Recursion and Genentech Join Forces to Develop Neuroscience Targets Using AI

Recursion has achieved a significant milestone in its collaboration with Roche and its Genentech subsidiary, announcing plans to co-develop a neuroscience small molecule discovery program. This initiative is based on the first validated target identified through an AI map designed to explore new targets in a complex therapeutic area. Genentech has exercised its option for this validated target after receiving a validation package from Recursion, triggering a $3 million milestone payment, which brings the total cash paid out to Recursion to $216 million since the partnership began in 2021.

Recursion stands to gain up to $300 million for each program developed in connection with this collaboration, which encompasses 40 programs focusing on key areas of neuroscience and an undisclosed gastrointestinal oncology indication. The specific target and the neurological disorders it may address have not yet been disclosed. However, Recursion noted that it took 15 months to move from the initiation of target validation to the creation of a validation package.

With over three billion people worldwide affected by neurological conditions, the need for new treatments is critical. Historically, only one in 40 neuroscience drugs that reach clinical trials receives approval, with only 8.4% of neurology drug candidates progressing from Phase I studies to approval, according to a study by the Biotechnology Innovation Organization (BIO).

“Finding new targets in neuroscience has historically been challenging, and this milestone highlights our ability to uncover novel biology in areas where conventional approaches have struggled,” stated Recursion CEO Najat Khan, PhD.

The advancement in neuroscience drug development is attributed to new tools aimed at studying the largely unexplored genome within living neurons. These tools include Recursion’s Data Factory, a comprehensive platform built over a decade to generate biological data tailored for AI models.

“We really took an approach to be unbiased in our full genome-wide scan of these opportunities to uncover new biology,” explained Christopher Winrow, PhD, Recursion’s vice president of neuroscience.

Through their partnership, Recursion and Genentech developed the first whole-genome CRISPR knockout map from over one trillion neuronal cells derived from induced pluripotent stem cells (iPSCs). This scale was necessary due to challenges faced by external cell manufacturers in meeting production demands.

“The cell context is important. We’re starting with human iPSCs and driving these into a very clearly homogenous population of neurons that we can test—and we do this at scale,” Winrow said. “You need a certain differentiation period for the iPSCs to form into the neurons that you want to study. There’s also a certain QC [quality control] that you need to do to make sure that those neurons are what we expect them to be, and that that’s robust and reproducible.”

Recursion implemented a rigorous validation process in collaboration with Genentech, advancing candidate targets through various stages of pathway, functional, and disease validation to ascertain their potential impact on neurological diseases. Researchers employed a large-scale perturbation set using whole-genome CRISPR-Cas9 knockouts to analyze interactions between genes and compounds.

“We’re looking at whole genome-wide knockout, not just a handful of areas or pathways of interest. That’s really a big game changer, in that we have this broad view,” Winrow added.

The subsequent imaging process aims to capture over 46 million cellular images, analyzing hundreds of features with the assistance of AI foundational models.

“What even is more mind-blowing beyond that is there are multiple features within each of those images,” Winrow noted. “Traditionally as a scientist, I’d go in and I’m really interested in mitochondria. So, I look at the image and I say ‘Wow, am I seeing mitochondrial fragmentation?’ Well, that’s great. Yes, I am, right? But if you’re looking at hundreds and hundreds of different features, that could be a mitochondrial shape, it could be nuclear capacity, it could be a whole bunch of different things that the AI models are trained upon, that then enables a real richness to come out of that dataset.”

The data generated is analyzed by Recursion’s BioHive-2 supercomputer, which was completed in 2024, to identify disease biology patterns and formulate hypotheses that guide further research.

“Each target is individual. They are related to a core disease process,” Winrow explained. “We have these core areas of biology that are looked at by a lot of different groups. We want to understand the unexplored biology around those areas, and that can be anything. It’s an unbiased approach, so all of those things that are associated have not necessarily been connected to this target in the past, so you come up with all sorts of different targets.”

Looking ahead, the partners plan to progress the discovery program from target identification to drug development, utilizing Recursion's chemistry platform for the design of potential first-in-class molecules while also identifying and validating additional targets.

Recursion intends to integrate its phenomics dataset with Genentech’s transcriptomics data to create further multimodal maps that will explore new potential targets and pathways.

The neuronal map is one of two whole genome neuroscience phenomaps produced by the Recursion-Roche-Genentech partnership, with the other focusing on specialized microglial immune cells, which also aims to uncover new targets in neurodegenerative diseases.

“This provides us really two distinct but CNS-focused maps that are rooted in human biology. I think the opportunities are there to really mine both of those,” Winrow remarked. “The neuro map is a little bit ahead, and so we’ve been able to take some of the learnings that we gained there and apply that to our work on the microglia side as well.”

Both neuro-focused maps are part of a total of six whole genome phenomaps created to facilitate the exploration of undiscovered targets and pathways, linking gene perturbations to cellular phenotypes.

“Because we can reuse these maps again and again, I might find something in the microglia map that’s really intriguing that I can now follow up in the neuron map, and vice versa,” Winrow concluded. “I think that’s a real power of this approach as well.”