Biodesign and biomaterials: a plain-language glossary

Letters with entries A B C D E G H L M P R S T
A

AI-led

Work with living matter where AI models are used everywhere and their outputs are taken at face value. The score becomes the verdict, the best-looking option becomes the choice, and the screen stands in for the living matter.

Often mistaken for AI-native, because the models are used throughout. What is missing is the reading: what the output cannot show, and who and what a change touches.

AI-native biodesign

Design work with living matter where AI models are part of the work throughout a project, their outputs are read for what they cannot show, and each change is weighed against what it would mean for the organisms and people involved.

Often taken to mean protein design. Proteins are one place it happens, not the whole of it.

AI-native biodesigner

A designer of living matter who works with AI models throughout a project, reads what their outputs cannot show, and asks what a change would mean for the organisms and people involved.

“Throughout” means present, not leading. The work often begins with the living matter itself. Using AI at every step is not enough: without that reading, the work is AI-led.

Not to be confused with “AI-native biotech”, a term biotech news and investment writing began using in 2026 for drug companies built around AI. That term describes a company. This one describes how a person works with living matter.

B

Bacterial cellulose

A mat of pure cellulose fibres built outside the cell by bacteria, usually Komagataeibacter, at the surface of a sugary liquid. It is the material behind kombucha leather.

Most claims about its strength describe the wet mat. Almost all of them are tested after drying.

Biocultural labels

Small digital tags, made through Local Contexts, that let Indigenous communities attach their own terms to genetic resources and the data made from them: where it came from, whether consent was given, and how it may be used.

Only communities apply them. Researchers and institutions have a separate tool, Biocultural Notices, to flag that Indigenous rights may apply to data they hold.

Biodesign

Design that works with living systems and the materials they grow. The designer sets up and tends the conditions in which an organism makes the material, rather than shaping one that arrives finished.

The word is shared with medical-device biodesign, taught at Stanford. Both apply design method to biology, and there the resemblance stops: different materials, different literatures, colliding search results.

Read: biodesign → Updated 31 Aug 2026

Biodesign AI footprint

The energy and water used, and the heat given off, by the data centres that run the AI tools in a biodesign project. Most of it happens far from the studio, in places the designer never sees.

An AI footprint is often taken to mean water per prompt. But a footprint also includes energy and heat, and it depends on where the work happens, which a per-prompt average cannot show.

Biodesign heat reuse

Using the low-temperature waste heat from data centres to keep living matter warm while it grows, or to dry what it makes. Examples include fungal mycelium, bacterial cellulose, and algae and cyanobacteria.

Heat reuse is often taken to mean heating homes. Growing living matter needs low, steady heat all year, including in summer, when homes need little.

Biodesign promise

The sentence about what a living material does that a design project has come to rely on. “This mycelium heals its own cracks.” It names an effect, and it usually arrives unchosen, picked up from a paper or a supervisor.

One word can hide five unrelated behaviours. “Self-heals” is the usual example.

Read: the biodesign promise → Updated 16 Jun 2026

Biofabrication

Letting an organism grow the material in conditions you tend, rather than shaping a material that is already made.

The word covers everything from a kombucha mat to printed human tissue. Wikipedia’s definition is automated and medical. Dezeen’s is anything grown.

Read: biofabrication → Updated 04 Aug 2026
C

Citrate (citric acid)

Citrate is the dissolved form of citric acid in a bacterial cellulose recipe. Recipes say citric acid, papers say citrate, and they are not the same molecule. Citrate is the one that bacteria meet.

Called a buffer in protocols since the 1980s, usually without a test.

Read: citrate (citric acid) → Updated 28 Aug 2026

Control interference

When something added to keep one part of an experiment steady also changes another biologically important part of the system. The control still does its job. That is what makes it easy to miss.

Changing the variable you are testing can change the control’s chemistry along with it. Holding a buffer’s concentration constant is not the same as holding its effects constant.

Read: control interference → Updated 01 Sep 2026
D

Dark proteome

The part of the protein world we know almost nothing about. The term first meant proteins with no known shape. Since AlphaFold, it’s used more for proteins with no known function.

A predicted shape can make a dark protein look understood. The shape says little about what the protein does, or what it touches.

