Question 1

Is Raganelė an LLM-based tool?

Accepted Answer

No. Raganelė is a mathematical model of coupled human physiology built from ordinary differential equations, paper-anchored parameters, and proprietary computational techniques rooted in quantitative finance. There is no large language model anywhere in the modelling pipeline. Every output is a numerical solution to a peer-reviewed equation system, not generated text.

Question 2

How is Raganelė different from drug-discovery AI startups?

Accepted Answer

Most drug-discovery AI uses statistical or deep-learning models trained on chemistry datasets. Raganelė instead solves the actual ODE systems that describe drug absorption, distribution, metabolism, excretion, and pharmacological effect across coupled organ systems. Parameters come from published clinical and pre-clinical studies, not from black-box training. Outputs are reproducible, auditable, and tied to specific peer-reviewed anchors.

Question 3

What can Raganelė produce a report on?

Accepted Answer

Raganelė delivers structured reports on: (1) personalised dosing schedules for a given patient covariate profile, including drug holidays and ramp profiles, (2) drug-drug interaction predictions for any pair or cocktail of approved compounds, (3) unexpected side-effect mechanisms via cross-organ coupling pathways, (4) repurposing candidates ranked by mechanism match against an under-served indication, (5) clinical trial replication and counterfactual extension to populations the original sponsor did not study, (6) adverse-event counterfactuals showing what the AE rate would have looked like under a different protocol, (7) formulation and route-of-administration comparisons, (8) blockbuster-candidate screening across the cross product of approved compounds against open indications.

Question 4

What mathematical framework does the body model use?

Accepted Answer

A composed system of ordinary differential equations describing the time evolution of state variables across cardiac, renal, hepatic, immune, marrow, endocrine, gastrointestinal, central nervous, and tumour subsystems. Roughly 1,949 state variables are integrated together with shared substrates, feedback loops, and pharmacological levers. The full system is solved end-to-end at sub-second wall-clock per scenario using proprietary techniques originally developed for derivatives pricing in quantitative finance.

Question 5

Can Raganelė help find blockbuster drugs?

Accepted Answer

Yes. The model evaluates approved-compound libraries against open commercial indications by computing predicted clinical response, predicted toxicity, and predicted exposure margins simultaneously. Output is a ranked candidate list with the mechanism story attached. The same machinery surfaces under-exploited dose schedules of existing molecules — the most common path to a new blockbuster is a new schedule, not a new molecule.

Question 6

Which drugs and diseases does the model cover?

Accepted Answer

Oncology (all 14 Hanahan-Weinberg cancer hallmarks wired, 10 tumour types including SCLC/NSCLC, breast, CRC, prostate, pancreatic, glioma, melanoma, RCC, HCC, haematologic), cardiovascular (hypertension, atrial fibrillation, heart failure, atherosclerosis, drug-induced TdP, stroke, anticoagulation, acute MI), metabolic (T2DM, T1DM, CKD, osteoporosis, thyroid, NAFLD), immune and inflammation (sepsis, ARDS, Long COVID, vaccine response, RA, IBD, asthma, COPD), CNS and behavioural (Alzheimer's, Parkinson's, depression, opioid use disorder, alcohol use disorder, epilepsy), with the underlying organ substrate covering cardiac electrophysiology, hepatic CYP450 enzymes, renal handling, plasma kinetics, immune signalling, marrow production, bone remodelling, gut absorption, and CNS pharmacology.

Question 7

How does the model handle drug-drug interactions?

Accepted Answer

Drug-drug interactions emerge as a consequence of shared metabolic and pharmacokinetic substrates rather than being hand-coded as rules. A drug that inhibits CYP3A4 will mathematically raise the AUC of any comedication metabolised by that enzyme via the competitive inhibition kinetics, and the downstream effect (such as QTc prolongation on the cardiac action potential) propagates through the coupled body system. The same logic surfaces overlapping marrow suppression, additive hepatotoxicity, and shared renal clearance bottlenecks.

Question 8

Who is Raganelė for?

Accepted Answer

Pharmaceutical R&D groups designing dose schedules and combination protocols, biotech companies evaluating repurposing candidates against under-served indications, academic clinical pharmacology and systems-pharmacology research groups looking to extend their published models into a composed body, and clinical investigators replicating and extending published trials. Engagement starts with a worked example on the partner's own published case before scaling.

Delivery & PK-PD

The peak determines the toxicity

Where this lands

Where this works less well

Send us an asset.
We’ll send back a schedule.

Delivery & PK-PD

The peak determines the toxicity

Where this lands

Where this works less well

Send us an asset. We’ll send back a schedule.

Send us an asset.
We’ll send back a schedule.