Week 1 HW: Final Project Ideas

(1) What you would like to biomanufacture and why?

Describe the target product or output you would like to make. Why are you interested in making it? Your motivation might be utilitarian—for example, addressing a problem in health, sustainability, food, energy, or manufacturing—but it does not need to be. Your project might instead have a social, cultural, artistic, or personal motivation. Clearly articulate what you want to make, who or what it is for, and why you think it is interesting or important.

This semester, for my project, I am tackling the problem of how to use wasted hay on the goat farm as a biomanufacturing medium. I believe a goat’s preference for living in a hay-filled environment dates back in evolutionary history to mountain patches at the edge of woodlands, particularly at higher elevations. The picture above isn’t my farm. Those goats live at the Archie Warder Tierpark in Northern Germany, a conservation park for historic domestic animal breeds. During my visit, the owner explained that some of the goats maintained there descend from the type of goats associated with the Galápagos populations described by Herman Melville in The Encantadas. Archie Warder taught me that conservation has inspired my desire to better engineer environments for animals on my farm to support animal health and further renewable agriculture objectives secondary to this project.

Let me back up a bit. My name is Charley Naney, and I am a goat and sheep farmer and a Ph.D. student in Bioscience at North Carolina Agricultural and Technical State University.

In the lab, I study bidirectional interactions between Metazoan host physiology and Microbiomes related to oxidative stress in aging and spaceflight (Naney, 2026).

On the farm, in second picture, I push hay and shovel waste from my goats and sheep. I also have chickens and turkeys, but cycling their waste is an entirely different challenge that I won’t address this semester. I spend a lot of time moving animals around the land and maintaining hay supplies when the goats and sheep have eaten all of the available forage. Much of that hay eventually ends up on the ground, where it becomes mixed with manure, soil, and moisture. From the perspective of farm maintenance, this material is waste that I have to move. From the perspective of biomanufacturing, however, it is a continuously generated biological feedstock.

That distinction is the starting point for my project. Rather than asking how to dispose of this material, I want to ask whether a living system can transform it into something useful for the animals that produced it. Specifically, I want to investigate whether waste hay, goat manure, and locally available mineral material can become feedstocks for a biomineralized composite that could eventually be used to construct an artificial climbing landscape for goats.

(2) What living system would you use? Propose at least one living-system manufacturing chassis that could produce your target—for example, a microbial, mammalian, or cell-free system. At a high level, describe how the system would make your target product. What goes into the system? What biological process or processes occur? What comes out? You do not need to design the complete biological system yet. Make reasonable assumptions where necessary and identify them.

I have already discovered that soiled goat hay is highly valuable for building biologically active soil rich in bacteria and earthworms. When waste hay becomes mixed with manure, urine, soil, and water, it does not remain waste for very long. Bacteria, fungi, earthworms, insects, plant roots, and other organisms gradually decompose and redistribute this material. I already take advantage of this process on my farm by burying organic material and allowing the existing soil community to cycle it back into the land.

For this project, I want to build on that living system rather than replace it. My goal is to create an artificial climbing landscape for goats—a small “Goat Mountain”—whose construction incorporates materials already moving through the farm. I envision a core constructed from locally available clay or compressed-earth blocks, potentially incorporating waste hay as fiber. Soil, hay, manure-derived material, and other suitable farm residues could then be incorporated around and within the structure where the farm’s existing decomposer community could continue breaking down organic matter and building soil.

The engineered biological component would introduce a second process: microbial biomineralization. A selected bacterial system capable of promoting calcium carbonate precipitation could be used to mineralize portions of the earthen composite. Microbial activity changes the local chemical environment so that calcium and carbonate can precipitate as solid calcium carbonate:

\mathrm{Ca^{2+}+CO_3^{2-}\rightarrow CaCO_3(s)}

Rather than asking bacteria to manufacture an entire mountain, I would use this process to help form mineral bridges between particles in selected blocks or surface layers. The clay or earthen core would provide bulk and support, while microbial mineralization could contribute cohesion, erosion resistance, and a harder climbing surface. Whether it actually provides those properties—and how much mineralization is necessary—would need to be experimentally measured.