Data provenance

The record of where data came from, who made it, and what was done to it. For a protein, it tells you whether a shape was measured, inferred from a relative, or predicted.

Data teams in medicine and drug discovery list provenance as one test of “AI-ready data”. Their concern is mostly whether a model can be trusted. Designers can also ask who and what the data came from.

Digital sequence information (DSI)

The genetic sequences of living things held in databases: DNA and RNA, and in many readings protein sequences too. It’s the term governments use when they argue about who should benefit when that data makes money.

There’s no agreed legal definition, which is part of the problem. If you have looked up a protein on UniProt or AlphaFold, you have used it.

E

Exposure legacy

A change caused by previous conditions that persists in a living material and can influence its response to later conditions. Two runs can share the same final recipe and begin with organisms prepared differently.

The tests that establish how a material behaves can become part of its preparation. Calibration can change the material being calibrated.

Read: exposure legacy → Updated 09 Sep 2026
G

Generic environmental claim

A broad green word about a product with nothing specific attached, such as eco-friendly, green or sustainable. Since 27 September 2026 these are banned in consumer advertising across the EU, unless the claim is specified in the same place or recognised proof of excellent performance exists.

For a biodesigner, it means the vague version of a material claim is now the risky one.

Growth medium

The liquid a culture is grown in: a sugar source, a nitrogen source, water, and whatever else the recipe inherited.

Updated 11 Aug 2026

Gypsum

Calcium sulfate dihydrate, added to mushroom substrate and mycelium recipes. It keeps grain from clumping, supplies calcium and sulfur, and is very often described as a pH buffer.

A buffer resists a change in pH. What was actually measured is gypsum lowering pH in one kind of compost, which is a different thing. The chemistry offers no buffering capacity either.

Read: gypsum → Updated 01 Sep 2026
H

HEPES

The buffer in many cyanobacteria and algae recipes. It is added to keep the pH from drifting as the cells feed and change the water around them. Under light, and with help from the medium, HEPES also makes hydrogen peroxide.

Whether HEPES matters depends on who is there to clear the peroxide. A crowded culture can. A thinly seeded one cannot, unless a neighbouring organism steps in.

Read: HEPES → Updated 24 Aug 2026
L

Living textile

A fabric whose cells are still alive, or can be woken up, after the sheet is made. Most grown textiles are dried and dead.

Keeping cells alive is what allows surface renewal or repair, and it is also what makes the material fussy. For a designer it moves the work from finishing a material to keeping one.

M

Mechanism accretion

A causal explanation in a protocol, growing more specific with each retelling while the evidence behind it stays the same. Nothing is invented. The explanation just thickens.

Walk one inherited sentence back through its sources. If every step added a sentence and none added a measurement, that is what you are looking at.

Read: mechanism accretion → Updated 01 Sep 2026

Molecular design literacy

The ability to read a claim about a living material and see the biology it depends on: what is doing the work, what sets it off, and what evidence would show it is real. It is a reading skill before it is a making skill.

Often taken to mean lab skill. It is closer to reading a label than to running a protocol.

P

Property capture

A behaviour that several conditions produced together, written down as a property of one ingredient or organism. Move it somewhere else and the behaviour does not come with it.

The short version fits a protocol table, a database field and a model variable. The version with the conditions attached fits none of them. That is why the short one travels.

Read: property capture → Updated 01 Sep 2026
R

Role overshadowing

A correct account of what an ingredient does in a living system, taken for the complete account. The label is right, which is what makes a swap look safer than it is.

Nothing here is false, so a fact check finds nothing wrong. A correct label still narrows what anyone thinks to test, and makes a swap look safer than it is.

Read: role overshadowing → Updated 29 Aug 2026
S

Structural colour

Structural colour is the colour produced by the physical arrangement of a material rather than by a pigment or dye. The arrangement reflects, redirects and scatters light so that some wavelengths reach the eye more strongly than others.

Usually offered as colour without dye, and so as the sustainable choice. Nothing here is coloured though, so the colour lasts only as long as the arrangement holds.

Read: structural colour → Updated 24 Aug 2026
T

Text-guided protein design

Designing a protein by describing in plain words what it should do, and letting a model turn that description into a sequence.

The description is not a specification, and the model’s shortlist is mostly duds. A fluent prompt and a confident output still leave the wet lab doing the deciding.