Approximately 90% of drugs fail during clinical trials. The primary causes of this failure include lack of efficacy, unexpected drug toxicity and poor drug-like properties. These failure rates can make drug discovery and development a costly and carbon-intensive process. But what if it were possible to breathe new life into some of these failed drugs?
At Ignota Labs, Dr. Jordan Lane and Dr. Layla Hosseini-Gerami are on a mission to revive shelved drugs. By harnessing the power of artificial intelligence (AI) and focusing on drugs that have already shown promise, they hope to accelerate development timelines, reduce clinical risk and cut carbon emissions.
To achieve their goal, the team at Ignota Labs has developed SAFEPATH, an AI platform that combines machine learning with bioinformatics and cheminformatics datasets to better understand toxicity mechanisms. The team recently demonstrated the power of this approach to gain new insights into the toxicity mechanisms of erlotinib and gefitinib, two cancer drugs used to treat non-small cell lung cancer.
Technology Networks caught up with Lane and Hosseini-Gerami at the 2025 ELRIG Drug Discovery conference to learn more about how SAFEPATH can help solve existing drug safety and efficacy issues and the impact this could have on the sustainability of drug discovery and development.
Blake Forman (BF):
Senior Science Writer
Technology Networks
Blake pens and edits breaking news, articles and features on a broad range of scientific topics with a focus on drug discovery and biopharma. He earned an honors degree in chemistry from the University of Surrey. Blake also holds an MSc in chemistry from the University of Southampton. His research project focused on the synthesis of novel fluorescent dyes often used as chemical/bio-sensors and as photosensitizers in photodynamic therapy.
Many companies have tried to tackle attrition, for example, by predicting toxicity earlier, but Ignota Labs is taking a different approach by reviving abandoned assets. What do you see as the main advantages of this “rescue” model compared to de novo discovery?
CSO and Co-Founder
Ignota Labs
Dr. Jordan Lane has extensive experience in the AI drug discovery space and is driven by a profound commitment to revolutionizing drug discovery, integrating artificial intelligence with bioscience to enhance drug safety and efficacy. At Ignota Labs, Lane leads innovation, leveraging his extensive background in biochemistry, genetics and AI to pioneer transformative treatments in global patient care.
Leaving good science on the shelf is heartbreaking. I’ve seen projects and companies fail during the clinical trial stages. Discovering that a promising drug is producing off-target effects during clinical trials means that it probably could have been saved at some point. There are many examples of this in the industry. Our ability to use AI to “rescue” these molecules is really powerful.
Chief Data Science Officer and Co-Founder
Ignota Labs
Dr. Layla Hosseini-Gerami bridges the worlds of chemistry, biology and AI in her role as chief data science officer at Ignota Labs. At Ignota Labs, she has built upon these sophisticated methods to understand the mechanisms of drug toxicity and how to mitigate them. Her goal is to leverage AI not just to understand, but also to prevent adverse drug reactions, ensuring safer therapeutic options for patients.
We see this as a bit of a contrast to other uses of AI in the industry, where many applications are focused on trying to get more drugs into the funnel, whereas we're really focusing on rescuing drugs when they drop out of the funnel.
BF:
Senior Science Writer
Technology Networks
Blake pens and edits breaking news, articles and features on a broad range of scientific topics with a focus on drug discovery and biopharma. He earned an honors degree in chemistry from the University of Surrey. Blake also holds an MSc in chemistry from the University of Southampton. His research project focused on the synthesis of novel fluorescent dyes often used as chemical/bio-sensors and as photosensitizers in photodynamic therapy.
How does SAFEPATH generate evidence that researchers and regulators can trust when making high-stakes decisions?
LHG:
Chief Data Science Officer and Co-Founder
Ignota Labs
Dr. Layla Hosseini-Gerami bridges the worlds of chemistry, biology and AI in her role as chief data science officer at Ignota Labs. At Ignota Labs, she has built upon these sophisticated methods to understand the mechanisms of drug toxicity and how to mitigate them. Her goal is to leverage AI not just to understand, but also to prevent adverse drug reactions, ensuring safer therapeutic options for patients.
We've built SAFEPATH to be completely explainable, so every prediction we make isn't a black box. We can trace exactly why that prediction is being made. Even drilling down into what part of the molecule is contributing to an off-target prediction, which means that a chemist can have a look, and they can very quickly understand: Is the AI just talking nonsense because that doesn't make sense chemically? That really helps build trust and ensure that all our predictions are as accurate as possible.
Additionally, with the very granular, mechanistic hypothesis that SAFEPATH generates, we don't just say we think it's hitting this pathway; we can clearly demonstrate through observations, such as level of gene expression, inference and protein-protein interactions, what's happening.
