Pharmacometrics · worked example

Population pharmacokinetics of drug mABC: base and covariate models, reproduced and reviewed

A. Analyst · QC (independent) · P. Lead (review) · CRO Pharmacometrics · 2026
Every figure carries a plankton lens — the logo in its corner. Click it to open the kton world: where the figure came from, re-verified in your browser. Because the whole analysis is one federated graph, different figures open different walks of it:

A one-compartment base model and a weight-on-clearance covariate model were fit to sparse concentration–time data for drug mABC. Diagnostics and predictive checks are shown below. None of it asks for trust: each figure's bytes re-hash to a recorded output, each producing record is signed, the base fit was independently reproduced, and the final model was reviewed — all visible by clicking the lens.

1. Base model — fit and goodness of fit

Concentration–time, base model
Figure 1. Concentration–time, base model.
Goodness of fit, base model
Figure 2. Observed vs predicted (base). Its lens descends to the base fit — where QC's reproduction sits.

2. Covariate model — the final model

Parameter forest, covariate model
Figure 3. Parameter estimates, 95% CI (final). Its lens carries the reviewer's sign-off and model-role: final.
Visual predictive check, covariate model
Figure 4. Visual predictive check (final model).

3. Reproducibility

The base fit was re-run independently by QC; the recorded output matched byte-for-byte (an L0 reproduction). The final model was accepted by the reviewer. Both attestations are signed claims attached to the same graph nodes the figures descend from — open any lens and walk to them.

Rendered with the kton lens · a figure with no logo is simply unknown to the connected planktons. The registry for this paper is delivered alongside it (lens-paper/data/); your own connected planktons, if any, are consulted too — so you may see more, or less, than the author shipped.