Week 1 HW: Final Project Ideas

Hannah Drawns, Williamsburg, USA - William & Mary Student HTBAA Homework #1

1. What I would like to biomanufacture, and why The big picture: I want to biomanufacture a persistent, self-renewing engineered symbiont system that manages the body’s chronic inflammatory and toxin load automatically, in the background, based on the body’s own internal signals. Instead of a pill you remember to take, it’ll be a living regulatory layer that’s just part of you. Concretely, this means the system doesn’t discriminate based on the source of an inflammatory or toxic input. If I drink water with trace heavy metals in it, or I’m exposed to something in my food or environment, or I simply make a choice that would normally cost my body something, the system detects and neutralizes it before it can do damage, then clears it with no residual trace — functionally, it’s as if the exposure never happened. I don’t have to identify the threat, time a response, or change my behavior to avoid it; the baseline layer is just handling it continuously, in the background.

On top of that automatic baseline, I want a toggle layer: something a person takes deliberately to override specific parts of that system when they choose to (not because the body signaled a need), so health becomes a foundation you build freely on top of rather than a constraint you’re constantly managing. This way the consumer can customize the effects if there are aspects an individual does not want without the item itself needing to be individually adjusted. This is directly inspired by a combination of myself and majority of the people close to me having chronic inflammation issues as well as a personal interest in non-traditional circuits and awe for a wide variety of fictional medicine meets magic and computation media, especially depictions of people whose physiology always heals no matter the issue using mechanisms that are slightly adjusted or recombined versions of real human biology (The Flash’s metabolism immediately detoxifying alcohol, Solo Leveling’s toggleable longevity and detoxification skills) because they describe a coherent target state (autonomous baseline health + consented override).

The reason this matters: chronic low-grade inflammation is a documented, mechanistic driver of most of the world’s leading causes of death (cardiovascular disease, cancer, diabetes, chronic respiratory disease), which collectively cause ~74% of global deaths (41 million people/year, per WHO). It also actively interferes with the body’s ability to heal and function well day to day, independent of any specific diagnosis. A platform that quiets chronic inflammation without requiring constant behavioral vigilance would be broadly useful — for people who manage chronic illness, struggle maintaining habits, can’t access medicine continuously, who simply want more freedom of choice around their health, and indirect improvements flowing downstream as the population becomes healthier (lower healthcare burden, higher functional capacity).

The buildable slice (my actual proof-of-concept target): a single, well-characterized mechanism inside that bigger system — an engineered probiotic that reduces metabolic endotoxemia, a documented process where lipopolysaccharide (LPS/endotoxin) from gut bacteria leaks across a permeable gut barrier into the bloodstream and triggers systemic inflammation via the TLR4 receptor. My organism does two things: (1) secretes intestinal alkaline phosphatase (IAP), an enzyme that dephosphorylates LPS’s toxic lipid-A component, reducing its inflammatory potency roughly 100-fold; and (2) produces butyrate, a short-chain fatty acid that reinforces the gut barrier itself, reducing how much LPS crosses over in the first place. This is deliberately narrower than the full vision, but it’s the first real brick: it’s a specific, catalytic mechanism (not indiscriminate binding, so it doesn’t interfere with nutrient or medication absorption the way charcoal/clay binders do), delivered as a capsule, requiring no behavior change from the user.

2. What living system I would use Chassis: Escherichia coli Nissle 1917 (EcN) — a non-pathogenic probiotic E. coli strain with decades of documented human use, a well-developed genetic engineering toolkit (plasmid and CRISPR-based), and existing precedent as a chassis for engineered live biotherapeutics currently in clinical trials for other conditions.

What goes in: a carbon source (glucose or another simple sugar) plus standard fermentation nutrients. What happens: engineered EcN carries two added genetic modules — one encoding intestinal alkaline phosphatase for secretion, one encoding the butyrate-production pathway (already demonstrated in engineered E. coli and related organisms in the literature). As the bacteria grow and colonize the gut, they continuously produce and release both outputs locally, at the site where LPS is generated and where the barrier needs reinforcing.

What comes out (at the manufacturing stage): bacterial cells grown to high density, harvested, and freeze-dried (lyophilized) into a stable, encapsulated powder for a standard oral capsule — shelf-stable, no refrigeration dependency, travel-friendly.

