Zeenome is backed by Ethereum, the world's largest public programmable blockchain. The cryptographic data that powers the protocol is freely available and never goes offline.
Any institution, clinic, sequencing company, researcher, or patient can join the network. Publicly available whitelists curated by trusted authorities let you interact with accredited network participants.
Zeenome uses frontier cryptography so patients can transmit the data points a researcher needs without exposing their underlying genome. Researchers can design studies that maximize privacy—the researcher never learns the patient's identity.
Most genomic data sharing networks are built around "queries". But what if you could run a program against an entire population? Test a model in the real world and compose past experiments into new ones to unlock additional insights.
Zeenome is a protocol, not a company. It is developed as an open standard and designed to maintain neutrality. It strives to be a technology that transcends nation-centric institutions to address global biases in genomics.
You open a letter—your premium just jumped because someone inferred serious disease risk from your DNA.
In genomics it's important to balance the benefits of data sharing with privacy.
Bridges and standards have been built between institutions so genomic data can stay with your healthcare provider while enabling access for research.
But this model has limitations.
Institutions create networks that allow researchers to query data, but an infiltration in one network can grant an attacker access to the entire system.
Genomics is one of the largest sources of data on earth. Institutions have difficulty absorbing the costs of hosting this data indefinitely.
Genomic research must cross national borders to capture the diversity of the human species, but building international networks introduces additional frictions.
When a large quantity of data is centralized it becomes a prime target for attack. Over time the expense of storing data grows, and vigilant monitoring is needed to prevent unauthorized access.
Networks skew toward clinical and biobank samples, so researchers lack broad access to healthy-population baselines. That gap warps controls, validation, and models that need a representative background.
Transfer a patient's raw genomic data to them on purpose-built hardware that connects to their home Wi-Fi. Patients could alternatively choose a cloud provider to custody their data.
Institutions publish attestations on Ethereum (signed messages using a private key) that represent the digital fingerprint of a patient's genome and phenotype. This proves they are the source of the data.
Attestations contain no information that identifies the patient.
Researchers craft programs that run within a zero-knowledge (ZK) virtual machine. These programs are simple functions that accept an array of requested alleles and phenotypes as parameters, perform arbitrary computation and return a result.
Example Code
pub fn predict(
&self,
snp_data: &[SnpData],
hpo_codes: Option<&[&str]>
) -> Result<InquiryResult, InquiryError> {
let mut hpo_match = false;
let mut hpo_confidence = 0.0;
let num_causal_snps = Vec::new();
for model in &self.snp_models {
...
}
Ok(InquiryResult {
hpo_match,
hpo_confidence,
num_causal_snps
})
}The researcher publishes their compiled program alongside metadata identifying them, their institution, and a description of the inquiry so patients can give informed consent.
The researcher can limit inquiry responses to patients whose genome and phenotype have been attested to by one or more trusted sources.
An authoritative body can examine the inquiry and give it a score so patients understand if the inquiry adheres to patient privacy best practices.
A patient's device detects the newly published inquiry and notifies them if they are eligible to respond.
The patient reviews the details of the inquiry and if they consent, their hardware device executes the inquiry and anonymously transmits the results with a zero-knowledge proof back to the researcher.
When a researcher receives a response, the provided zero-knowledge proof is verified. This verification guarantees that the patient's hardware ran the inquiry program correctly.
The data sent by the patient does not include any stable identifier that would link them across subsequent queries. It does however prevent patients from submitting duplicate responses.