Shane Woods, PhD: The term “TechBio” is gaining significant traction. Could you unpack what this philosophy means for drug development and how it represents a fundamental shift from traditional R&D?

Sid Jain: While traditional R&D is reliable, it can be slow to change, often tacking new technology onto existing, sometimes broken, processes. The TechBio advantage is building from the ground up with data-first workflows that let us iterate at machine speed. This allows Recursion to move faster and find novel biology at scale. If you think of the traditional model as a single ship liner, our approach is more like a modern fleet with better instrumentation, allowing us to try multiple routes and course-correct very quickly.

Shane Woods, PhD: This partnership brings together an immense scale of data. Beyond the headline number, how does combining Tempus’ real-world patient data with Recursion’s experimental data unlock scientific programs that were previously out of reach?

Forward genetics alone gives you correlation, while reverse genetics provides functional clarity but lacks patient context. The magic happens when you fuse them. The patient data tells you what genes are associated with a disease, and our perturbational data tells you why. Layering these two views into our causal AI models allows us to finally move from correlation to causation. This helps us amplify signals that would normally require hundreds of thousands of patient samples to detect, which is a massive unlock for rare diseases and small patient cohorts.

Shane Woods, PhD: Drug discovery faces undeniable challenges, from long timelines to high attrition rates. How are you applying this integrated TechBio approach in early research to tackle those core problems head-on?

By combining forward and reverse genetics, we can pursue novel biological targets with higher confidence because the insights are already grounded in patient data from day one. For patient populations, our causal AI models provide insights into not just causality, but also the context of where we should intervene. This has allowed us to identify novel expansion opportunities for preclinical programs and refine patient stratification by identifying clinically actionable subgroups and their associated biomarkers. We can kill bad ideas earlier and double down on the ones with a real shot at clinical relevance.

Shane Woods, PhD: As programs advance from discovery into the clinic, the stakes get higher. How are you leveraging this data-driven approach to de-risk key clinical development decisions, such as validating patient selection strategies or refining biomarker hypotheses?

We have used these integrated datasets to reprioritize indications for in-development assets, giving us more confidence in where to proceed and, just as importantly, where to stop. This is critical from both a probability of success and an ethical perspective. If we can better predict who will or will not respond, we can target trials to patients most likely to benefit. We have also used the data to refine inclusion and exclusion criteria in our protocols. For example, modifying a hemoglobin criterion can open the trial to a wider patient population, including minority groups that may have been excluded previously, increasing both the speed and the potential impact of the trial.

Shane Woods, PhD: Integrating massive, diverse datasets from two different organizations sounds great in theory, but the reality is often complex. Could you share some of the key technical and organizational hurdles your teams had to overcome to make this partnership work effectively?

Sid Jain: On the organizational side, culture is a true differentiator for Recursion. We have a nimble team that is willing to try new ideas at a speed that would be difficult in most other organizations. For example, we built a team and took an idea for our ClinTech approach from concept to use in 90 days. That culture of wanting to do things differently is a critical part of how we innovate. Any strong partnership goes through bumps, but we have always had the conviction that bringing our capabilities together would lead to amazing things.

Shane Woods, PhD: In an industry where success is measured over a decade, demonstrating near-term impact is a major challenge. How does Recursion approach measuring the ROI of these new tools, and what leading indicators do you track to build confidence along the way?

Hayley Donnella, PhD: We measure impact in two layers: scientific and operational. Scientifically, we look at our ability to recover known biological signals, which builds confidence in the novel insights our models generate. Operationally, we measure our “learning velocity”—how much validated insight we generate per cycle and how quickly that insight feeds back into the system. These small accelerations stack up and create a compounding advantage, teaching the system how to get better at discovery and translation for the next program.

Shane Woods, PhD: Looking ahead, what is your vision for where the TechBio revolution will lead the industry, and what can organizations do today to prepare for that future?

Sid Jain: My vision is for precision medicine to become a reality, whether that eventually looks like an N-of-1 trial or not. We want to industrialize development in a way that we can simulate and predict response, run adaptive trials, and find the right patient cohorts before the first patient is even enrolled. That is what the future of precision medicine looks like.

Please note that the content in this document has been revised for clarity and conciseness. Some language and formatting may have been adjusted to enhance readability while preserving the original meaning and intent of the discussion.