.
Stat power: the USP7 organoid assay used repeated‑measures ANOVA (p < 0.001), three biological replicates, and L1000 connectivity mapping to confirm pathway shifts—robust by ex‑vivo standards.
5. Why this matters
Mutation‑agnostic: REVERT cares about state, not genotype; theoretically it can sidestep the heterogeneity that torpedoes many targeted drugs.
Lower collateral damage: instead of carpet‑bombing dividing cells, you nudge them back into line—think immunotherapy‑level precision without cytokine storms.
Portable: any tumour from which you can grab matched single‑cell transcriptomes could be run through the same pipeline in weeks.
6. What’s next?
Pan‑cancer benchmarking on the Human Tumour Atlas Network.
USP7 PROTACs entering mouse xenografts to test systemic tolerance.
Patient‑specific avatars: feed an individual’s organoids into REVERT, rank their personal switches, and choose the gentlest cocktail that drives CS below 0.1.
If successful, oncology may pivot from “seek and destroy” to “seek and restore.”
Sources
D. Shin et al. “Attractor Landscape Analysis Reveals a Reversion Switch in the Transition of Colorectal Tumorigenesis,” Adv Sci 12 (8): e2412503 (2025) DOI 10.1002/advs.202412503. (sbie.kaist.ac.kr)
KAIST News Centre press release, 5 Feb 2025.
Conexiant coverage, 7 Feb 2025.
#Hashtags
#CancerReversion #SystemsBiology #USP7 #SingleCell #Organoids #PrecisionOncology #Bioinformatics #TTTimes
Subscribe to my newsletter
Read articles from Tthe Tech Times directly inside your inbox. Subscribe to the newsletter, and don't miss out.
Written by