A real lab, in a real building.
Four stations for plant tissue culture: sterilization, laminar flow, incubation, grow-out. We run micropropagation lines, work on engineered pigment, and photograph every jar.
The lab — wide shot of all four stations — 21:9
Prep and sterilization
Media formulation, decanting, and autoclaving. Everything that enters a jar passes through here first.
Laminar flow
All transfers and subcultures. HEPA-filtered laminar airflow keeps the working field sterile.
Incubation
Temperature and photoperiod control for culture initiation and multiplication.
Grow-out
Acclimatization from jar to soil — the stage where most micropropagation losses happen.
Autoclave — real photo — 4:3
Flow hood, working field — 4:3
Jars in the incubator — 4:3
Cotton that grows its own blue.
Denim is dyed with synthetic indigo, and the dyeing is the dirtiest step in the supply chain. We are working on cotton that makes a blue pigment in the fibre itself — expressing the bacterial bpsA and sfp genes in the Coker 312 line, driven by fibre-specific promoters.
Validation runs in tobacco first, because it is fast and cheap to be wrong in. Freedom-to-operate has been reviewed and the relevant promoter patents appear to have expired.
Coker 312 explants in culture — 4:3
Blue cotton is the R&D.
The blue spider lily is the mission.
Lycoris radiata — the higanbana — blooms scarlet along Japanese graveyards and rice-paddy borders at the autumn equinox, on a bare stem with no leaves. It has been a marker for the boundary between worlds for centuries.
A blue one has never existed. The genus has no pathway to the pigment, so no amount of breeding will produce it. That is exactly why it is worth aiming at: it is a clear, honest, very hard target, and everything we learn getting there is useful somewhere else first.
Built on higanbana folklore, which is public domain and considerably older than anything it has since appeared in.
A labeled image dataset for micropropagation.
Every jar is photographed on a fixed rig at every subculture and once a week between, paired with a record of medium, growth regulators, contamination, and outcome at transfer. Contamination detection and vigour scoring in tissue culture are still done by eye. There is no public dataset for it. We are building one.
Capture rig — jar on the copy stand — 4:3