BRENNEN

Industries / Health / Pharmaceuticals

Biotech

Biotech uses living cells and biological molecules, such as DNA, enzymes and antibodies, to make medicines, tests and materials.

1973

Stanley Cohen, Annie Chang, Herbert Boyer and Robert Helling joined pieces of DNA from two different plasmids, small rings of bacterial DNA, in a test tube, and showed that the joined DNA worked when put back into the bacterium E. coli. They published in PNAS in November 1973.

pmc.ncbi.nlm.nih.govsciencehistory.org

42

Milestones achieved across 6 domains.

16

To watch for: announced, not yet confirmed outside.

15

Open deliverables. Nobody has shipped them.

In one minute

Where it started

In 1973 Stanley Cohen and Herbert Boyer showed that a gene could be cut from one DNA molecule, joined to another and copied inside bacteria. Boyer and the investor Robert Swanson founded Genentech in 1976 to make human proteins this way. Human insulin made by bacteria, sold by Eli Lilly as Humulin, was approved by the US Food and Drug Administration (FDA) in 1982; the Science History Institute calls it the first biotechnology product to reach the market.

Where it stands

Reading DNA has become routine: the US National Human Genome Research Institute (NHGRI) estimates that a draft human genome cost about USD 14 million in 2006 and under USD 1,500 by late 2015. Proteins made in hamster cells now reach several grams per litre of culture, copies of cell-made drugs are approved in Europe and the US, and the FDA lists 55 approved cell and gene therapy products, among them the first medicine made with the gene-editing tool CRISPR. The field has also had public failures, from a gene therapy trial death in 1999 to the Theranos blood-testing fraud and deaths that narrowed the use of a muscle-disease gene therapy in 2025.

What is open

Much of what the industry knows sits in PDFs, private files and lab notebooks: how accurate each approved test is, which gene therapies were approved where and at what price, how bioreactor runs turned out, and which experiments have been repeated. The open jobs below turn that material into public tables, test sets and leaderboards. None of them needs a lab. Each can be started this month with public data, open-source software and current AI models that read documents and write code.

Domains

AchievedWatch forOpen

Open in biotech

3 of 15 open deliverables

Open · Reading and writing DNA

A rolling test for DNA models

A public leaderboard for DNA language models, AI models trained on raw DNA letters, such as Arc Institute's Evo 2, GPN-MSA and Nucleotide Transformer. Each model is scored only on variants from ClinVar, the US public archive of DNA changes labelled harmful or harmless, whose labels first appeared after that model's training data was collected. It is rebuilt with each new ClinVar release.

Built with what exists

Monthly ClinVar releases from NCBI; open model weights (Evo 2 is on Hugging Face in sizes from 1 billion to 40 billion parameters, with up to 1 million letters of context); VEP-eval's scoring notebooks; one rented cloud GPU; Python. Open DNA models with million-letter context are the advance that lets a small team score many models across the genome.

Credit · unclaimed

Open · Protein design

One rule for counting a hit

A public script and table that recomputes the success rate of every open binder-design competition under a single written rule, saying what counts as made, what counts as a binder and what counts as designed from scratch, so that numbers from different events can be compared.

Built with what exists

The public ProteinBase collections; the Adaptyv EGFR competition data on GitHub; the Anthropic dataset on Hugging Face (CC BY 4.0 data, MIT-licensed scripts); Python and pandas.

Credit · unclaimed

Open · Cell and gene therapy

A register of approved therapies

A free, downloadable table with one row per cell or gene therapy product. Each row gives the approval date from each regulator (FDA, EMA, MHRA, Japan, China, South Korea), the published list price where there is one, withdrawals and expiries, and safety actions such as boxed warnings or narrowed uses. Every field links to its source.

Built with what exists

Python (requests, BeautifulSoup) to collect FDA and EMA product pages; pdftotext for approval letters; a CSV file in a public GitHub repository with a change log.

Credit · unclaimed

Where it sits