What interests me most is that construction would not necessarily end when the structure was built. I could continue using the biological processes already operating on my farm by adding or burying appropriate organic residues in designated soil-filled areas of the structure. Earthworms and soil microorganisms could decompose those materials, plants could colonize portions of the surface, roots could stabilize soil, and nutrients could continue cycling through the system. Meanwhile, selected mineralized regions could provide the harder surfaces necessary for hoof traffic and climbing.

The resulting structure would therefore contain two complementary biological processes. One promotes decomposition and nutrient cycling, turning organic waste back into biologically active soil. The other promotes mineralization, turning part of the mineral matrix toward a more persistent solid state. Instead of treating these as opposing processes, I want to investigate whether they can occupy different compartments of the same structure.

The full-scale structure I have in mind resembles the climbing landscape shown in the photograph from Arche Warder. Rather than manufacturing that entire volume as an experimental biomaterial, I would use clay or earthen blocks for much of the core and reserve biomineralized composite for the portions where its properties are useful. This should reduce the quantity of biologically manufactured material required while allowing the structure itself to remain integrated with the biological cycling already occurring on the farm.

At the beginning of the project, I do not yet know which microbial species, mineral composition, calcium source, or farm-derived feedstock will produce the best material. I also do not know how the mineralized and biologically active regions will interact over time. Those uncertainties are part of what I want to measure rather than assumptions I want to make.

What interests me most is that construction would not necessarily end when the structure was built. I could continue using the biological processes already operating on my farm by adding or burying appropriate organic residues in designated soil-filled areas of the structure. Earthworms and soil microorganisms could decompose those materials, plants could colonize portions of the surface, roots could stabilize soil, and nutrients could continue cycling through the system. Meanwhile, selected mineralized regions could provide the harder surfaces necessary for hoof traffic and climbing.

The shape of the mountain could itself contribute to this cycling process. Goats using the elevated climbing surfaces would inevitably deposit manure, wasted hay, hair, and other organic material on the structure. Some of this material would remain in cracks and soil-filled pockets, while some would be transported downslope by gravity, animal movement, and rainfall. Rather than allowing that material simply to accumulate or wash away, the perimeter of the mountain could be designed as a biologically active capture zone containing soil, buried organic matter, plant roots, earthworms, and the microbial communities already present on the farm. Organic material moving off the harder mineralized surface would therefore be intercepted at the periphery, where it could decompose and be reincorporated into the soil.

This would make Goat Mountain more than a static climbing structure. It would function as a small engineered landscape with different biological compartments: a durable mineralized surface for animal use, protected areas where organic matter can accumulate and decompose, and a living perimeter that captures material moving off the structure and returns nutrients to the surrounding soil.

(3) How would you use automation? Identify at least one automation technology that you would combine with your living system. This could include a large-scale Cloud Lab, a desktop liquid-handling robot such as an Opentrons, a programmable microfluidics platform such as Nuclera’s, or another automation technology. Describe what part of the biomanufacturing workflow you would automate and why. For example, you might automate liquid handling, culturing, assembly, screening, sampling, measurement, optimization, or another step.

I would initially use automation to optimize the biological process rather than to manufacture the final structure. The composition of the farm-derived material and the conditions required for microbial mineralization contain many variables that could influence calcium carbonate precipitation. These could include the proportion of waste hay, manure-derived material or ash, soil and clay composition, calcium availability, pH, moisture, microbial inoculum, and incubation time.

At the laboratory scale, an automated liquid-handling platform such as an Opentrons robot could prepare and maintain a matrix of small experiments in which these variables are systematically changed. Automation could dispense microbial cultures, nutrients, calcium-containing solutions, and other liquid components; collect samples at predetermined intervals; and prepare samples for measurements of pH, microbial growth, and mineral precipitation. This would allow many candidate formulations to be compared under reproducible conditions rather than developing the material through trial and error.

