HTGAA 2026 · Week 14 · May 5–13, 2026
Final Week

Bio Design & Bio Fabrication

The semester closing as an opening: completing the computational models, finalizing the DNA designs, vectorizing the figures, and presenting Füzi Poiesis to the Committed Listeners of the SynBio USFQ Node.

Suzanne Lee Christina Agapakis Presentation · May 13 Füzi Poiesis · Aim 1 Complete
Lecture · May 5, 2026

Bio Design & Bio Fabrication

The final week of HTGAA 2026 was curated by Suzanne Lee and Christina Agapakis — two figures who embody different traditions of the same question: what does it mean to design with living biology? Lee from material biofabrication (Biocouture, mycelium, kombucha leather), Agapakis from the interface between science, art, and scientific writing. The week functioned as a conceptual frame for what each student had built over the semester: not mastery of techniques, but the ability to articulate a project of one's own with rigor and with meaning.

Suzanne Lee — Biofabrication
Designing with organisms, not just for them
Lee's work on biologically grown materials — bacterial leather, structural mycelia, fermented fibers — illustrates the same principle guiding Füzi Poiesis: the organism is not a passive reactor but a design agent. In Biocouture, Komagataeibacter xylinus produces bacterial cellulose that the designer shapes during the growth process, not after. In Füzi Poiesis, the topological architecture of the auxotrophic hypercycle converts metabolic cooperation into a containment mechanism — the consortium is not contained by an external system but by its own population dynamics. Both projects treat biology as a design material with its own agency.
Christina Agapakis — Bio Art & Communication
Writing as an epistemic instrument
Agapakis has consistently articulated that the separation between science and the humanities is an institutional construction, not a cognitive distinction. Her work on bioperfumes (Gingko Bioworks), on fermentation as cultural practice, and on art-science as legitimate research resonated directly with the ethical dimension of Füzi Poiesis: the Mapuche-Lafkenche principle of itrofill mogen — the right of all life to perpetuate its cycle — is not rhetorical decoration on the project, but the criterion that subordinates any deployment decision to community sovereignty. Agapakis offers a framework for understanding why this matters epistemologically, not just ethically.
Connection to Füzi Poiesis

The week's thesis — that bio-design and bio-fabrication are not stages that follow scientific discovery but independent forms of knowledge generation — describes exactly what Aim 1 of Füzi Poiesis attempts to demonstrate. The mathematical model is not a post-hoc validation of the design: it is the design. The three-strain ring topology is not proposed because it is known to work — it is proposed because the equations demonstrate that its containment properties are qualitatively distinct from any two-strain system. The computational design is the experiment.

Final Sprint · May 5–12, 2026

Completing Füzi Poiesis — The Last Hours

Week 14 was not only a week of reading and reflection — it was the week that closed the cycle of work begun in January. In the hours before the May 13 presentation, following a late-night review session with Benjamín Arias (tutor of the SynBio USFQ Node), the final substantive revisions to the presentation were made. What follows is the log of that work.

