What I learned leading a data science & software team in biotech
You
You are a technical leader at a biotech startup. Your title is Director, or VP of Data Science / Machine Learning / Engineering, or CTO / CDO. Your remit is to develop a "platform" or "engine", an exciting opportunity that is now your responsibility. These posts are for you! I cover operations, culture, tactics, and more. I wish I had these experiences when I started, and I hope you find them useful. As always, feedback welcome!
Table of Contents
- Background
- Scientific or clinical results > platforms
- Biweekly digest
- Monthly retrospectives
- Build vs Buy
- Biotech & Tech have different vocab & mental models
- Track work in a single source of truth
- Build an advisory board
- Friction from legacy systems
Acknowledgements
Thanks for advice, intros, and support: @gpbatra, @dshaywitz, @BenGlicksberg, @jejomath, @merrickfurst, Anthony Philippakis, Brian Johnson, and more.
Opinions expressed below are solely the author’s and do not reflect the opinions and beliefs of the companies mentioned.
Background
In early 2021 I joined Prometheus Biosciences, a small biotech startup, as an early employee and the head of data science and software engineering ("DSE"). Our mission was to develop precision therapeutics for autoimmune disease, with an initial focus on inflammatory bowel disease (IBD). Fresh off a $130M fundraise, we had a validated drug target (TL1A), a pipeline of antibody candidates, and an exclusive license to data and samples in the Cedars-Sinai Medical Center's IBD biobank. Our ambitions in precision medicine were apparent in press releases at the time:
> The company's precision medicine platform, Prometheus 360 (P360), combines proprietary bioinformatics discovery methods with one of the world's largest gastrointestinal bioinformatics databases to identify novel therapeutic targets and develop therapeutic candidates to engage those targets.
My job was to build this. Over the next two years, we defined the product roadmap, recruited a small and stellar team of data scientists and software engineers, stood up our cloud architecture, and ingested TBs of multimodal biological data. Our infrastructure, capabilities, and expertise enabled target validation and companion diagnostics (CDx) development.
In March 2021, we IPO'd, announced positive phase 1 data, and expanded indications to systemic scleroderma. At the end of 2022, we announced positive phase 2 data in IBD. In 2023, after two years as the best-performing biotech stock, we were acquired by Merck for $10.8B (). This outcome was a win for patients and shareholders, but the journey was not without adversity, tribulation, and failure. Although your context will differ from mine, I hope this list of things I wish I knew serves as a helpful resource.merck.com/news/merck-com…
Apr 18, 2019 • 13 tweets • 9 min read
1/ Why is lactate elevated in sepsis or septic shock?
