As drug developers work toward getting their products approved, 30% of drugs fail clinical trials due to toxicity concerns. While many datasets exist with valuable information about toxicity that could assist developers’ efforts, these resources are often overlooked and underutilized.
Sara Selitsky, who has spent her career to date focusing on biotech and bioinformatics, founded FlashPath to organize and evaluate these datasets to more easily predict toxicity.
FlashPath curates, consolidates and analyzes various datasets regarding toxicology within its platform. Some of these datasets are sourced from the U.S. Food and Drug Administration, which makes them publicly available, and some are from proprietary sources.
With relevant and comprehensive information fed into the startup’s platform, companies using Flashpath can browse different chemicals and see what kind of adverse effects or safety hazards are associated with them.
“I was inspired to make FlashPath because there was a lot of data that existed across the toxicology space that has a lot of value and can really be trained on to make strong predictions for drug development, agriculture and other kinds of sectors,” Selitsky said. “But people weren’t really looking at it.”
Providing access and analytics
FlashPath is seeking to create the most comprehensive platform for toxicology data. The startup differentiates itself both in its organization of a variety of datasets and through its use of AI models to analyze those datasets, Selitsky said.
The startup is developing its own AI models based on the datasets already at its disposal, and those models will in turn be used to analyze information. This will enable FlashPath’s client companies to enter the exact chemical structures relating to their drugs and receive in turn an accurate prediction of toxicity (or lack thereof).
Companies will also be able to use their own models to analyze certain datasets on FlashPath’s platform, provided they have a license allowing them access.
“Our ultimate goal is that they can put in any chemical structure and get out a prediction of toxicity,” Selitsky said.
Being able to predict toxicity using FlashPath will save companies between tens of millions to a couple billion dollars, Selitsky said. That assertion is based on the fact that drug developers often have to conduct many iterations of tests to predict toxicity, and the expenses of that process add up quickly. There is also harm involved, given that the tests are often conducted on animals prior to human clinical trials.
Toxicity can also be detected later in the process, once clinical trials begin, causing companies to potentially lose millions of dollars on testing for a product that cannot go to market in its current form.
“If you could have a better predictor of toxicity ahead of time, you could maybe know better what will be safe,” Selitsky said.
Focusing on the biotechnology market
Drug companies can access the FlashPath platform, datasets and AI models through a subscription. The startup is seeking to attract pharmaceutical companies and drug developers to use their platform, with a particular emphasis on targeting smaller drug developers.
FlashPath also aims to partner with both private companies and government entities to use their data to improve the AI models.
Selitsky, meanwhile, recently wrapped up stint in the Center for Entrepreneurial Development’s GRO incubator. She pitched the startup to investors and others from the Triangle entrepreneurial ecosystem at the incubator’s fall “Demo Day” (recapped here).
Selitsky said participating in the incubator helped her learn how to pitch FlashPath as a business, rather than just a science startup.
“They’ve just given us instruction and tools, and hearing pretty complex concepts into digestible concepts has been a huge help,” Selitsky said.
QUICK BITS
Startup: FlashPath
Founder: Sara Selitsky
Founded: 2025
Team size: 2
Location: Durham
Website: flashpathai.com
Funding: Raising pre-seed
While FlashPath is currently focused on honing in on the biotechnology market in the form of toxicology data, Selitsky said the platform’s organization and analysis of datasets has potential applications beyond pharmaceutical companies and drug developers.
FlashPath’s technology could also be applied to fields such as artificial intelligence companies or agricultural technology, she said.
The startup is currently seeking pre-seed funding to further develop and expand the preliminary FlashPath platform.
“If we can make those predictions and save those lives and save time and money, like all of that, I think is really powerful, as well as saving human lives,” Selitsky said.