Senior Science Writer
Technology Networks
Blake pens and edits breaking news, articles and features on a broad range of scientific topics with a focus on drug discovery and biopharma. He earned an honors degree in chemistry from the University of Surrey. Blake also holds an MSc in chemistry from the University of Southampton. His research project focused on the synthesis of novel fluorescent dyes often used as chemical/bio-sensors and as photosensitizers in photodynamic therapy.
Where do you think AI-driven repurposing or redesign can make the biggest sustainability gains across the discovery and development process?
JL:
CSO and Co-Founder
Ignota Labs
Dr. Jordan Lane has extensive experience in the AI drug discovery space and is driven by a profound commitment to revolutionizing drug discovery, integrating artificial intelligence with bioscience to enhance drug safety and efficacy. At Ignota Labs, Lane leads innovation, leveraging his extensive background in biochemistry, genetics and AI to pioneer transformative treatments in global patient care.
The drug discovery industry is carbon, water and plastic intensive. There’s a lot of fantastic work underway to reduce the amount of solvents and plastics used during the process, but ultimately, this is likely to never reach zero. If your drug doesn’t make it through clinical trials, which happens to 90% of drug candidates, you must then repeat the process from the start. With AI-driven repurposing, you are essentially saving on carbon that would be used to start the process over.
Chief Data Science Officer and Co-Founder
Ignota Labs
Dr. Layla Hosseini-Gerami bridges the worlds of chemistry, biology and AI in her role as chief data science officer at Ignota Labs. At Ignota Labs, she has built upon these sophisticated methods to understand the mechanisms of drug toxicity and how to mitigate them. Her goal is to leverage AI not just to understand, but also to prevent adverse drug reactions, ensuring safer therapeutic options for patients.
We can also reduce the number of experiments that we need to carry out as part of the discovery process, because we have AI models that allow us to prioritize candidates. Without those models, we perhaps would have needed to perform additional screening experiments in the lab.
BF:
Senior Science Writer
Technology Networks
Blake pens and edits breaking news, articles and features on a broad range of scientific topics with a focus on drug discovery and biopharma. He earned an honors degree in chemistry from the University of Surrey. Blake also holds an MSc in chemistry from the University of Southampton. His research project focused on the synthesis of novel fluorescent dyes often used as chemical/bio-sensors and as photosensitizers in photodynamic therapy.
Resurrecting a previously shelved compound raises practical questions around ownership. What kinds of hurdles do you anticipate in getting these revived candidates back into development pipelines?
JL:
CSO and Co-Founder
Ignota Labs
Dr. Jordan Lane has extensive experience in the AI drug discovery space and is driven by a profound commitment to revolutionizing drug discovery, integrating artificial intelligence with bioscience to enhance drug safety and efficacy. At Ignota Labs, Lane leads innovation, leveraging his extensive background in biochemistry, genetics and AI to pioneer transformative treatments in global patient care.
We've been asked: Why don’t you patent the compound? Why do you go to the companies and in-license these instead? And our argument for that is that there’s a huge amount of benefit to understanding the work that’s gone on prior. The amount of time and effort the developers have put into understanding the compound means that we can stand on their shoulders and hopefully do the smallest amount of work possible to rescue these drugs. We’ve seen many times with our internal projects how beneficial that background of work is. It allows us to innovate both within and outside the developer's IP, with the confidence that we won't end up in conflict.
LHG:
Chief Data Science Officer and Co-Founder
Ignota Labs
Dr. Layla Hosseini-Gerami bridges the worlds of chemistry, biology and AI in her role as chief data science officer at Ignota Labs. At Ignota Labs, she has built upon these sophisticated methods to understand the mechanisms of drug toxicity and how to mitigate them. Her goal is to leverage AI not just to understand, but also to prevent adverse drug reactions, ensuring safer therapeutic options for patients.
It makes the turnaround process easier having that data and being able to model the structure-activity relationship.
BF:
Senior Science Writer
Technology Networks
Blake pens and edits breaking news, articles and features on a broad range of scientific topics with a focus on drug discovery and biopharma. He earned an honors degree in chemistry from the University of Surrey. Blake also holds an MSc in chemistry from the University of Southampton. His research project focused on the synthesis of novel fluorescent dyes often used as chemical/bio-sensors and as photosensitizers in photodynamic therapy.
Looking ahead, could AI-powered drug development become a mainstream part of pipelines?
JL:
CSO and Co-Founder
Ignota Labs
Dr. Jordan Lane has extensive experience in the AI drug discovery space and is driven by a profound commitment to revolutionizing drug discovery, integrating artificial intelligence with bioscience to enhance drug safety and efficacy. At Ignota Labs, Lane leads innovation, leveraging his extensive background in biochemistry, genetics and AI to pioneer transformative treatments in global patient care.
A lot of AI models are being democratized, so more people are using them. I believe that in 10–15 years, AI will become an integral part of standard workflows everywhere. I think the impact of AI is initially overstated, but in the long term, it is understated. It happens time and time again: a new technology becomes something you use daily, and people don’t realize or remember how much of an impact that innovation has had.
