BIO 2026: Generative genomics helps biotechs design better, not just faster

Generative genomics uses AI to analyze genetic sequences for innovative treatments.

Every industry is racing to optimize AI in their operations. For many biotech research teams and investors that means “generative genomics.”

Generative genomics is an emerging field of study using AI to analyze genetic sequences. Researchers hope that, with AI, they can efficiently assess the world’s collected stores of DNA/RNA data, discovering overlooked patterns that lead to new treatments. 

Experts believe this technology will lead to faster results and increase the success rate of clinical trials. The possibilities were discussed during a panel at the BIO International Convention (BIO 2026) entitled “How Can Generative Genomics Help Us Design Biology Better, Not Just Faster.”  

Cause for excitement

When it comes to reducing time and costs of developing medicine, there is potential for major efficiency gains. It can take more than a decade for a new product to go from the planning stage to rollout, and according to Genentech CEO Ashley Margagee, the cost of developing a drug is $2 billion on average.

The most optimistic proponents of generative genomics say AI could shorten drug development time from years to months, creating money savings too.

Accuracy could also improve. Currently, 90% of clinical trials fail, a figure that has remained stubbornly stable despite decades of advancements in the industry. Any change in that situation could be impactful.

“Just increasing success by five percent is a massive deal,” said John Androsavich, General Manager of Ginkgo Datapoints at Ginkgo Bioworks. “Investments that move the success rate by small increments would be huge.” 

Generative genomics could be the technology that finally moves the needle, according to experts. AI categorizes and processes data at a speed never seen before. As the global biotech industry continues producing more data, AI can recognize patterns from across different sources and present its own conclusions. It will become easier than ever before for scientists to learn from research conducted by others. 

Ensuring trust and quality

Biotech is a high-stakes industry: a literal matter of life and death. For biotech researchers to engage with generative genomics to its full potential, their model must be able to earn their trust.

“If the most beautiful, accurate model doesn’t change a decision, it’s not useful,” said Julie Rytlewski, a Senior Director at Bristol Myers Squibb. 

She explained the real proof of success will not come from generating accurate hypotheses, but instead when those hypotheses start changing researchers’ decisions. 

“There has to be trust in what these models are doing, especially when it comes to treating people,” Rytlewski said. “As we build trust, people value the output of the models more. It won’t be a switch overnight.”

The panelists agreed that developers cannot—and should not—rush the process in which AI wins over its skeptics. 

“There is no single modality that can unlock biology,” said Justin Guinney, Senior Vice President of Cancer Genomics at Tempus AI. “It requires a multimodal perspective.”

‘We’re seed-planters, not tree-buyers’

It is no secret that AI is a darling of investors, and their enthusiasm extends to generative genomics’ potential. But with so many options to choose from, companies have to be strategic about which platforms they rely upon.

Danjuma Quarless, Senior Director of AI Innovation at Lilly Ventures gave the BIO 2026 panel insight on the industry’s decision-making process. Lilly Ventures, as the investment branch of biotech titan Eli Lilly, is tasked with “intersecting with the next great generation of biotechs” ahead of their rise to prominence. His strategy is to partner with as many promising developers as possible, a process which involves much speculation.

“I like to say that we are seed-planters, not tree buyers. We foster many small options without fear of redundancy.” 

Rytlewski agreed, and explained why large companies are taking chances with early-stage models.

“One of key ideas is blended innovation: it’s not ‘build’ or ‘buy’ but partnering. It’s not uncommon for us to place multiple bets. It’s all about finding the right tool for the right ambition.”

This tactic of backing several developers at once is indicative of the optimism that established biotech companies have in generative genomics. It reflects the confidence that the industry will prove incredibly lucrative, and that it is better to pick the eventual winning companies early on.

“Most foundation models are in the ‘preclinical stage,’” said Jeff Leek, Co-CEO of Synthesize BIO, which is developing its own generative genomics model. “We are at the very beginning of the game.”

Much has been said about the ways AI will reshape the world, but its introduction into biotech could lead to unprecedented streamlining of the discovery to approval pipeline. And in biotech, time saved means lives saved.

The post BIO 2026: Generative genomics helps biotechs design better, not just faster appeared first on Bio.News.