Night of May 11 · Tutor review session
Feedback from Benjamín Arias on the presentation
The late-night review session identified four concrete problems with the slides: (1) the computational figures were rendered as low-resolution raster PNGs and pixelated when projected; (2) the final slide lacked a figure to visually anchor the narrative of progression toward community-led restoration; (3) the second slide had excessive visual density that compromised readability; (4) the cover slide included neither a photograph of the lake nor its map, visually disconnecting the biological system from its real geographic and ecological context.
Early morning May 12 · ~2 hours of work
Vectorization of computational figures
The four ODE model panels (monoculture, closed pair, three-strain hypercycle, cascade extinction) and the structural containment figure (exponential decay + eigenvalue spectrum) were remastered as scalable SVG vectors. This involved regenerating the plots from the JupyterLab notebook with an SVG backend instead of PNG, and adjusting palette, typography, and proportions to maintain coherence with the dark visual scheme of the presentation. The figures are now sharp at any projection resolution.
Early morning May 12 · Cover slide
Integration of lake photograph and basin map
A photograph of Lake Budi from the southern shore (the littoral zone where fecal contamination has been documented) and the basin map with Quesille (2022) sampling points were added to the cover slide. The challenge was integrating both images without visually overloading the slide or losing coherence with the central narrative structure: presenting the problem (stratification-driven eutrophication) and the proposed solution (the auxotrophic hypercycle). The solution was to position both images as contextual elements in the lower-left margin, freeing the central space for the consortium diagram.
Early morning May 12 · Slide 2
Restructuring the computational slide
The second slide was redesigned structurally to improve legibility. A three-column grid was established: left column for the ODE panels (topological evidence), center column for the transfer functions and truth table (circuit evidence), right column for the plasmid maps (DNA design evidence). The key quantitative results — n* ≈ 0.629, Re(λ_max) ≈ −0.215, t½ = 6.9 h — were made visible without requiring dense text reading.
Early morning May 12 · Slide 3
Progression figure: from computation to community-led restoration
The final slide lacked a figure to make the progression from the computational model to field deployment under Lafkenche governance visually legible. The existing illustration was painted over so that the progression lines — E. coli K-12 → Halomonas elongata → native Budi strains → Bokashi del Budi in the littoral zone — were no longer abstract lines but elements representing a concrete and readable trajectory. The result was a three-level figure with biological and geographic iconography that narratively closes the arc of the project.
Deliverables · Aim 1 Complete

Füzi Poiesis — Computational Proof of Concept

At the close of Week 14, Aim 1 of Füzi Poiesis is complete. The four computational and DNA design deliverables have been produced, internally validated, and documented with sufficient detail for independent reproduction.

Complete plasmid
pFP-C · 4,238 bp
AND-gate PhoA · ts-ori · KanR · ΔleuB coupling · R-M shielded · Benchling Academic
Functional cassettes
pFP-A + pFP-B
4,626 bp (mcjABCD anti-coliform) · 3,677 bp (sqr-pdo anti-H₂S) · GenBank exported
ODE model · interior fixed point
n* ≈ 0.629
s* ≈ 0.628 · Re(λ_max) ≈ −0.215 · all 6 eigenvalues Re<0 · Lyapunov stable
Structural containment
t½ = 6.9 h
Cascade extinction of all 3 strains vs. single loss in the closed pair · escape estimated ~decades
AND-gate · Boolean truth table
Validation of the Strain C logic circuit
The LuxR/AHL transfer function (Hill, K_d = 10 nM, n = 2) produces a digital-like sigmoidal response for signal X₁. The soluble reactive phosphorus sensor (SRP ≥ 0.5 mg/L — the eutrophication threshold documented for Lake Budi by Quesille 2022) acts as X₂. PhoA output is <5% of maximum in the three OFF states — (0,0), (1,0), and (0,1) — and >90% of maximum in the ON state (1,1). The circuit responds only when the consortium is present and the lake is eutrophied: the two simultaneous conditions that define a real environmental emergency.
Technical note · Topology vs. kinetics

The containment property of the three-strain ring does not improve the kinetic escape time relative to a two-strain system — the order of magnitude is the same (~decades with N=10⁸, μ_mut=10⁻⁸, f_escape=10⁻⁴). The advantage is topological: in a closed pair, a mutant that restores its auxotrophy can persist parasitically while its partner continues supplying metabolite. In the ring of three, restoring independence from one partner still leaves the mutant without the metabolite of the third strain, whose disappearance is a direct consequence of the cycle breaking. The ring fails closed; the pair fails open. This is the structural distinction the project demonstrates.

Presentation · May 13, 2026

Committed Listeners — SynBio USFQ Node

On May 13, 2026, Füzi Poiesis was presented to the Committed Listeners of the SynBio USFQ Node. The presentation consisted of three slides, following the logic of the course: problem → computational evidence → progression toward the real world. Total duration was approximately five minutes, followed by audience questions.