The most promising conditions could then be transferred from liquid or slurry screening into larger material coupons or blocks containing the actual farm-derived substrate. At this stage, automation would shift from liquid handling toward process control. Pumps and valves could regulate water, nutrients, or mineralization solutions; sensors could monitor moisture, temperature, and pH; and imaging could document changes in cracking, shrinkage, and surface structure over time.

Automation at manufacturing scale would therefore look very different from automation during discovery. A laboratory robot could help identify the biological and chemical recipe, whereas a larger reactor or controlled delivery system would reproduce those conditions while mineralizing kilograms of material. The goal would not be to have a robot construct the mountain one pipette tip at a time, but to use automation to determine and maintain the environmental conditions under which the microorganisms manufacture the mineral cement.

(4) How much would you need to make, and how long would it take? Perform a back-of-the-envelope calculation estimating what it would take to manufacture your target at a meaningful scale. (a) How much? Define a scale relevant to your project and calculate approximately how much of your target product you would need to manufacture. For example, if you are producing a medicine, you might estimate the amount required to treat 1,000 people, everyone in Massachusetts, or everyone in the world. If you are producing a building material, you might calculate the amount required for one structure or an entire neighborhood. If you are producing an art or design artifact, define an appropriate scale yourself. Clearly show your calculation, including units, assumptions, and sources for important numbers.

For this project, I define meaningful manufacturing scale as enough biomineralized composite to cover one goat climbing structure comparable to the one shown at Arche Warder. Using the goats in the photograph as approximate scale references, I estimate the structure to be approximately 10–12 m long, 6–8 m wide, and 2–3 m high. Rather than manufacturing the entire mound from biomaterial, I propose constructing a bulk earthen core and manufacturing an approximately 10-cm-thick biologically mineralized climbing surface. A preliminary geometric estimate places the required composite volume on the order of 10 m³.

(b) How long? Using your proposed living system and automation technology, estimate how long it would take to manufacture that amount. Consider quantities such as production rate, yield, volume, number of parallel experiments or reactors, automation throughput, or whatever parameters are most appropriate for your system. You may discover that the automation technology you would use to prototype your system is very different from what you would use to manufacture at scale. If so, describe both. Again, we are looking for a reasonable engineering estimate. Show your assumptions and calculations. 5. What don’t you know yet? After completing your analysis, identify the 2–3 biggest technical unknowns, bottlenecks, or assumptions in your proposed biomanufacturing system. What would you need to learn, measure, or experimentally determine to know whether your idea could actually work?

Two observations provide the starting conditions for this project. First, goats waste a substantial amount of hay. Second, on my farm I have repeatedly observed that when this hay is incorporated into soil with animal waste and sufficient moisture, biological decomposition can convert it into soil-like organic material over a period of months. This process does not depend on whether my proposed biomineralized composite ultimately succeeds.

The farm also already contains other materials that could contribute to construction. Clay-rich soil is abundant in North Carolina and on my land, providing a readily available mineral material that could be evaluated for compressed-earth blocks, fill, or the matrix of the mountain itself. Naturally fallen trees provide another material stream. Rather than removing decomposing wood from the system, suitable pieces could be incorporated into non-load-bearing portions of the mountain or its peripheral decomposition zones, where they could retain moisture, provide organic carbon, create habitat for decomposer organisms, and gradually become incorporated into the developing soil. Waste hay, manure, fallen wood, clay, soil, microorganisms, earthworms, water, and eventually plant roots would therefore become components of the same engineered landscape.

The engineering problem is not simply how to decompose goat waste, because that biological process already occurs. The challenge is to balance containment with moisture removal. Organic material must remain in the system long enough for microorganisms, fungi, earthworms, and other decomposers to act on it, while excess water must be able to drain away. I envision the harder climbing surfaces shedding water and loose organic material downslope toward biologically active peripheral capture zones. These zones could contain soil, waste hay, manure, and decomposing wood and be designed to retain organic matter while permitting infiltration and drainage.

