Presets
Playback
Parameters
Camera
Layers
Quality
Model check
The science
What S phase is. A chromosome is copied from many origins at once. Each origin is first licensed (a ring-shaped MCM helicase is loaded in G1), then fires at a random time in S phase; two forks leave it in opposite directions and run until they meet the forks of the neighbouring origins. Because forks are slow (about 2 kb per minute in yeast, 1 to 2 kb per minute in human cells) a chromosome of millions of base pairs could never be copied from one origin in time: S phase is set by how many origins fire and how fast forks move.
The model. One chromosome of length L kb with licensed origins scattered at random at density rho per 100 kb. An origin that is still in unreplicated DNA fires with rate k per minute; an origin overrun by a fork before it fires is lost (passive replication). Fork speed is v = v0 · Q10(T-30)/10 · [c/(Km+c)]/[c0/(Km+c0)] · H(T) with Q10 = 2 (a typical enzyme temperature coefficient), Michaelis-Menten dependence on the dNTP pool c (Km = 12 µM, normalised so that 60 µM gives 1.8 kb/min) and a heat-shock factor H that collapses above about 38 °C; these constants are illustrative, chosen to give a 30-40 minute S phase for yeast-like parameters, not measurements of one organism. A lesion stalls the first fork that meets it for the repair time; with the checkpoint on, the firing rate is multiplied by exp(-stalled forks / 3) (the real checkpoint kinase Rad53/Chk1 restrains late origins when forks stall).
Exact result. With no damage and no checkpoint, the chance that a point is still unreplicated at time t is the chance that no origin has reached it. Origins within distance d reach it if they fired before t - d/v, so the expected number reaching it is 2 ρ v [t - (1 - e-kt)/k] and the unreplicated fraction is its Poisson exponential: f(t) = 1 - exp(-2 ρ v [t - (1 - e-kt)/k]) (a Kolmogorov-Johnson-Mehl-Avrami result). The simulated cells average onto it; this is why S-phase length falls as 1/√(ρ v k) when firing is slow and as 1/(ρ v) when it is fast.
Hydroxyurea and UV. Hydroxyurea depletes dNTPs, so forks crawl and the checkpoint holds back late origins; UV lesions stall forks and the checkpoint again slows firing. Both lengthen S phase, but the checkpoint also keeps stalled forks from collapsing in real cells, which this model does not represent.
Honest simplifications. One-dimensional chromosome (the 3D coil is only a picture of it, drawn as a confined random walk, not the real chromatin fold); random rather than clustered or timed origin firing (real cells fire early and late origin domains in a programme, and replication timing correlates with chromatin state); no dormant origins that fire near a stall; no fork collapse, termination delays or sister-chromatid cohesion; constant temperature and dNTP pool during the run (the real pool is made by ribonucleotide reductase and rises in S phase).
Checks. (1) stochastic cells against the exact KJMA-type formula; (2) the speed of a single fork crossing lesions against v / (1 + lambda v tau); (3) the rate-law arithmetic (Q10, half-saturation at c = Km, heat stress); (4) every bin of a damaged chromosome replicated exactly once, origin bins plus fork-copied bins equal to the genome.