Pharmacokinetic & Pharmacodynamic (PK/PD) Modelling

What it does

• Non-compartmental analysis: AUC, Cmax, lambda_z, half-life, CL/F, Vz/F, with lambda_z windows chosen by adjusted R², Tmax-and-earlier points excluded, and automatic findings when >20% of AUCinf is extrapolated or the terminal phase spans too few half-lives.
• Compartmental fitting and model selection: 1/2/3-compartment comparison via AIC, BIC and F test; log-scale estimation with asymmetric CIs; RSE and correlation reporting to expose non-identifiability; runs test to separate structural misspecification from a wrong error model.
• Population PK dataset checks: catches the silent NM-TRAN defects — BLQ text read as a real zero, blank covariates becoming 0 kg, ADDL without II, duplicate timestamps resolved by file order.
• Regimen simulation: steady-state peak/trough plus population target attainment with between-subject variability.
• Exposure-response and C-QTc: Emax/EC50 with plateau-outside-data flags, and the ICH E14 question (upper bound of the two-sided 90% CI vs 10 ms).
• Bioequivalence: ABE, EMA ABEL and FDA RSABE kept distinct, plus sample size and power.
• Allometry, paediatrics and first-in-human: maturation-adjusted scaling, NOAEL-derived MRSD and MABEL.
• Static DDI: ICH M12 basic and mechanistic static models with their cut-offs.
• Bayesian TDM: MAP individual parameter estimation with a warning when a single level cannot separate CL from V.

Who it's for

• Pharmacometricians and clinical pharmacologists in pharma or CROs
• NONMEM / Monolix / nlmixr2 users who want dataset hygiene and model plausibility checked first
• Teams preparing regulatory deliverables (NCA reports, population analysis plans)
• Graduate-level PK/PD teaching and self-study

Example uses

1. Phase 1 NCA report: python3 nca.py -i profile.csv --dose 100 --route extravascular --partial-auc 0-24 returns AUC(0-24), Cmax and t½ while flagging that 25% of AUCinf was extrapolated, prompting a switch to AUC(0-tau) at steady state.
2. Avoiding a model-selection trap: fit_compartmental.py --compare 1cmt,2cmt,3cmt shows AIC picking three compartments while BIC and the F test reject it and Q3 carries 98% RSE — the two-compartment model wins.
3. Stress-testing a regimen: simulate_regimen.py --simulate 2000 --omega-cl 0.35 --target-trough 4.0 reveals a typical trough of 3.6 but only 44% population attainment, forcing a dose or interval change.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/K-Dense-AI/scientific-agent-skills/HEAD/skills/pkpd-modeling/SKILL.md

What I want: (describe your task here)

Install the skill I found on AI Skill Mart.
Copy the skills/pkpd-modeling folder from the GitHub repo K-Dense-AI/scientific-agent-skills into my ~/.claude/skills/pkpd-modeling/.
When it's done, tell me in one line what this skill can do.