uid stringclasses 10
values | dataset_id stringclasses 1
value | year int64 2.02k 2.02k | therapeutic_area stringclasses 6
values | program_stage stringclasses 1
value | lead_candidate stringclasses 10
values | efficacy_model_relevance stringclasses 3
values | pk_exposure_adequate stringclasses 2
values | tox_margin_clear stringclasses 2
values | safety_pharm_findings_consistent stringclasses 2
values | species_selection_justified stringclasses 2
values | package_concordance stringclasses 3
values | sponsor_package_summary stringclasses 10
values | question stringclasses 1
value | ground_truth_label stringclasses 2
values | ground_truth_rationale stringclasses 10
values | outcome_signal stringclasses 4
values | source_citation stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
PP-TR-001 | pharma-preclinical-package-coherence-risk-v0.1 | 2,024 | Oncology | IND_enabling | CAND_A | high | yes | yes | yes | yes | high | Efficacy aligns with exposure; tox margin clear; safety pharmacology clean. | Is preclinical package coherence intact. Answer coherent or incoherent. | coherent | All package elements align with IND story. | fiH_ready | Synthetic |
PP-TR-002 | pharma-preclinical-package-coherence-risk-v0.1 | 2,024 | Oncology | IND_enabling | CAND_B | medium | no | no | no | no | low | Strong efficacy claims but exposure not demonstrated; tox margin unclear; species rationale absent. | Is preclinical package coherence intact. Answer coherent or incoherent. | incoherent | Claims exceed PK and tox support. | ind_hold_risk | Synthetic |
PP-TR-003 | pharma-preclinical-package-coherence-risk-v0.1 | 2,023 | Cardiology | IND_enabling | CAND_C | high | yes | yes | yes | yes | high | Dose response in disease model with matched PK; tox supports margin. | Is preclinical package coherence intact. Answer coherent or incoherent. | coherent | Concordant PK tox efficacy. | fiH_ready | Synthetic |
PP-TR-004 | pharma-preclinical-package-coherence-risk-v0.1 | 2,023 | Cardiology | IND_enabling | CAND_D | low | yes | no | no | no | low | Safety package patched from healthy animals; disease model weak; tox signal near therapeutic exposure. | Is preclinical package coherence intact. Answer coherent or incoherent. | incoherent | Model mismatch and no safety margin. | ind_hold_risk | Synthetic |
PP-TR-005 | pharma-preclinical-package-coherence-risk-v0.1 | 2,022 | Neurology | IND_enabling | CAND_E | high | yes | yes | yes | yes | high | Transgenic model matches disease; CNS exposure confirmed; tox acceptable. | Is preclinical package coherence intact. Answer coherent or incoherent. | coherent | Translation package strong and aligned. | fiH_ready | Synthetic |
PP-TR-006 | pharma-preclinical-package-coherence-risk-v0.1 | 2,022 | Neurology | IND_enabling | CAND_F | medium | no | no | yes | no | low | Behavioral efficacy reported but brain exposure not shown; tox OK but relevance unclear. | Is preclinical package coherence intact. Answer coherent or incoherent. | incoherent | Missing exposure bridge breaks package. | translation_failure_risk | Synthetic |
PP-TR-007 | pharma-preclinical-package-coherence-risk-v0.1 | 2,021 | Rheumatology | IND_enabling | CAND_G | high | yes | yes | yes | yes | high | Immune model matches mechanism; tox margin clear; QT clean. | Is preclinical package coherence intact. Answer coherent or incoherent. | coherent | Package internally consistent. | fiH_ready | Synthetic |
PP-TR-008 | pharma-preclinical-package-coherence-risk-v0.1 | 2,021 | Rheumatology | IND_enabling | CAND_H | high | yes | no | no | yes | medium | Efficacy strong but tox margin narrow and safety pharm flags ignored. | Is preclinical package coherence intact. Answer coherent or incoherent. | incoherent | No margin and ignored safety pharmacology. | ind_hold_risk | Synthetic |
PP-TR-009 | pharma-preclinical-package-coherence-risk-v0.1 | 2,020 | Infectious Disease | IND_enabling | CAND_I | high | yes | yes | yes | yes | high | Pathogen clearance aligns with exposure; tox margin supports dosing. | Is preclinical package coherence intact. Answer coherent or incoherent. | coherent | Clear exposure-response and margin. | fiH_ready | Synthetic |
PP-TR-010 | pharma-preclinical-package-coherence-risk-v0.1 | 2,020 | Metabolic Disease | IND_enabling | CAND_J | medium | yes | no | no | no | low | Weight loss in animals but tox near therapeutic; species not relevant; package discordant. | Is preclinical package coherence intact. Answer coherent or incoherent. | incoherent | Discordant package and weak justification. | portfolio_risk | Synthetic |
What this repo is about
This repo evaluates whether a drug’s preclinical package is internally consistent enough to justify first-in-human entry.
It checks whether efficacy, pharmacokinetics, toxicology, and safety pharmacology actually support the same story.
Many programs fail because preclinical data exists but does not align. Studies are run. Signals appear. But the elements do not form a coherent bridge to humans.
This repo detects that gap.
What it looks for
efficacy only at exposures not achievable in humans
toxicology margin too narrow for planned dosing
species used that do not match human biology
safety pharmacology signals ignored
PK and efficacy disconnected
IND narrative stronger than underlying data
What the model must decide
Given one program row, return
coherent or incoherent
How to interpret
coherent means the preclinical package forms a viable bridge to humans
incoherent means the package contains structural gaps that put FIH or IND at risk
Why it matters
Preclinical incoherence is one of the biggest causes of
IND holds
failed FIH trials
wasted development spend
late safety surprises
This repo provides a structured early warning system.
Who uses it
translational medicine teams
toxicology leads
clinical pharmacology
governance boards
investors doing diligence
- Downloads last month
- 12