Can AI Rescue Failed Drugs?
Drug development has an expensive problem. Around 90% of drug candidates never make it to market, with many failing late in development after millions of pounds and years of research have already been sunk. Usually, these failures are not due to a lack of efficacy, but due to an unexpected safety issue like liver toxicity or other adverse reactions.
A Cambridge-based biotech company, Ignota Labs, believes AI may have the potential to give some of these abandoned therapies a second chance. Founded in 2021, Ignota has developed an AI-driven platform called SAFEPATH that aims to identify why promising drug candidates failed and determine whether those problems can be fixed. Rather than discovering entirely new medicines from scratch, the company focuses on reusing existing compounds that were shelved after safety concerns.
The concept is deceptively simple. If a drug demonstrated strong therapeutic potential but failed due to a specific ADR, could AI identify the root cause and suggest a small modification that preserves efficacy while removing the ADR?
To answer this question, the software combines multiple layers of biological and chemical data like molecular structure and protein interactions. By analysing these relationships simultaneously, AI can uncover patterns that would be far more difficult for researchers to identify manually.
According to the company, this technology can identify a range of issues like unintended off-target binding, poor drug distribution, or elimination problems that may have contributed to a candidate’s failure. The goal is targeted redesign rather than repurposing, so that the underlying cause of failure can be addressed while retaining the same therapeutic intent.
While much of the conversation around AI in the pharma industry centres around discovering new drugs, this innovation shows there may be significant value in revisiting compounds that have already undergone substantial development. In many cases, the biological promise is already proven but the real challenge is understanding why things went wrong.
If successful, AI-powered drug rescue could reduce R&D costs, minimise wasted investment and potentially bring valuable medicines back into the clinic much faster than starting from scratch. AI is not only helping discover new drugs, it may also be helping investors by recovering the ones we almost lost.
“Our core aim is to take these drugs, make minimal chemical changes and get them back into the clinic as efficiently as possible”
Layla Hosseini-Gerami, Co-Founder of Ignota Labs