Slide 1 — Cover
The lake, the problem, the consortium architecture
The cover integrated the project name (Füzi Poiesis: Topological Biocontainment for the Bioremediation of Lake Budi), the photograph of the lake from the southern shore, the basin map, and the central consortium diagram: the three-strain auxotrophic hypercycle layered over the lake's stratification — oxic surface zone, halocline, anoxic zone with H₂S production, sediment with PO₄³⁻ release. The diagram simultaneously presented the ecological problem and the proposed solution: each layer of the lake corresponds to one strain's remediation function.
Slide 2 — Aim 1
Computational proof of concept
The second slide presented the ODE model results (four SVG vector panels), the AND-gate transfer functions and truth table, and the three plasmid maps (pFP-A, pFP-B, pFP-C). The three-column restructuring allowed a left-to-right reading: topological containment evidence → circuit logic evidence → DNA design evidence. The key quantitative results — n* ≈ 0.629, Re(λ_max) ≈ −0.215, t½ = 6.9 h — were legible without requiring dense text reading.
Slide 3 — Aims 2 & 3
From computation to community-led restoration
The third slide presented the three-level progression: Level 1 (E. coli K-12, computational proof, Aim 1 complete), Level 2 (Halomonas elongata + native Budi strains, bioprospecting + wet-lab validation, postgraduate), Level 3 (native chassis consortium in Bokashi del Budi matrix deployed in the littoral zone under Lafkenche governance, Aim 3, contingent on FPIC and regulatory approval). The progression figure designed in the preceding hours anchored this trajectory visually.
Good! They look great, Nicolás. Excellent work. They're very beautiful. Go ahead and upload them.
Benjamín Arias Almeida · Tutor, SynBio USFQ Node · May 12, 2026, 10:11 AM
Reflection · Validation as closure

The phrase "go ahead and upload them" — the final permission from a tutor who accompanied the project from the ODE models to the lake map — is not trivial in the context of a course where scientific validation and community validation are equally serious. Füzi Poiesis does not yet have permission from the Lafkenche community to enter the lake. It has, for now, its tutor's permission to upload the slides. Those two validations operate at entirely different scales — and the project recognizes them as such.

Aim 1 is complete. The lake is still waiting. I have given it everything I had.

End of semester

What it means to design with biology

Week 14 of HTGAA does not close just a semester — it closes an argument that began in January with the question of what is possible with synthetic biology when you have access to the tools. Füzi Poiesis's answer is a specific proposal: design the mathematical conditions under which biogeochemical resilience can be encoded into the evolutionary architecture of a synthetic consortium.

Lesson of the semester
Topology as a design argument

The methodological contribution of Füzi Poiesis is not in any individual tool — Benchling, Python, SciPy, the RBS Calculator — but in using Lyapunov stability theory to establish a topological distinction as a rigorous mathematical result, not as an intuition. That is bio-design in the sense that Agapakis and Lee proposed this week: not applying biology to a problem, but letting the structure of the problem reveal which biological architecture belongs to it.

Lake Budi has three documented pathologies: fecal coliform contamination, hydrogen sulfide accumulation, and internal phosphorus eutrophication. The consortium has three strains, each targeting one pathology, coupled in a dependency cycle that replicates the cooperative structure of a healthy ecosystem. The solution is not proposed because it is known to work — it is proposed because the equations demonstrate that its properties are qualitatively distinct from any two-strain alternative. The design is the argument.

What remains
The honest limitations of Aim 1

The model assumes a well-mixed environment and kinetic parameters from E. coli K-12 under amino acid restriction — a laboratory strain under controlled conditions, not a halotolerant strain in brackish water at 5–15 g/L NaCl with benthic anoxia. If the model is recalibrated with Halomonas elongata or native Budi strain parameters in Aim 2 and the interior fixed point loses stability, the three-strain ring requires redesign. The benthic metagenome of Lake Budi does not exist in the published literature: horizontal gene transfer risk cannot be quantified without it. That too is a result of Aim 1 — knowing precisely what is not yet known.

Full project documentation

Füzi Poiesis — Final Project

Genetic design, mathematical modeling, Lyapunov stability analysis, and the ethical framework for deployment under Lafkenche governance. All of Aim 1 documented.

View final project →