This creates two coupled but independently useful systems. The first is an existing biological process in which farm residues are captured, decomposed, and incorporated into soil. The second is experimental: using microbial biomineralization to precipitate calcium carbonate and produce or strengthen selected surface materials. If the biomineralized composite is unsuccessful, the mountain can still function as a landscape for capturing and cycling organic matter. If it succeeds, the same system gains a biologically manufactured, more durable climbing surface.

My longer-term goal is therefore not to import enough manufactured material to build an artificial rock and place it on the farm. It is to determine how much of the structure can be built from materials and living processes already present within the farm system, while identifying the minimum external inputs necessary to make the structure durable and safe.

AI

Feel free to use AI and our learning tools to be shared in recitation to assist you, but remember that you will present in class, so make sure you have a solid understanding of your responses.

Please also cite your AI client and prompts.

AI Use Statement: OpenAI ChatGPT (GPT-5.6 Sol) was used to help organize the project concept, refine written explanations, and explore potential biological manufacturing approaches. The author evaluated and revised the resulting material and is responsible for the final content.

Representative prompts: “Can you help me organize this biomanufacturing project around using wasted hay from my goat farm?” “What living system could transform farm-derived material through microbial mineralization?” “How could microbial calcium carbonate precipitation be incorporated into an artificial climbing landscape for goats?” “How could automation be used to prototype and optimize this biomanufacturing process?”

Readings:

Asin-Garcia, E., Fawcett, J. D., Batianis, C., & Martins dos Santos, V. A. P. (2025). A snapshot of biomanufacturing and the need for enabling research infrastructure. Trends in Biotechnology, 43(5), 1000–1014. https://doi.org/10.1016/j.tibtech.2024.10.014 Averesch, N. J. H., Berliner, A. J., Nangle, S. N., Zezulka, S., Vengerova, G. L., Ho, D., Casale, C. A., Lehner, B. A. E., Snyder, J. E., Clark, K. B., Dartnell, L. R., Criddle, C. S., & Arkin, A. P. (2023). Microbial biomanufacturing for space-exploration—what to take and when to make. Nature Communications, 14, 2311. https://doi.org/10.1038/s41467-023-37910-1 Karim, A. S. & Jewett, M. C. (2021). “Toward sustainable, cell-free biomanufacturing.” Current Opinion in Biotechnology 69, 136–144. DOI: 10.1016/j.copbio.2020.12.012. OpenAI. (2026). ChatGPT (GPT-5.6 Sol) [Large language model]. ⁠https://chatgpt.com/ Petzold, C. J., & Mukhopadhyay, A. (2025). From bench to biofactory: High-throughput technologies and automated workflows to accelerate biomanufacturing. Current Opinion in Biotechnology, 94, 103320. https://doi.org/10.1016/j.copbio.2025.103320 Yap, W. S., Choudhury, D., & Suntornnond, R. (2023). Towards biomanufacturing of cultured meat. Trends in Biotechnology, 41(3), 292–294. https://doi.org/10.1016/j.tibtech.2023.01.014 Ye, J.-W., Lin, Y.-N., Yi, X.-Q., Yu, Z.-X., Liu, X., & Chen, G.-Q. (2023). Synthetic biology of extremophiles: A new wave of biomanufacturing. Trends in Biotechnology, 41(3), 342–357. https://doi.org/10.1016/j.tibtech.2022.11.010 Zhang, Y.-H. P., Sun, J., & Ma, Y. (2017). Biomanufacturing: History and perspective. Journal of Industrial Microbiology & Biotechnology, 44, 773–784. https://doi.org/10.1007/s10295-016-1863-2

