2018
Daphne Koller's conversations with the investors who founded insitro began in early 2018 as she was transitioning out of Calico.
“they really started in early 18 as I was kind of starting to transition out of Calico.”
Topic
Insitro is a company led by Daphne Koller that applies AI to drug design and human biology, and the podcasts here cover how it started, how it works and what drugs it is developing. Koller told Luke Timmerman that insitro was founded in 2018, and she explained to NEJM AI Grand Rounds that the name joins "in silico" (computer) and "in vitro" (lab).
The conversations also cover insitro's partners, including Gilead Sciences, Eli Lilly and Bristol Myers Squibb, and its work on liver disease, ALS and eye disease. Koller told Timmerman that three programs are expected to enter the clinic within about 18 months. She also described how agentic AI systems shortened its experiment cycles from six to nine months to about eight weeks.
40 facts from 6 episodes, in date order. These are what speakers said on the shows, each linked to the moment it was said. They aren't independently verified.
2018
Daphne Koller's conversations with the investors who founded insitro began in early 2018 as she was transitioning out of Calico.
“they really started in early 18 as I was kind of starting to transition out of Calico.”
2018
Jim Tannenbaum of Foresight Capital was insitro's first investor, followed soon by Krishna Yeshwant of GV and Bob Nelson of Arch Venture Partners.
“So the first investor was Jim Tannenbaum. And then very soon thereafter, Krishna Yeshwant from GV and Bob Nelson from Arch.”
2018
insitro's initial investor group had five major investors, including Third Rock Ventures.
“So we ended up with five major investors, actually, in that initial group, with Third Rock being the fifth.”
2018
Daphne Koller left Calico in February 2018 and launched insitro.
“I ended up leaving Calico in February of 2018 and launching a company in situ”
2018
The name insitro combines 'in silico' (in the computer) and 'in vitro' (in the lab), according to Daphne Koller.
“it's the synthesis of in silico, which means in the computer and in vitro, which means inside the lab”
2018
Daphne Koller says insitro started as a one-person company funded with about $100 million from investors who had been seeking a machine learning drug discovery company.
“they said, yeah, here's $100 million, do something.”
2018
Insitro was founded in 2018 and has raised $750 million from venture capital and pharma partnerships.
“Insitro was founded in 2018 and has secured $750 million in funding from venture capital firms and through partnerships with big drug makers”
2020
insitro raised a $143 million Series B financing in May 2020, led by Andreessen Horowitz.
“as it raised a $143 million Series B financing in May 2020, led by Andreessen Horowitz, and which included earlier investors like ArchVenture Partners, Foresight Capital, GV, and Third Rock Ventures.”
2026
Daphne Koller said insitro's first drug, for the fatty liver disease MASH, would enter the clinic the year after the panel.
“Our first drug is heading into the clinic next year. Um, it's in the disease called MASH, which is a terrible fatty liver disease.”
2026
Insitro has three programs expected to enter the clinic within about 18 months.
“So we have three programs that are heading into the clinic in the next 18 months.”
2027
Insitro plans to file for its MASH program this year, with the first patient dosed in early 2027.
“we're going to file this year and first patient in will be early in 27.”
2028
Insitro hopes to bring its ALS antisense oligonucleotide into the clinic in early 2028.
“then we hope to get that into the clinic in early 2028.”
Undated
Daphne Koller says many insitro hires are roughly 70 percent life scientists who can program in Python, an approach she calls recruiting 'damp' scientists.
“They are 70 % life scientists, but they know how to program in Python and analyze their own data.”
Undated
insitro makes induced pluripotent stem cells from donors' cells and differentiates them into cell types such as neurons or liver cells for experiments.
“where we can take cells from you or me, convert them into pluripotent status. stem cells and then differentiate them into a Daphne neuron or a Daphne hepatocyte or any one of the set of lineages that the community has developed differentiation protocols for.”
Undated
insitro's liver disease work came out of its partnership with Gilead Sciences on NASH, non-alcoholic steatohepatitis.
“It came out of the partnership that we have with Gilead on NASH, non-alcoholic steatohepatitis.”
Undated
Daphne Koller says insitro relies heavily on unsupervised and self-supervised learning because there are too few successful drugs to use as training examples.
“We also have a much greater reliance on unsupervised and self-supervised machine learning algorithms where the need for supervised examples is much, much lower.”
Undated
Pande says he sits on the board of Insitro, a company led by Daphne that applies AI to drug design and human biology.
“And so InCitro is her vision for bringing AI to drug design.”
Undated
Daphne Koller says the name insitro reflects the integration of in silico (computer-based) work and in vitro (lab-based) work.
