About BNTRG
Biology can now be measured at a depth and frequency that was impossible when most laboratory workflows were designed. Genetics exposes changes inside a cell. Robotics repeats precise work. Sensors preserve each observation. Artificial intelligence and cloud computing compare more conditions than one team could hold in its head.
BNTRG was founded by a computer scientist and genetic engineer who studied at Oxford and Cambridge. The group applies software engineering methods to biological research: automated execution, repeatable workflows and complete experimental records. Cancer research is the first focus because the time required to test an idea has a direct human cost.
A manual experiment produces one result at the speed of the people running it. A programmed experiment can preserve every step, repeat across parameter sets and gather comparable data continuously. The components now exist to give a small scientific team an experimental reach that previously demanded a far larger institution.
Every delay between a promising idea and reliable evidence uses time that people waiting for better diagnosis and treatment do not have. Controlled automation can run more experiments, capture richer measurements and preserve reproducible records. Strong ideas receive more chances to become useful evidence.
Science already in progress
Scientists are testing these fields now in people, animals and cultured cells. Repeatable experiments and comparable data increase the number of conditions they can test and shorten the path from one result to the next. Each panel below sets out where the work has reached and the question a larger experimental system could take on next.
Targets and treatment response
Treatment matched to a tumour's own biology needs a test that reads the tumour quickly and a decision made before the disease moves on. In 2024 the Nobel Prize in Chemistry went to the team behind AlphaFold, for solving a problem that had held biology for fifty years: their models predict the shape of more than two hundred million proteins, and every one of them is free to use. A Sydney programmer sequenced the tumour of his own dog, Rosie, designed a personalised cancer vaccine for her with AlphaFold, and had it made in a university lab. The tumour shrank by three quarters, for about two thousand pounds. In people the same idea has now worked: a personalised melanoma vaccine, built from the fingerprint of each patient's own tumour, passed its Phase 3 trial in 1,137 people, cutting the risk of the cancer returning by 42 per cent.
A precise result only counts when it changes what happens to the person. The questions a larger experimental system could answer: which differences in a tumour expose it to treatment, how quickly that answer can be produced, and whether it can be produced the same way for every patient rather than only for the cases a laboratory already knows about.
Evidence: Nobel Prize in Chemistry 2024 · Rosie, and the vaccine designed for her (UNSW, 2024) · INTerpath-001 Phase 3, ESMO (2026)
Cells, animals and early human research
Partial reprogramming uses three of the Yamanaka factors to reset the chemical marks on a cell's DNA without pushing the cell all the way back to an embryonic state. It has improved tissue function in mice, and the first human step is under way: ER-100, an epigenetic therapy built on OSK reprogramming, opened a Phase 1 study for age-related optic nerve conditions in March 2026 and is recruiting.
The risk is why this stayed in the laboratory for two decades. Reprogrammed cells can lose their identity or grow into tumours, and a treatment has to be shown safe before it is shown to do anything. Working out which changes can be reversed, and which cannot, is a measurement problem before it is a therapy problem.
Evidence: The long and winding road of reprogramming-induced rejuvenation, Nature Communications (2024) · ER-100 Phase 1 registry
First human safety study complete
Teeth do not regrow once they are lost, which makes the mouth a clean place to test whether a programme can travel from cell biology to a human safety study. TRG035, a treatment that stimulates tooth formation, has been through a Phase 1 double-blind, placebo-controlled, dose-escalation study in healthy adults missing back teeth. Registered in October 2024 at Kyoto University Hospital, the study aimed for 30 participants and has completed recruitment.
What has not been shown is the part a patient cares about: whether an early safety study can be followed by a tooth that regrows and works. That is a longer trial, and the kind of work that falls behind when every experiment is run once, by hand, for one person at a time.
Evidence: TRG035 Phase I registry
Mouse lifespan and human ageing studies
A multi-site study found that rapamycin extended median and maximal lifespan in genetically diverse mice, even when treatment began late in life. Human studies have since tested rapamycin against safety and healthspan measures in older adults. Whether it can extend human life remains unanswered.
That creates one valuable research question: why did it work in mice? Mapping the molecular and epigenetic effects could identify which changes drove the lifespan result. Rapamycin is already used as an immunosuppressant in people and has recognised risks, including infection, impaired wound healing and metabolic effects. The research therefore targets the useful biological effects downstream of rapamycin and tests how to reproduce them safely in human cells. The same experimental system can investigate cancer targets, treatment response, cell regeneration and other questions across genetics and biology.
Evidence: Harrison et al., Nature (2009) · PEARL trial results, Aging (2025) · EMA medicine overview
Retinal cells, optic nerves and stem cells
Sight is lost when cells that do not replace themselves go missing. A research consortium has proposed tackling that by repopulating the retina, transplanting stem cells and attempting to regenerate the optic nerve, and a Phase 1 epigenetic therapy, ER-100, is recruiting for open-angle glaucoma and non-arteritic anterior ischaemic optic neuropathy.
Whether any of that restores sight in people is not yet known. The field needs a way to test many candidate approaches, measure recovery properly and compare them against each other rather than against a single promising result.
Evidence: RReSTORe consortium, Ophthalmology Science (2023) · ER-100 Phase 1 registry
How the work scales
A precise experimental method can be versioned, reviewed, run repeatedly and improved. Every successful run adds measurements that the next researcher can use.
Record the inputs, steps, controls, measurements and outputs needed to perform the experiment and understand its results.
Connect the method to compatible instruments, laboratory providers and data services, then vary defined parameters without rebuilding the whole process.
Compare every run, preserve the full record and use the growing body of data to decide what deserves to be tested next.
Build with BNTRG
We are bringing together geneticists, cancer researchers, engineers, students and founders to turn important scientific questions into programmes that can be run, measured and expanded.