Industries / Frontier science / Synthetic biology
Biomanufacturing
This work takes an engineered organism or enzyme and makes product with it at working scale, in fermentation tanks or in reactors with no cells at all, then separates the product from everything else in the broth.
6
Milestones achieved, from 1915 to 2018.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1915
Acetone brewed for cordite
Cordite, the propellant for naval guns, needed acetone, and wood distillation could not supply enough. Chaim Weizmann isolated the bacterium now called Clostridium acetobutylicum, which turns maize starch into acetone and butyl alcohol, and the British government gave him a pilot plant at Nicholson's Three Mills gin distillery in Bow, London, in 1915. The process then ran at a Scottish distillery, a works at King's Lynn and a new Admiralty plant in Dorset.
Lessona gin distillery was never built to keep other microbes out, and contamination dogged the Bow site, in what the magazine The Microbiologist calls the first microbiological process to require asepsis on an industrial scale.
Credit
Chaim Weizmann; his patent describes the fermentation
1988
A cloned enzyme in detergent
Novo Industri put Lipolase on the market, a lipase, an enzyme that breaks down fats, for use in laundry detergent. Novonesis, the company's successor, dates to 1988 the launch of what it calls the world's first fat-breaking detergent enzyme made by microbial fermentation. A Unilever patent filed in May 1988 describes the same product as the lipase of the fungus Humicola lanuginosa, cloned by recombinant DNA technique and expressed in the mould Aspergillus oryzae, and names Lipolase as the version sold by Novo. Nature Biotechnology reported in April 1988 that two European companies were marketing lipases.
Credit
Novo Industri A/S
1989
Impurities in a fermented supplement
An outbreak of eosinophilia-myalgia syndrome, a painful blood and muscle disorder, struck the United States in 1989. Of six makers of the supplement L-tryptophan, only the product of Showa Denko K.K. was clearly associated with illness, exposure was almost universal among cases and rare among controls, and cases fell abruptly once the products were recalled. Contaminants in that product were later identified as compounds formed when tryptophan joined fatty acids from the bacterial cell membrane.
Lessonin a fermented product that people swallow, the traces carried over from the organism decide safety, so a change in purification is a change in the product.
Credit
Showa Denko K.K. made the implicated product; the link was set out by Edwin Kilbourne and colleagues at the US Centers for Disease Control and Prevention, and the contaminant structures by Klarskov and colleagues
2010
An enzyme replaces rhodium
The diabetes drug sitagliptin was made with a step that used a rhodium catalyst under hydrogen pressure. A team replaced that step with an engineered transaminase, an enzyme that moves an amine group onto a carbon skeleton. They began from an enzyme that had the right chemistry but no activity at all on the drug's intermediate, then changed its substrate step by step and evolved it until it worked in manufacturing.
Lessonthe enzyme found in nature was the wrong starting point, and the work was in the rounds of redesign that followed.
Credit
Christopher Savile and colleagues at Codexis and Merck, in Science
2016
A plant for bio-based butanediol
Novamont opened its Mater-Biotech plant at Bottrighe, in north-east Italy, on 29 September 2016 to make 30,000 tons a year of 1,4-butanediol, a chemical used in plastics, by fermenting sugar with Genomatica's GENO BDO process instead of making it from fossil feedstock. The capacity was set 70% above the original target of 18,000 tons, and Novamont put the investment at EUR 100 million.
Credit
Novamont and Genomatica
2018
Ethanol from steel-mill gas
A joint venture of LanzaTech and the Chinese steelmaker Shougang Group started a plant at the Jingtang steel mill in Caofeidian, Hebei, on 3 May 2018, which ferments the mill's waste gases with anaerobic bacteria to make 46,000 tons of fuel ethanol a year. The ethanol met the American blending standard ASTM D4806 and China's GB 18350. The same organism, the engineered acetogen Clostridium autoethanogenum, was later reported in Nature Biotechnology making acetone and isopropanol from waste gas at industrial pilot scale, at up to about 3 grams per litre per hour and about 90% selectivity, with a life-cycle analysis that found a negative carbon footprint.
Credit
LanzaTech and Shougang Group; the pilot work by Fungmin Liew, Michael Köpke and colleagues
Watch for
Five chemical steps become one
WuXi AppTec says its biocatalysis team replaced a five-step chemical route to a drug intermediate with a single enzymatic step, raising the yield from about 40% to 76%, and then, after two rounds of directed evolution in a month, simplified the system to one enzyme and one coenzyme with a yield of 90% and a total cost more than 70% below the chemical route. The account names neither the client, the molecule nor the scale.
Achieved whenthe route and the figures appear in a paper or a regulatory filing, or the client confirms them.
