insitro

What can AI do for drug discovery today, and what can’t it do yet? This week on his podcast, The Long Run, @ldtimmerman asked @DaphneKoller for a reality check, and she gave him one.

Daphne explains why AI companies are building wet labs and how human genetics links experiments in cells to what happens in people. She walks through the programs insitro is advancing in liver disease, ALS and brown fat, and closes with the case for public and philanthropic investment in richer human data.

Luke calls it a conversation worth listening to “for a balanced view from someone in the trenches who is both optimistic and realistic.”

🎧Apple Podcasts: podcasts.apple.com/us/podcast/ep2…

🎧Spotify: open.spotify.com/episode/3uv72l…

#DrugDiscovery #AI #Biotech

Out now: “Getting Data to Grow in a Dish,” the third issue of Deep Phenotype, insitro’s Substack on scaled biology and deep causality.

Generating biological data at scale is one challenge. Ensuring that data is reliable, reproducible and meaningful enough to power AI-driven drug discovery is another.

In our third installment, insitro founder and CEO @DaphneKoller and insitro director of integrated technology exploration, @ATGmsalickTAG, take readers behind the scenes of building insitro's biological data factory, where automation, AI, and human expertise come together to generate high-quality data and deepen our understanding of disease biology.

Through the evolution of our Pooled Optical Screening in Human cells (POSH) platform, Daphne explores the extensive scientific and engineering advances required to turn complex cellular experiments into scalable, reproducible workflows.

Along the way, the share of cells we could match to their genetic perturbation more than doubled, from 34% to above 71%, and developing a new cell-segmentation model went from six to twelve months to about a month.

“The most important thing we have built is not any single platform. It is the ability to build the next one.”
╰┈➤ Read the latest installment of Deep Phenotype: deepphenotype.substack.com/p/getting-data…
#DrugDiscovery #AI #Biology #CausalBiology

At insitro, we proved again we don't spook easy at our annual Halloween costume contest last week. Every team that participated shared the same stretch goal: win.

Protein Sciences modeled the central dogma of molecular biology with paper bases and arrows. Read the costumes from left to right, and you could follow the path from DNA to RNA to protein. MASH rounded up the classic monsters, Op Ex and EH&S suited up as the OSHA-Busters, and our distributed LIMS team grew a field of broccoli one video tile at a time.

The fiercest rivalry was between the Legal and People teams. Legal’s take on “The Wizard of Oz” and “Wicked” featured an AI-generated Tin Man inspired by pooled optical screening and cellular perturbation. The People team deployed Ted Lasso and AFC Richmond, setting off a wave of spirited banter and tactical maneuvering. Then came the twist. As “Ted Lasso” writer and actor Brett Goldstein has explained, the show is a retelling of “The Wizard of Oz,” with Ted as Dorothy and Roy Kent as the Tin Man.

The competition was intense, but when the yellow brick road met the pitch, camaraderie was the real winner.

#Halloween #Halloween2026 #costume #sciencepuns #insitro

Pattern recognition tells you what’s associated with disease, not what drives it. At #AIDDD2026, insitro CSO Philip Tagari presents “Generative AI for Target Hypothesis Design: Moving Beyond Pattern Recognition,” plus Workshop 5 on building techbio at scale, Oct 27.

Today, insitro announced a strategic collaboration with @AstraZeneca for the AI-enabled discovery of novel therapeutics for metabolic dysfunction-associated steatohepatitis (MASH) and cardiomyopathies.

insitro will receive $40 million upfront and is eligible for up to approximately $1.46 billion in potential, discovery, development and commercial milestones, plus tiered royalties.

Both diseases are defined by the problem the collaboration is built to address. In each case, a single diagnostic label spans patients whose underlying biology differs. The heterogeneity across these diseases has made them difficult to treat, and the condition under which insitro’s machine learning capabilities on human data is most useful.

Under the agreement, insitro will identify causal disease biology in both indications and advance AstraZeneca-nominated targets from discovery through to development candidate; AstraZeneca will lead clinical development, regulatory work, and commercialization.

“Partnering with AstraZeneca allows us to pair our deep biological precision with their deep pharmaceutical expertise and scale with potential for accelerated advancement of novel programs through development and to patients.” — @DaphneKoller, founder and CEO of insitro

Sharon Barr, Executive Vice President and Head of BioPharmaceuticals R&D, AstraZeneca: "Partnering with insitro, we are using AI to identify the biological drivers of disease to support the rapid discovery of high-quality novel molecules, advancing our ambition to deliver the next generation of medicines to patients faster.”

Read the announcement: insitro.com/news/insitro-a…

More on our approach to causal biology: deepphenotype.substack.com
#AIDrugDiscovery #MASH #Cardiomyopathies