platoseedplatoseed
AI and Data for Cancer Therapeutics
Creating the largest patient-derived cancer single-cell dataset for licensing to pharmaceutical and AI companies. These datasets are unique as they screen the effects of various drugs and genetic medicines under the control of our DNA switches.
Origin Bio develops perturbation biology in real human tumor tissue, using patient-derived samples to map cancer biology and build precision therapeutics. They collect tumor tissue, create precision-cut slices, perturb them with various agents, and apply multimodal readouts to train AI models for target identification, in-silico screening, and biomarker discovery.
Tissue is collected from partner hospitals into anonymized, annotated datasets. Fresh tissue is processed into precision-cut tumor slices that preserve the tumor microenvironment. Slices are divided into control and perturbed arms, with compounds including checkpoint inhibitors, chemotherapeutics, cytokines, and more. Multimodal readouts (transcriptomic, proteomic, histological assays) are performed to capture perturbation responses. The resulting data trains foundation AI models to enable target identification, in-silico drug screening, patient stratification, and biomarker discovery, with each experiment informing subsequent data collection and perturbations.
Who it’s for: biopharma companies and research institutions focused on cancer therapeutics, precision oncology labs, and translational research groups leveraging tumor perturbation data
Backed by Y Combinator; notable research activity and AI model development indicate traction and early-stage funding signals
CEO @ Origin | Computer Science @ UIUC, Computer Vision and Reinforcement Learning Research, First Prize 2022 OpenCV AI Research Competition.
CTO @ Origin | Prev: Computer Science @ UIUC, ML Research @ Wadhwani AI, Automorphic (YC S23). Published disease modeling research done in high school in Nature Scientific Reports.
AI designed regulatory DNA sequences to program gene expression patterns
Origin announces Axis, an AI model that designs regulatory DNA sequences to activate therapeutic genes in target disease cell-states, aiming to make cell and gene therapies safer and more effective. They are building a large proprietary dataset of synthetic regulatory elements and claim Axis outperforms Google DeepMind’s AlphaGenome in predictive benchmarks.
The agentic drug company. We simulate human biology.
Multimodal foundation models to predict uncollected patient biology