Week 2 HW: DNA Design

Homework — Tuesday September 29, by 2:30p ET

Elowitz, M. B. & Leibler, S. A synthetic oscillatory network of transcriptional regulators. Nature 403, 335–338 (2000). https://doi.org/10.1038/35002125 Stanton, B. C. et al. Genomic mining of prokaryotic repressors for orthogonal logic gates. Nature Chemical Biology 10, 99–105 (2014). doi.org/10.1038/nchembio.1411 Carr, P. A. et al. Protein-mediated error correction for de novo DNA synthesis. Nucleic Acids Research 32, e162 (2004). https://doi.org/10.1093/nar/gnh160

Gibson, D. G. et al. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nature Methods 6, 343–345 (2009). https://doi.org/10.1038/nmeth.1318

Kosuri, S. et al. Scalable gene synthesis by selective amplification of DNA pools from high-fidelity microchips. Nature Biotechnology 28, 1295–1299 (2010). https://doi.org/10.1038/nbt.1716

Kong, D. S. et al. Parallel gene synthesis in a microfluidic device. Nucleic Acids Research 35, e61 (2007). https://doi.org/10.1093/nar/gkm121

Tian, J. et al. Accurate multiplex gene synthesis from programmable DNA microchips. Nature 432, 1050–1054 (2004). https://doi.org/10.1038/nature03151

Readings Folder: https://drive.google.com/drive/folders/1Mdc1KPKCZgLyiN42loqJSqG71fRpdffN?usp=sharing

Week 3 HW: Cell Free

HTBAA Week 3: Designing a Cell-Free Reaction Composition

Homework Due: October 6, 2026 Homework’s Associated Website: coopetition.htgaa.org Coopetition: coopetition specifications Homework created by: Ronan Donovan and Michelle Yue

1. Timeline & Overview Tue, Sept 29: Reshma Shetty lecture + coopetition introduction
   Thu, Oct 1: Ginkgo field trip + homework assignment introduced by Ronan
   Tue, Oct 6: Kate Adamala lecture + 30-min review of initial cell-free designs
   Thu, Oct 8: Recitation focused on global reaction composition ideation
   Week, Oct 12: Virtual office hours (TBD) to plan 80 unique compositions (no class time) Tue, Oct 20: HTBAA submits the official autonomous lab experiment request to Ginkgo

2. INTRODUCTION

A. HOMEWORK OVERVIEW

In this assignment, you will design new reaction compositions for synthesizing three proteins (sfGFP, PETase, and Reteplase) through cell-free protein synthesis (CFPS). In lysate-based CFPS, the protein-synthesis machinery is extracted from cells and used outside of intact living cells. The reaction combines a cell lysate containing the transcription and translation machinery with a DNA template, energy substrates, and other required reaction components to produce proteins in vitro. Because protein synthesis is decoupled from cell growth and viability, CFPS is highly tunable, supports rapid experimental iteration with results often available within hours, and is well suited for automated workflows (Silverman et al. 2020; Rasor et al. 2021).

This homework was designed based on an exciting ongoing project at Ginkgo Bioworks aimed at comparing the performance of human scientists and AI agents in designing cell-free protein synthesis reactions. Teams of researchers and AI agents are independently designing reaction compositions for the same three proteins (sfGFP, PETase, and Reteplase), with the goal of optimizing the cost of producing one gram of functional protein.

A central challenge in reaching this goal is in optimizing the reaction composition in which the CFPS takes place, as this strongly influences both CFPS yield and cost. Cell extract is typically the largest single expense in a CFPS reaction composition, while changes in the concentrations and combinations of other reaction components can alter protein yield and thus cost per gram by one to two orders of magnitude (Olsen et al. 2026; Smith et al. 2026). Because the optimal reaction composition for these proteins has not yet been established, identifying improved formulations requires highly iterative experimental screening.

However, the number of possible combinations of reaction components is too large to evaluate exhaustively. To address this challenge, Ginkgo Bioworks uses a staged screening strategy to efficiently evaluate candidate formulations. Reaction conditions are first tested in 384-well plates at 20 µL-scale volumes to minimize reagent use, and the best-performing conditions are then re-tested in 96-well plates to assess reproducibility. Final candidates are evaluated in 2 mL tubes. Each stage of screening narrows the candidate reaction composition while testing reproducibility and scale-up viability.

