Si-gul combines target intelligence, computational chemistry and experimental validation to create de-risked, licensable drug assets for pharma partners.
Protein structure model, ligand bound. Medicine begins as something small enough to hold.
Targets are chosen on partial evidence. Molecules are designed against them. The answer arrives in the clinic, years and billions later.
Each gate removes candidates. Target selection, molecule design, validation, then the clinic, which removes almost all of them.
Most programs fail late, after most of the money has been spent. The failure was knowable earlier. Nobody asked.
Targets are scored on genetic, clinical and structural evidence before any chemistry begins. Weak biology is set aside while it is still cheap to set aside.
Generate candidates only against targets that hold. Validate them in the lab. Fewer dead ends reach the clinic, and the ones that reach it are better.
Illustrative. The field shows the shape of the argument, not a measured attrition rate.
Four stages, each producing evidence the next one is measured against. The output is not a report. It is a program a partner can take on.
Genetic association, clinical evidence, druggability, safety signal and literature strength scored together for every candidate target in an indication.
Output · a ranked target with an evidence file, not a hunch.
Candidate molecules designed and optimised against the ranked target, filtered for potency, selectivity and developability before anything is synthesised.
Output · a small, defensible series worth making.
Binding, cellular and early ADME assays run with contract research partners. Every prediction is tested, and every result is fed back into the models.
Output · experimental evidence a partner's own scientists can audit.
A validated program packaged for partnering: target rationale, chemistry, data package and a development plan, ready for in-licensing or co-development.
Output · an asset pharma can build on.
Each stage is designed to end a program early if it should end. That is where the value is made.
Most discovery platforms predict once. Si-gul's models are retrained on the results of their own predictions, so the error in the next selection is smaller than the error in the last one.
The economics of drug discovery are set by how late failure is discovered. Moving that discovery earlier changes every number below.
Phase I to approval
Roughly one in eight programs that reach first-in-human studies becomes a medicine.
From target to patient
Long enough that the biology chosen at the start is rarely re-examined before the end.
Estimated cost per approved drug
Including the cost of the programs that failed on the way. Most of that spend is late-stage.
Figures are industry-wide estimates supplied by Si-gul for this page and should carry their published sources at launch.
Si-gul identifies promising biology, designs and optimises candidates, generates experimental evidence, and packages validated programs for partnering and licensing. The platform is how we work. The asset is what we hand over.
Cross-border biotech licensing value, 2025
Pharma increasingly buys its pipeline rather than growing all of it. Programs originated outside the traditional hubs, with the evidence to survive diligence, are what that market is looking for.
Origination in India. Partnering wherever the medicine is built.
A great deal of disease biology is under-studied because it was never commercially obvious. Much of it matters most in the places we come from. We think evidence, not geography, should decide what gets built.
If you lead search and evaluation, external innovation or early R&D at a pharma or biotech, we would like to show you what we are building and hear what you need.
