CP-01..CP-15 Clinical Pharmacology Agents

The 15 agents

Lifecycle coverage.

CP-01

FIH Dose Selection

First-in-human dose calculation, MTPC estimation, safety window optimization. Allometric scaling + species-bridge models.

CP-02

Exposure-Response (E-R)

Concentration-vs-effect modeling. Both efficacy E-R and safety E-R; supports time-to-event and continuous endpoints.

CP-03

QTc Prediction

Concentration-QTc modeling per ICH E14. Identifies torsadogenic risk before dedicated TQT study.

CP-04

Transporter Substrate Scoring

P-gp, BCRP, OATP1B1/3, OAT/OCT substrate scoring. Predicts DDI from transporter biology.

CP-05

Pharmacogenomics (PGx)

CYP2D6/2C19/2C9/3A5 stratification. Identifies populations needing dose adjustment by metabolizer phenotype.

CP-06

Special Populations

Pediatric, hepatic impairment, renal impairment dosing recommendations. PBPK-anchored.

CP-07

Therapeutic Drug Monitoring (TDM)

Target concentration windows, sampling protocols, dose-adjustment algorithms.

CP-08

Food Effects

Predicts food-effect magnitude (high-fat vs fasted) from molecule properties + formulation.

CP-09

Immunogenicity Risk

ADA risk stratification for biologics. T-cell epitope prediction + structural similarity to self.

CP-10

Bioequivalence (BE) Design

BE study design for generics, line extensions, formulation bridges.

CP-11

In Silico ADME

Absorption, distribution, metabolism, excretion prediction from molecular structure. Pre-DMPK triage.

CP-12

Pharmacovigilance

Signal detection in post-market AE data. Disproportionality metrics + Bayesian shrinkage.

CP-13

Metabolite ID & Safety

Active/toxic metabolite identification, MIST analysis, contribution-to-effect estimation.

CP-14

Chronopharmacology

Circadian timing of dosing — when matters in addition to how much.

CP-15

Polypharmacy Risk Engine

Multi-drug interaction stack analysis for patients on 5+ medications. Identifies highest-risk pairs.

Architecture

Each agent: HTML prime directive + Python module.

Every agent ships in two halves:

Styled HTML prime directive — human-readable specification of what the agent does, its inputs, outputs, and known limitations. Useful for reviewers, regulators, and onboarding.
Python module — __init__.py + core class file with the actual implementation. Useful for orchestration, batch runs, and pipeline integration.

Source path on disk: agents/Clinical_Pharmacology/CP_NN_*/. 35 files total (15 MD + 20 HTML).

Integration with the combo framework

CP agents feed the predictive framework.

The CP agents are not standalone — they feed the predictive framework with the inputs it needs. For example:

CP-01 (FIH Dose) provides the dose-scaling assumption for the lenvatinib safety prediction.
CP-03 (QTc) feeds the QT-prolongation row in the safety AE table.
CP-04 (Transporters) flags DDI risk for combination therapies.
CP-06 (Special Populations) provides the population-fragility coefficient inputs.
CP-12 (Pharmacovigilance) is how new safety signals enter the model post-deployment.

Need clinical-pharm coverage for a program?

We can run any subset of CP-01..CP-15 against a candidate molecule, regimen, or population. Most engagements run 4–6 agents in a single briefing.