“insitro is the synthesis, the deep integration of in silico, which means in silico on the computer, and in vitro, which means in the lab.”
Undated
Daphne Koller says drug candidates fail more than 90% of the time once they enter clinical trials.
“And then you push that into the clinic and it fails over 90 % of the time.”
Undated
Daphne Koller says Eli Lilly makes a particular type of molecule for insitro because Lilly has capacity and capabilities insitro lacks.
“We have actually an kind of inverse collaboration with Lilly where they have extra capacity and capabilities in making a particular type of molecules that isn't within our sweet spot, and they're making that for us.”
Undated
Daphne Koller says regulatory toxicity requirements for a drug candidate take about a year to complete and are hard to speed up.
“there are a certain set of requirements that regulators require, for example, on toxicity that just take as long as it takes a year to do these experiments.”
Undated
Daphne Koller says insitro's stated values include 'create together' and 'engage with respect'.
“Create together is one of our values. Engage with respect is another one.”
Undated
Daphne Koller says insitro grew to about 50 people just as the pandemic began, which distracted the company from adding decision and governance processes.
“I will use as an excuse that when we hit 50 was also when the pandemic hit.”
Undated
Daphne Koller says insitro's 'pipeline through platform' strategy aims to build a repeatable engine that produces multiple drugs, not just one.
“we are committed to something we call pipeline through platform. And that means that we don't just make a drug that makes people better.”
Undated
Daphne Koller said competitors in physical AI have no shortcuts and must build custom hardware, data and models, which she described as a competitive moat.
“You have to build the custom hardware, you have to collect the custom data, you have to build the custom models that understand the physical world and causality.”
Undated
Insitro has onboarded about 28 petabytes of data, mostly cell-based.
“we have at this point onboarded, I think, 28 petabytes of data.”
Undated
Three to four petabytes of Insitro's data are human clinical data.
“we also have somewhere between three and four of those petabytes are human clinical data.”
Undated
Insitro partners with Moorfields Hospital in the UK, which holds one of the world's largest collections of OCT eye scan data.
“our partnership with the Moorfields Hospital in the UK that has one of the world's largest collections of OCT data, which is a form of eye scans.”
Undated
Insitro's partnership with Eli Lilly produced models predicting in vivo pharmacokinetic properties of small molecules.
“our really exciting partnership with Lilly that allowed us to predict in vivo pharmacokinetic properties for small molecules.”
Undated
Koller says agentic LLM systems cut Insitro's experimental design and execution from six to nine months to about eight weeks.
“took a process that used to take our scientists like six to nine months and shrunk it down to, you know, eight weeks.”
Undated
Gilead Sciences, Insitro's first partner, decided to exit liver disease and focus on virology and oncology.
“they are exiting the liver space and focusing on virology and oncology”
Undated
Insitro's disclosed MASH target is IRS-1.
“This one is IRS-1, and the genetics are incredibly strong for this.”
Undated
Insitro's MASH drug is a liver-targeted siRNA, designed because inhibiting IRS-1 throughout the body would be harmful.
“A liver-targeted siRNA, which is actually really critical because you would not want to inhibit IRS-1 systemically.”
Undated
Insitro disclosed IRS-1 because its patent on the oligonucleotide sequence revealed the target.
“We filed a patent in order to protect the composition of matter, And for an oligo therapeutic, the sequence that you are disclosing in the patent tells you the target.”
Undated
Koller says the siRNA for MASH could be dosed about every three to six months.
“SIRNAs have very long durability. This one is easily three to six months injection”
Undated
Koller says Insitro has found three ALS programs aimed at the roughly 97 percent of patients with TDP-43 pathology.
“we have been, we believe, able to find not one but three programs that truly are disease-modifying in the 97 % of ALS patients that have what is called TDP-43 pathology”
Undated
Bristol Myers Squibb chose to pursue the first ALS target as a small molecule, while Insitro advances it as an intrathecal oligonucleotide at its own cost.
“BMS elected to pursue the first of those targets, ALS-1, as a small molecule.”
Undated
Insitro's AI measurement of brown fat from whole-body MRI found people scanned in winter had more brown fat.
“if the MRI was collected in winter, you have more brown fat.”
Undated
A mouse model of one Insitro brown fat target lost 25 percent of its fat mass without losing lean mass or eating less.
“that mouse lost 25 % of its fat mass. It didn't lose any lean mass.”
Undated
Insitro says it trained the world's largest foundation model for OCT eye scan data, which yielded its dry AMD target.
“we were able to train the world's largest foundation model for oct data”
Episodes