Claimed by
WuXi AppTec, May 2026
A tougher biodegradable plastic
A group reinforced the gene for a lactate-polymerising enzyme in the bacterium Cupriavidus necator and reports 68 grams per litre of a high-molecular-weight lactate-based polyester within 48 hours, which it calls the highest titre reported for that polymer, with about 20 MPa tensile strength, 190% elongation at break and over 75% breakdown in natural seawater in five weeks. The work was done at laboratory and bench scale.
Achieved whenanother group reaches the titre, or Kaneka reports the polymer made at tonne scale.
Claimed by
Seiichi Taguchi and colleagues at Shinshu University with Kaneka Corporation and Japan's National Institute of Advanced Industrial Science and Technology, online December 2025 and in print on 1 April 2026
Open
01
A register of enzyme preparations
A public, machine-readable register of the enzyme preparations sold for industrial use, one row per product, giving the supplier, the enzyme, the organism it is made in, whether that organism is engineered, the legal basis for its use, the date and a source link. Europe's food enzymes are the model; the gap to fill is everything outside that, from United States food use to detergents, textiles, pulp, animal feed and drug manufacture.
The briefCredit · unclaimed02
A table of biocatalysis cases
One public table of the industrial cases where an engineered enzyme replaced a chemical step, with the fields that decide whether a claim means anything: the product, the step replaced, how many steps before and after, the yield before and after, the scale the new route ran at, and whether the numbers appear in a peer-reviewed paper or only in the company's own account.
The briefCredit · unclaimed
Open 01 · the brief
A register of enzyme preparations
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public, machine-readable register of the enzyme preparations sold for industrial use, one row per product, giving the supplier, the enzyme, the organism it is made in, whether that organism is engineered, the legal basis for its use, the date and a source link. Europe's food enzymes are the model; the gap to fill is everything outside that, from United States food use to detergents, textiles, pulp, animal feed and drug manufacture.
Why it is still open
Europe has such a register. Sciensano's Food Enzyme Database lists the food enzyme preparations on the European market and lets anyone filter them by protein name, source, intended use, production method and whether the strain is genetically modified. The main United States record, the Food and Drug Administration's GRAS Notice Inventory, is organised by notice: it gives the substance, the file number, the notifier and the dates, and the production organism sits inside each notice's own documents rather than in a field anyone can sort. Outside food, the nearest thing is LICEDB, published in 2025 and described by its authors as the first database for the core enzymes of light industry, which gathers enzyme properties and sequences for feed, papermaking, textiles, detergents, leather and sugar rather than the products sold and the strains they are made in.
Built with what exists
The GRAS Notice Inventory and the notice documents behind it; suppliers' published product and specification sheets; a long-context AI model to read each notice and pull the production organism, with a person checking a sample; Python; a CSV in a public Git repository with a stated schema.
First step
Download the GRAS Notice Inventory, filter it to enzyme preparations, and code the production organism for the first 50 notices by hand to settle the columns.
Done looks like
At least 300 enzyme preparations in one CSV, each with supplier, enzyme, production organism, whether it is engineered, legal basis, date and link; the tagging rules published; and a hand check of 100 random rows finding at least 95% of production organisms correct.
Open 02 · the brief
A table of biocatalysis cases
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
One public table of the industrial cases where an engineered enzyme replaced a chemical step, with the fields that decide whether a claim means anything: the product, the step replaced, how many steps before and after, the yield before and after, the scale the new route ran at, and whether the numbers appear in a peer-reviewed paper or only in the company's own account.
Why it is still open
The best-documented case is now 16 years old, the sitagliptin transaminase published in Science in 2010. Newer cases circulate as company accounts: WuXi AppTec's own post describes a five-step route cut to one step with the yield rising from about 40% to 90% and the cost falling by more than 70%, while naming neither the client, the molecule nor the scale. No public list sets such cases side by side with what each one actually discloses. The field's own curation effort stops short of this: the RetroBioCat Database, published in 2023, collects biocatalytic reactions with synthetic substrates because they are missing from enzyme databases such as BRENDA and KEGG, and it is organised by reaction for picking an enzyme, not by industrial process with the steps replaced, the yield before and after, and the scale the route ran at.
Built with what exists
Europe PMC and publishers' process chemistry journals for the published cases; company posts, vendor case studies and conference abstracts for the rest; an AI model to pull the fields from each document, with a person checking every row; a CSV in a public repository.
First step
Enter 20 cases by hand, starting with the sitagliptin process and the WuXi account, and fix the column definitions before going further.
Done looks like
At least 100 cases in one file, each with the fields above and a source link, a written rule for what counts as an industrial case, at least 30 rows carrying a peer-reviewed source, and a published count of how many cases state the scale they ran at.