Synthesize Bio’s generative genomics models predict the results of gene expression experiments with unprecedented accuracy

Funding led by Madrona with participation from AI2 Incubator, Sahsen Ventures, Inner Loop Capital and Point Field Partners

Synthesize Bio is actively engaging with biopharma partners to enable more confident drug development decisions using generative genomics

GEM-1 foundation model now available at Synthesize Bio and through R and Python API clients

SEATTLE–(BUSINESS WIRE)–Synthesize Bio, a biotechnology company pioneering biological foundation models to simulate the results of gene expression experiments, today announced a $10 million seed round led by Madrona, with participation from Sahsen Ventures, Inner Loop Capital, Point Field Partners, and AI2 Incubator. The funding will accelerate the development of Synthesize Bio’s generative genomics models and expand access for biopharmaceutical companies and researchers.

“Biopharma companies need rich, representative data to identify new drug targets, validate drug safety and efficacy, and power foundation models. Our models help scientists predict the results of new and currently impossible experiments to move their science forward much faster and much more cost effectively,” said Rob Bradley, PhD, Co-founder of Synthesize Bio.

Synthesize Bio has built Generate Expression Model-1 (GEM-1), a generative genomics foundation model trained on one of the most deeply curated RNA-seq datasets ever assembled. The company’s recent publication demonstrates GEM-1’s ability to generate in silico data that matches wet lab experiments directly from experimental design descriptions. This achievement represents a new paradigm of “generative genomics” in which high fidelity generative models augment laboratory results, predict future experimental results and clinical trial outcomes, and ultimately accelerate innovation.

Founded by Dr. Bradley and Jeff Leek, PhD, Synthesize Bio empowers researchers to accelerate drug discovery and other applications that rely on gene expression data, without being limited by the scarcity or bias of lab or clinical datasets.

Dr. Leek is the J Orin Edson Foundation Endowed Chair and Chief Data Officer at Fred Hutchinson Cancer Center where his work focuses on RNA informatics. He has spearheaded major efforts to assemble, normalize, and combine disparate RNA datasets from researchers around the globe into the largest combined dataset available. He was named to the Time AI 100 2025 list for his work bringing cancer centers together to federate patient data.

Dr. Bradley is the McIlwain Family Endowed Chair and Director of the Translational Data Science Integrated Research Center at Fred Hutchinson Cancer Center. His research has revealed that RNA dysregulation is a common cause of cancer initiation and identified new therapeutics for treating these cancers.

“Madrona’s investment builds on our longstanding thesis at the intersection of AI and life sciences. We believe Synthesize Bio represents a transformative shift in how biopharma and researchers efficiently discover and develop new therapies,” said Matt McIlwain, Madrona Managing Director. “Rob, Jeff, and the team are uniquely equipped to bring generative genomics into practice, and we are thrilled to partner with them.”

The Generate Expression Model-1 (GEM-1)

The performance of Synthesize Bio’s first-generation foundation model is described in a preprint available on bioRxiv. It describes GEM-1’s successful prediction of the results of laboratory experiments that were performed after its training data cutoff as well as the AI-powered generation of data from large clinical cohorts, providing the first indication that novel studies performed fully in silico can produce results comparable to those from a wet lab or clinical trial.

The Synthesize Bio team is developing partnerships with biopharma teams to accelerate drug development using their foundation models. These partnerships leverage the company’s foundation models to de-risk the entire clinical pipeline, from identifying high-confidence targets to simulating therapeutic responses and optimizing clinical trial design. Access to the GEM-1 foundation model is now available at Synthesize Bio and through R and Python API clients.

About Synthesize Bio

Synthesize Bio is pioneering the use of AI for generative genomics. Its first foundation model, Generate Expression Model-1 (GEM-1), is built on the world’s most comprehensive annotated gene expression data. GEM-1 can generate gene expression data based on experimental design descriptions of future and even currently impossible experiments for scientific discovery and drug development. Founded by Fred Hutchinson Cancer Center scientists Rob Bradley and Jeff Leek, the company’s platform provides the world’s leading generative genomics RNA-seq data to power generative genomics and AI-native biopharma innovation. Additional information can be found at www.synthesize.bio.

About Madrona

Madrona (www.madrona.com) is a Seattle-based venture capital firm with 30 years of experience investing in early-stage technology companies. Madrona partners with founders from day one, providing long-term support to build innovative, industry-leading businesses for the long run. The firm has more than $3 billion under management, and was an early investor in companies such as Amazon, Snowflake, UIPath, Rover, Redfin, and Smartsheet. Rooted in the Pacific Northwest, Madrona also backs companies across the country, focusing on sectors such as AI, intelligent applications, and life sciences.

Contacts

Business Development: partnerships@synthesize.bio
Press: synthesizebio@tenbridgecommunications.com

(c)2025 Business Wire, Inc., All rights reserved.