Fig 1. Ginkgo’s staged screening strategy. Testing begins in 384-well plates with best CFPS compositions further evaluated in 96-well and 2 mL tubes

B. THE CENTRAL DOGMA IN CFPS

Cells produce proteins through the central dogma of molecular biology: DNA ––transcription→ mRNA ––translation→ protein

During transcription, a sequence of DNA is used as a template to produce messenger RNA (mRNA). During translation, ribosomes read the mRNA sequence in three-nucleotide codons and use this information to assemble amino acids, the building blocks of proteins, into a polypeptide chain. The resulting chain then folds into a specific three-dimensional structure that determines protein function, such as sfGFP producing fluorescence, PETase hydrolyzing PET, and Reteplase dissolving clots.

Cell-free protein synthesis carries out these same processes outside of living cells.

For this project, E. coli BL21 Star (DE3) cells are first grown, lysed, and clarified by centrifugation to remove membranes and cellular debris. The resulting lysate, or cell extract, retains functional ribosomes, tRNAs (which transport amino acids to the ribosomes), and enzymes required for transcription and translation. The lysate typically comprises approximately one-quarter of the reaction volume.

A DNA template encoding the desired protein is then combined with the lysate and a chemical mixture containing nucleotide triphosphates (or nucleotide monophosphates) for mRNA synthesis, twenty amino acids for protein synthesis, an energy system that regenerates ATP, and salts such as magnesium and potassium that are required for proper function of the transcriptional and translational machinery. Once these components are combined, transcription and translation proceed within the reaction.

Importantly for this project, the gene is placed downstream of a T7 promoter and transcribed by T7 RNA polymerase, a fast and robust single-enzyme transcription system derived from a bacteriophage. The polymerase is supplied by the source E. coli strain—which is a BL21 Star (DE3) derivative that carries the T7 RNA polymerase gene—and is therefore already present in the lysate. Because transcription is carried out by a single polymerase, the same reaction framework can be used for nearly any gene placed downstream of a T7 promoter, contributing to the flexibility of CFPS programming.

Fig 2. DNA is transcribed into mRNA which is translated by ribosomes into a folded protein, all in a cell free reaction tube.

These features make CFPS well suited for automation, optimization, and bioproduction. Because the reaction does not contain an intact cell membrane and does not require maintenance of living cells, individual components can be added, removed, or adjusted to directly evaluate their effects on protein production. Automated systems are particularly advantageous in that they can prepare and analyze hundreds of reaction compositions in parallel.

2. CFPS EXPERIMENTAL FRAMEWORK

A. OVERVIEW OF PROTEIN TARGETS

Each reaction composition is evaluated using three proteins (i.e., sfGFP, PETase, and Reteplase), which span a range of expression and folding difficulty. Why these three proteins?

sfGFP is a widely used fluorescent reporter protein that folds robustly and contains no disulfide bonds; its two cysteines remain as free thiols. It therefore provides a relatively simple baseline for evaluating whether a CFPS reaction is functioning.

PETase is a cutinase from Thermobifida fusca (TfCut2) that hydrolyzes polyethylene terephthalate (PET). TfCut2 contains one disulfide bond, making it somewhat more challenging to fold than sfGFP.

Reteplase is a clot-dissolving therapeutic derived from human tissue plasminogen activator (tPA). It contains nine disulfide bonds that must form in the correct pairings for the protein to function, making it the most challenging of the three proteins to produce in an active form.

A reaction composition that performs well for one protein may therefore perform poorly for another. Testing the same composition across proteins with increasing folding difficulty provides a measure of how broadly the formulation performs. Of note, in one study, a cost-optimized CFPS formulation was tested across twelve additional proteins; six were ultimately detectable by gel electrophoresis and nine were detectable using a more sensitive assay, illustrating how CFPS performance remains partly protein-dependent (Smith et al. 2026).

