Why a Startup Just Gave Away Four Million Living Human Neurons for Free: Crownlands, Olfactory Cells, CZ CELLxGENE, and the Open Data Land Grab Reshaping Alzheimer’s Drug Discovery and AI Biology
Video Preview
🎧 Podcast episode for paid subscribers only. Also available on Spotify.
To listen to paid episodes in Apple or Spotify, link your Substack subscription via the show settings on those platforms (instructions inside the Substack app under Subscriptions → Podcast).
Abstract
What happened: Crownlands, a San Francisco discovery-tools startup backed by Caffeinated Capital, open sourced a 4 million cell single-cell dataset built from olfactory neurons collected from 200-plus living donors, including healthy controls plus Alzheimer’s, Parkinson’s, and other neurodegenerative patients. Posted free to CZ CELLxGENE. Largest single-cell dataset on that platform from living humans.
How: Gateway platform = proprietary collection device + 10x Genomics Flex Apex probe-based single-cell sequencing + centralized processing across multiple IRB-approved sites over a 10-month window.
Why it matters for this audience:
The brain has been a black box because you cannot biopsy living human cortex at scale. Olfactory neurons are the workaround.
The data, not the algorithm, is the bottleneck for AI drug discovery in neuro. This is a data-supply move dressed as philanthropy.
Open sourcing the static dataset while keeping the device, the longitudinal pipeline, and the models is textbook commoditize-the-complement.
CZ CELLxGENE distribution turns a press release into a standard. Atlas positioning is the actual prize.
Bottom line: free data is the loss leader. The longitudinal, patient-linked, repeatable collection engine is the business.
Table of Contents
The thing that actually happened
Why nobody had four million living brain cells lying around
The nose as a back door to the brain
What is actually in the dataset and why experts should care
The amyloid casino and why target discovery has been starved
Open sourcing as a business model, not charity
The CZ CELLxGENE play and the atlas wars
The co-scientist agent and the new shape of a moat
What could go wrong, and what skeptics will say
What to watch next
The thing that actually happened
Strip away the launch-day adjectives and here is the event. A previously stealthy company came out of the gate by dumping 4 million single cells onto a public repository, free, no login wall, no academic-collaboration handshake, no data-use agreement that takes six months and a lawyer to clear. The cells came from more than 200 living people. Some healthy, some carrying Alzheimer’s, Parkinson’s, and other neurodegenerative diagnoses. The collection ran across multiple clinical sites under IRB approval and took roughly 10 months end to end. They built it on 10x Genomics Flex Apex, which is the probe-based, fixed-sample chemistry, and they parked the whole thing on the Chan Zuckerberg Initiative’s CZ CELLxGENE platform where any grad student with a laptop can pull it.
The framing in the release is that this is the largest single-cell dataset from living human donors on that platform. That qualifier is doing a lot of work, and it is the whole point, so hold onto it. There is a mountain of single-cell brain data already public. Almost none of it is from living people. That distinction is the entire reason this is interesting and not just another atlas drop nobody outside a sequencing core would notice.
They also shipped a natural-language query agent at chat.crownlands.com that lets you interrogate the full dataset by gene, pathway, or cell type without writing a line of Python. And there is a companion bioRxiv preprint, “Gateway: patient olfactory neurons for large-scale discovery in neurodegenerative disease,” carrying the claim that olfactory neurons pulled from a living nose actually reflect a meaningful chunk of the gene expression you would see in living brain tissue. CEO Nate Dalva is selling the thesis that the next wave of progress in neuro follows data availability, not model cleverness. Co-founder Kevin Zhu, the Head of Discovery, is the one putting his name on the biology claim. That division of labor tells you something about where the risk sits.
