Genomic Intelligence · Promoter Atlas
Name the cell your therapeutic gene has to work in, and the cell it has to stay quiet in. We search thousands of natural and designed promoters in silico and hand you a short, prioritised set of sequences — ready to synthesise and place in your cassette.
—
on/off pairs in the catalogue today
—
promoters designed by our models
—
human promoters screened and ranked
On request
a pair we do not hold yet is a design project
Why this matters
The therapeutic gene determines what the treated cell can produce. The promoter determines where, when, and how much of the drug is produced. For a successful therapy it has to solve three problems at once:
Promoter choice does not replace capsid engineering, dose optimisation, biodistribution studies, or immune management. But it is one of the central control layers built directly into the therapeutic cassette. FDA guidance explicitly highlights tissue-specific promoters as a way to control expression in target cells and limit activity in non-target tissues, and emphasises evaluating transgene expression in both target and non-target tissues during preclinical development.
In first-generation gene therapy trials for X-linked severe combined immunodeficiency, strong regulatory elements in the retroviral vector activated genes near the vector's integration site. Five of 20 treated children developed T-cell leukaemia between 23 and 68 months after treatment, and one died; in several cases the vector activated the nearby LMO2 oncogene. The field responded with self-inactivating vectors and safer regulatory elements.
In a ten-patient trial of the haemophilia B gene therapy FLT180a, one patient on a high dose reached factor IX activity of 310% of normal after four weeks, required months of anticoagulation, and was hospitalised after an arteriovenous fistula thrombosis. Investigators reduced the dose for later participants. The objective is not maximum expression, but expression within the correct therapeutic range.
Celladon began testing its cardiac gene therapy Mydicar in people by 2009. In 2015 a 250-patient Phase 2b study failed to improve clinical outcomes; later analyses found low or absent vector DNA in cardiac samples and identified insufficient delivery as a leading explanation. Celladon's shares lost almost 80% of their value, development was suspended, and the company ultimately cut roughly 70% of its peak workforce.
These examples are not arguments against gene therapy. They show that expression control is part of the drug: a promoter that is too broad, too strong, or too weak can undermine an otherwise promising programme.
How it works
Promoter development usually starts with a literature search. A team picks a few familiar promoters, clones them into vectors, and tests them in whatever cell models are available. If expression is too weak, too strong, or not specific enough, the team picks another set of sequences and begins again. Every cycle costs design, synthesis, cloning, vector production, cell assays and analysis — and animal studies can still show that a sequence which looked promising in vitro behaves differently in vivo.
Promoter Atlas moves the broad search into software. Our models evaluate thousands of natural and designed promoters before synthesis, so you enter the laboratory with a focused set of candidates already prioritised for the expression profile you need.
7–12 months
3–5 Design-Build-Test-Learn cycles · ~8–10 weeks each · ≈$2–7M in direct R&D
2–3 months
One focused validation cycle · ≈$0.5–1.5M under the same assumptions
These are our current planning assumptions for a typical programme, not a quotation. Reaching a lead cassette four to eight months earlier may preserve tens of millions of dollars in programme value, depending on burn rate and development stage.
You still need a laboratory to validate the final therapeutic cassette. You do not need many laboratory design cycles to discover that a promoter is broadly active, too weak, or active in the wrong cells.
Candidate generation and prioritisation move from repeated experimental cycles to an in silico search completed before synthesis.
Weak, non-specific and otherwise unsuitable candidates are removed before you spend on synthesis, cloning, vector production and animal studies.
Candidates are selected for strong target-cell expression and low off-target activity — the right amount of protein in the right cells.
Promoters can be evaluated against both preclinical and human requirements, so you can prioritise shared sequences or matched pairs early.
What you get
Each Promoter Atlas entry is an actionable promoter candidate, delivered with:
The promoter goes into an AAV transfer plasmid, lentiviral construct, conventional plasmid or another expression vector, next to your therapeutic gene. Running the search yourself would mean a genomic model, a scoring pipeline and a screening operation; this is the output of ours.
The catalogue
Pick the cell type your therapeutic gene must be active in, and the one it must stay quiet in. Pairs we already hold are shown straight away. Anything else, we design.
★ marks a pair already in the catalogue.