B. EXPRESSION CHALLENGES OF EACH PROTEIN TARGET

sfGFP: Fluorescent Reporter

Fig 3. Superfolder GFP (PDB 2B3P; UniProt P42212). The β-barrel structure surrounds the fluorescent chromophore.

Superfolder green fluorescent protein (sfGFP) is ~240 amino acids long and 27 kDa. Its green fluorescence is produced by a chromophore formed autocatalytically from Thr65, Tyr66, and Gly67 through cyclization and dehydration, followed by an oxygen-dependent oxidation step. sfGFP contains two cysteines, but they do not form a disulfide bond. Because sfGFP folds reliably, it is commonly used as a CFPS reference protein; failure to produce fluorescent sfGFP generally indicates that the reaction itself is not functioning properly.

PETase: PET-degrading enzyme

Fig 4. TfCut2, a thermostable PET-degrading cutinase (PDB 4CG1; UniProt Q6A0I4).

The PETase used in this project is TfCut2, a cutinase from Thermobifida fusca that is ~260 amino acids and 28 kDa in its active form. TfCut2 hydrolyzes ester bonds in PET, the plastic used in beverage bottles and polyester. It is a serine hydrolase with a catalytic triad composed of Ser170, Asp216, and His248. TfCut2 contains one disulfide bond, making its folding requirements more demanding than those of sfGFP, providing a test of whether a reaction composition can produce an active enzyme rather than only a fluorescent reporter.

Reteplase: therapeutic protease

Fig 5. The catalytic domain of tissue plasminogen activator (PDB 1RTF; tPA UniProt P00750). Reteplase is a truncated form of tPA; yellow sticks indicate disulfide bonds.

Reteplase is a 355-amino-acid, ~39 kDa therapeutic protein derived from human tPA and used to dissolve blood clots. Full-length tPA contains five domains and 17 disulfide bonds, while Reteplase retains the kringle-2 and serine protease domains and contains nine disulfide bonds. These bonds must form in the correct pairings for the protein to adopt its active structure.

It is therefore substantially more difficult to fold than sfGFP or PETase. Industrially, it is produced in E. coli as inactive inclusion bodies and refolded in a separate chemical step. A standard reducing CFPS formulation does not support formation of its required disulfide bonds, so active Reteplase requires an oxidizing CFPS formulation.

C. DISULFIDE BONDS: REDUCING AND OXIDIZING CONDITIONS

Disulfide bonds are covalent bonds formed between the sulfur atoms of two cysteine residues. Formation of an S–S bond requires oxidation of two cysteine thiol groups and helps stabilize a protein’s three-dimensional structure. Differences in disulfide-bond requirements are therefore a major reason the three target proteins require different CFPS conditions.

Standard CFPS reactions are generally maintained under reducing conditions using DTT (dithiothreitol), which keeps cysteines in their free-thiol state and limits unwanted disulfide-bond formation. This environment is appropriate for sfGFP and remains compatible with TfCut2.

For Reteplase, however, DTT would prevent formation of its nine required disulfide bonds and therefore prevent correct folding. Reteplase instead requires an oxidizing, or disulfide-bond (DSB), formulation with three major modifications:

1. Modified lysate: Lysate is prepared without DTT and treated with iodoacetamide (IAM). IAM irreversibly caps free thiols and inhibits enzymes that would otherwise reduce newly formed disulfide bonds. Excess IAM is quenched with an equal amount of cysteine.

2. Redox buffer: Oxidized glutathione (GSSG) and reduced glutathione (GSH) are added in a controlled ratio to establish an oxidizing environment that favors disulfide-bond formation.

3. Folding helper: DsbC, a disulfide-bond isomerase, is added to rearrange incorrectly paired disulfide bonds. The lysate is also prepared from an E. coli strain lacking glutathione reductase (Δgor), helping preserve the oxidizing environment.