Assumption I’m flagging: that both functional modules (IAP secretion + butyrate production) can be stably co-expressed in one strain without excessive metabolic burden reducing growth/yield. Precedent exists for each individually; combining them is the part that would need real lab validation.

3. How I would use automation Strain development (design-build-test-learn cycle): liquid-handling robots (e.g., Opentrons-class platforms) to build and screen a library of promoter/gene-expression variants in parallel, rather than manually testing one construct at a time — dramatically cuts the time to find a strain with good yield of both IAP and butyrate without excessive growth penalty.

Production-scale fermentation: automated fed-batch bioreactor control — feed rate, pH, dissolved oxygen, and temperature managed by control software rather than manual monitoring, which is standard in industrial E. coli fermentation and lets one facility run many batches with consistent output.

Downstream processing: automated harvesting/centrifugation and freeze-drying (lyophilization) lines to convert liquid culture into stable capsule-ready powder at scale.

For the bigger-vision version of this project, automation would also need to include biosensor-based dynamic control circuits — genetic switches that turn specific outputs on/off in response to internal signals (LPS level, pH, inflammatory markers) rather than constant constitutive expression. This kind of stimulus-responsive genetic control already exists in limited form in current synthetic biology research; extending it to a multi-signal, multi-output “circuit” inside a living gut symbiont is the genuinely unproven, ambitious part of the vision.

4. How much I would need to make, and how long it would take Defining a scale: enough capsules to supply 1 million people with one daily dose for a year (a defensible starting scale for a proof-of-concept product aimed at a chronic-inflammation-linked population).

Dose: based on real dosing data from comparable engineered live biotherapeutics currently in clinical development (e.g., MH002, an engineered EcN-based therapy: 1×10¹⁰–4×10¹⁰ CFU/day), I’ll assume 1×10¹⁰ CFU per daily capsule as a working estimate.

Total CFU needed per year: 1,000,000 people × 1×10¹⁰ CFU/day × 365 days = 3.65×10¹⁸ CFU/year

Achievable fermentation titer: E. coli is one of the best-characterized industrial fermentation organisms; high-density fed-batch fermentation routinely achieves on the order of 1×10¹⁰ CFU/mL (1×10¹³ CFU/L) — I’m flagging this as an assumption based on typical high-cell-density E. coli culture benchmarks that I’d want to confirm against a specific published process before finalizing.

Volume needed per year: 3.65×10¹⁸ CFU ÷ 1×10¹³ CFU/L = ~365,000 L (365 m³) of culture per year For context, this is well within a scale already demonstrated for other engineered bacterial biomanufacturing (comparable extremophile fermentation has been scaled to 200 m³ single reactors).

Time: assuming a 24-hour fed-batch growth cycle plus ~12 hours turnaround (cleaning, prep) between batches = 36 hours/batch. Using a single 10,000 L (10 m³) production reactor: 365,000 L ÷ 10,000 L = 36.5 batches needed → 36.5 × 36 hours ≈ 1,314 hours, or about 55 days of production time per year to hit the full 1-million-person target — meaning even one mid-sized industrial reactor could meet this scale with significant capacity to spare, or the target population could be scaled up substantially within the same reactor.

5. What I don’t know yet Whether IAP and butyrate production can be stably co-expressed in one EcN strain without one pathway suppressing the other or creating excessive metabolic burden that tanks growth/yield — this is the single biggest unknown and the first thing I’d want to test experimentally.

Whether enzymatically neutralizing LPS locally in the gut meaningfully reduces systemic inflammation in practice, versus just local gut effects — this would need to be measured (e.g., blood inflammatory markers) rather than assumed from the mechanism alone.

Real achievable fermentation titer and batch cycle time for this specific engineered strain (my numbers above are based on general E. coli fermentation benchmarks, not this construct specifically) — actual values could shift the scale-up math meaningfully in either direction. (For the bigger vision): whether a genetic circuit responsive to multiple internal signals, with a safe and reliable external “toggle” mechanism, is achievable at all with current synthetic biology tools, or whether it requires fundamentally new control systems that don’t exist yet.

AI-citation: AI assistance (Claude) was used to help research mechanisms, structure this proposal, and draft calculations; prompts covered analysing sources, brainstorming project direction, researching gut-inflammation biology, and building the scale-up math.

Final Project