Startup Advances AI Models for Tumor RNA Analysis

Blank Bio, a developer of high-quality, highly accurate sequencing solutions, is building RNA foundation. Models that learn directly from the molecular complexity of tumor transcriptomes to improve patient-level prediction across oncology. Proceeds from the financing will support continued model development, expanded collaborations with pharmaceutical and diagnostic companies, and new long-read RNA-seq datasets to support applications in biomarkers, clinical trial design, and diagnostics.

“Bulk RNA-seq is one of the most clinically accessible and information-rich assays in oncology, but much of its signal is still reduced to simplified gene-level summaries,” said Jonathan Hsu, CEO and Co-Founder of Blank Bio. “Blank Bio was founded to apply foundation models to the full molecular detail contained in each patient’s tumor transcriptome and turn that information into more precise, clinically useful predictions. This financing will support continued model development and partnership expansion, while our collaboration with PacBio will generate the high-resolution long-read RNA data needed to train and evaluate these models in patient tumor samples.”

Advancing RNA foundation models with high-resolution long-read data

For its strategic collaboration with PacBio, Blank Bio will generate PacBio HiFi long-read, bulk RNA sequencing data from up to 100 fresh frozen patient tumor samples spanning multiple cancer indications. Sequencing will be done at Seattle Children’s Research Institute, where the Kinnex RNA libraries are automated on the SPTLabtech firefly®+ platform. Blank Bio will use the data to further train and evaluate its models, focusing on oncology applications where incorporating RNA-level signals may improve patient stratification, biomarker discovery, and clinical interpretation.

“PacBio HiFi long-read sequencing was built to resolve biology that other technologies miss, and nowhere is that more consequential than in the complex transcriptomes of patient tumors,” said David Miller, Global Vice President of Marketing at PacBio. “Blank Bio’s foundation models demonstrate how high-resolution RNA data and machine learning can advance the next generation of precision oncology applications, from biomarkers and diagnostics to clinical trial design.”

Expanding the analytical foundation for clinical RNA-seq Bulk RNA-seq is increasingly used across oncology research, drug development, and clinical care because it captures the molecular state of tumor samples at a scale and cost that support clinical deployment. However, standard analytical workflows often compress RNA-seq data into per-gene count summaries, limiting their ability to capture isoform architecture, mutational complexity, and other features of patient-specific tumor biology. Blank Bio is using the seed financing to build the models, datasets, and partnerships that turn this signal into precision oncology products.

Seed financing to support model development, datasets, and partnerships

The oversubscribed round included Define Ventures, Leonis Capital, Nova Threshold, Ripple Ventures, SignalFire, Y Combinator, and others. Proceeds will support continued model development, expanded pharma and diagnostic collaborations, and new data generation efforts. 

“Blank Bio is building at the intersection of two major shifts in biology: the expanding clinical use of RNA-seq and the emergence of foundation models capable of learning complex biological patterns at scale,” said Sahir Raoof, TechBio advisor to SignalFire. “The company brings together deep scientific and technical expertise in RNA biology, machine learning, and oncology, with a platform that has the potential to turn transcriptomic data into a more powerful layer of patient-level insight for drug development and diagnostics.”

The Blank Bio team includes AI scientists and engineers from companies including Recursion, Deep Genomics, DeepMind, and Amazon, and institutions including Memorial Sloan Kettering Cancer Centre, Stanford, and the Vector Institute. The team has published work in Nature Methods, Nature Genetics, Nature Biotechnology, ICML, and NeurIPS, including Orthrus, its prior academic work on RNA foundation models, recently published in Nature Methods.