Health IT & data platforms · Microbial genomics

One Codex

A cloud platform that takes a raw sequencing file and names the microbes inside it. Started in a Y Combinator batch in 2014, bought by Invitae for $83.4 million in 2021, sold back out nineteen months later, and still shipping.

Three thousand E. coli

In early 2013 a data scientist whose last job had been fraud prevention software sat down with a computational geneticist who had just finished a PhD in genome sciences at the University of Washington. Nick Greenfield and Nik Krumm were not looking for a biology problem. They were looking at a filing problem.

Sequencing had gotten cheap, reference libraries enormous. The field’s workhorse search tool, BLAST, had been written in the 1990s assuming those libraries would stay small.

“A lot of genomics has operated on this model of one default reference, or a few default references,” Greenfield told Bio-IT World in 2015. That works when there is one canonical genome per species. In microbiology there is not.

“You had 3,000 E. coli references, and a further 2,000 Salmonella references, and those numbers are ballooning very quickly.”

So they built an index instead of a search. The One Codex classifier chops a read into overlapping k-mers and looks each one up in a key-value store where the keys are k-mers and the values are the organisms they came from. It never reassembles anything. It narrows until it runs out of confidence, then stops at the last taxonomic rank it can defend.

Speed was the payoff. The company told TechCrunch in 2014 that the comfortable number against BLAST was “somewhere in the 1000 to 1500 [times faster] plus range.” By early 2015 the index held roughly 28,000 genomes and more than 1,000 user accounts, and it was free.

The government noticed early

In late January 2015 the CDC awarded Reference Genomics $200,000 in its “No-Petri-Dish” challenge, for strain typing Shiga toxin producing E. coli straight out of clinical samples. The agency said the platform “can rapidly identify STEC from complex clinical samples and provide meaningful information about its strain type and characteristics, even when the pathogens are present at levels too low to support assembly-based methods.”

Two years later the Defense Threat Reduction Agency booked a small SBIR Phase I purchase order, $149,871, running March to October 2017. In 2020 the US Geological Survey opened a purchase order for bioinformatic analysis of amplicon sequences on which no money was ever obligated. Those two contracts, the prize and a trademark are the entire federal paper trail. No NIH awards, no NSF awards, no FDA clearances, no registered trials.

Greenfield’s pitch to public health was about what sequencing had quietly taken away. “We used to live in a world where the primary diagnostic tool involved culturing, and people sent those cultures to CDC, and they built big biobanks of freezers,” he said. Molecular tests replaced the freezers with files. “The second-order intelligence is often being lost.”

Standards, a pandemic, then a buyer

In May 2017 ATCC, which maintains one of the world’s largest microbial strain collections, paired its physical sequencing standards with the platform so labs could score their own runs against a known answer. In late 2019 the two launched the ATCC Genome Portal, an ISO 9001 compliant reference database powered by One Codex.

In March 2020 One Codex made its SARS-CoV-2 analysis free and automatic on any sample where the virus turned up, asking only that users share metadata and consensus sequences. In January 2021 it signed Transnetyx as a sequencing channel. Four weeks later, Invitae bought the company.

Nineteen months inside a public company

Invitae acquired 100 percent of Reference Genomics in February 2021 for $17.3 million in cash and 1.4 million Invitae shares. The purchase price landed at $83.4 million, of which $63.4 million was goodwill and $23.8 million was assigned to the classifier and the platform behind it.

In August 2022 Invitae announced a strategic realignment. In September it sold One Codex, and the filing does not soften it: “In September 2022, we sold our equity interest in One Codex and certain related assets for an immaterial amount of cash proceeds, as part of our strategic realignment.” Invitae booked a loss of about $19.8 million. Y Combinator names the buyer plainly: the One Codex team, in an employee buyout.

“On behalf of the whole One Codex team, I’m delighted to share that we have spun out One Codex from Invitae as an independent company!” wrote Christopher Smith that October. Greenfield did not return right away. He spent 2022 and 2023 as VP of Product and Tech for biosecurity at Ginkgo Bioworks, another YC S14 company, before coming back.

What is proven, and what is still claimed

Read plainly: the technology is real, independently tested, and has been priced twice by a public company, at $83.4 million on the way in and at approximately nothing on the way out. Those numbers say less about the software than about its fit inside a hereditary cancer testing company that took $2.3 billion of goodwill and IPR&D impairments that year. The benchmarks show a fast, broad classifier that calls more organisms present than its competitors do. That is an advantage in discovery and a liability at a lab director’s desk.

What to watch

• Whether Genome Library, validated on the 3,000 plus assemblies in the ATCC Genome Portal, turns an analysis tool into the system of record for a customer’s strain bank.
• Clinical validation the company publishes itself. The independent benchmarks ran on software versions now years old.
• Whether the 4bases CE-IVD route yields a regulated product in Europe, and whether anything equivalent is attempted in the United States.
• A Form D, or any public signal that the post buyout company has raised outside capital. As of September 2026 there is none.

In their words

Related companies

• Lamar Health: Also in Healthcare IT. A San Francisco startup that builds AI software to automate the referral, benefits verification and prior authorization paperwork behind expensive specialty drugs
• Nest Genomics: Also in Healthcare IT. Software built to carry a genetic test result through a health system, from finding the patient to reminding the family years later
• Candid Health: Also in Healthcare IT. Billing software built to get a medical claim right before it ever leaves the clinic, now sold to more than 200 provider organizations
• Apero Health: Also in Healthcare IT. An API driven practice management platform that folds scheduling, visit notes, insurance billing, patient payments and e-prescriptions into one system for medical

Sources

1. Public recordInvitae Corp Form 10-Q, Q1 2021, business combinations noteSEC EDGAR, CIK 0001501134 · May 4, 2021
2. Public recordInvitae Corp Form 10-K FY2022, disposition of One CodexSEC EDGAR · Feb 28, 2023
3. Public recordFederal awards HDTRA117P0016 and 140G0120P0339 to Reference Genomics, Inc.USAspending.gov · 2017 and 2020
4. Public recordTrademark ONE CODEX, serial 87384650, registration 5778376USPTO · registered Jun 18, 2019
5. Peer reviewedComparison of Three Commercial Tools for Metagenomic Shotgun Sequencing AnalysisThoendel et al., J Clin Microbiol · Feb 24, 2020
6. Peer reviewedBenchmark of thirteen bioinformatic pipelines for metagenomic virus diagnosticsde Vries et al., J Clin Virol · Aug 2021
7. Peer reviewedEvaluation of tools for taxonomic classification of virusesCadenas-Castrejón et al., Brief Funct Genomics · Nov 5, 2022
8. IndependentOne Codex and Microbiology’s Search ProblemBio-IT World, Aaron Krol · Feb 9, 2015
9. IndependentStartup One Codex Building Platform to ID Pathogens from NGS DataGenomeWeb, Tony Fong · Feb 10, 2015
10. CompanyOne Codex is Joining InvitaeOne Codex blog, Nick Greenfield · Feb 19, 2021
11. CompanyA Bright Future for One CodexOne Codex blog, Christopher Smith · Oct 19, 2022
12. PartnerATCC and One Codex Form Collaboration to Address Challenges of Microbiome ResearchPR Newswire · May 31, 2017
13. CompanyOne Codex website and Y Combinator company and founder profileonecodex.com and ycombinator.com · accessed Sep 23, 2026