The two reaction regimes can therefore be summarized as follows:

Feature Reducing: sfGFP, PETase Oxidizing/DSB: Reteplase DTT in lysate Yes No Additional lysate treatment None IAM alkylation, then cysteine quench Redox buffer None GSSG + GSH Folding helper None DsbC E. coli strain BL21 Star (DE3) BL21 Star (DE3) Δgor Best suited for Proteins with few or no disulfide bonds Proteins requiring multiple disulfide bonds

C. MEASURING PROTEIN PRODUCTION AND ACTIVITY

Each protein is evaluated using an assay matched to its function. Across all assays, a standard curve converts an experimental signal into an estimated concentration (a proxy for activity), while positive and negative controls determine whether the measurement is reliable. A positive control should produce a strong signal, whereas a negative control (e.g., a reaction lacking DNA), should produce little or no signal.

(a) sfGFP: direct fluorescence Because sfGFP is intrinsically fluorescent, protein production can be measured directly. The completed reaction is excited near 485 nm, and green fluorescence is measured near 510–520 nm. This is an endpoint assay performed after the reaction is complete.

Greater fluorescence indicates a greater amount of correctly folded, active sfGFP. A standard curve generated using purified sfGFP converts fluorescence into protein concentration. Because only properly folded and active sfGFP fluoresces, the assay measures functional folded protein rather than total protein production.

(b) PETase: substrate hydrolysis at 405 nm PETase activity is measured using a soluble p-nitrophenyl ester substrate. Hydrolysis releases p-nitrophenol, which is present primarily as the yellow p-nitrophenolate ion at the assay pH and absorbs strongly near 405 nm.

Absorbance at 405 nm is measured repeatedly over time. The initial rate of increase in absorbance reflects the amount of active PETase present: a steeper slope indicates greater enzymatic activity. A standard curve using the colored product converts absorbance into product concentration. Samples are diluted as needed so that measurements remain within the linear range of the assay.

(c) Reteplase: fluorogenic substrate cleavage Reteplase activity is measured using the fluorogenic peptide substrate IPR-AMC. IPR refers to isoleucine–proline–arginine, while AMC is the fluorophore 7-amino-4-methylcoumarin. While AMC remains attached to the peptide, its fluorescence is quenched. Reteplase cleaves the substrate after arginine, releasing fluorescent AMC. The substrate is added directly to the reaction plate, and fluorescence is measured over time using excitation ~345 nm and emission ~445 nm.

The initial fluorescence rate is proportional to the amount of active Reteplase. Because Reteplase is active only when its nine disulfide bonds are correctly formed, enzymatic activity also provides a functional measure of correct protein folding.

D. Staged screening of reaction compositions

Candidate reaction compositions are evaluated through a staged 384-well → 96-well → 2 mL screening workflow. Each stage reduces the number of candidates while increasing reaction volume.

In the first stage, reaction compositions are screened in 384-well plates using microliter-scale volumes, allowing many conditions to be tested while minimizing reagent use. Each 384-well plate contains 80 reaction compositions, with each composition tested in four replicate wells; the first four columns are reserved for standard curves and controls, while columns 5–24 contain the experimental conditions. Promising compositions are then re-tested in 96-well plates to determine whether their performance is reproducible, and final candidates are evaluated in 2 mL tubes. The study proceeds through three rounds, each using this 384-well → 96-well → 2 mL screening workflow.

Fig 6. High-throughput CFPS screening workflow.

Fig 7. 384-well input plate.The initial 3 columns contain standard curves; column 4 contains positive controls and negative controls, the remaining 320 wells contain DNA templates and experimental samples.

E. Automated experimental platform

High-throughput CFPS screening is performed using Ginkgo’s automated Nebula platform. The microplates are transferred between instruments using robotic arms and a MagneMotion magnetic track.

An Echo 525 acoustic dispenser uses sound waves to transfer reagents in droplets as small as 25 nL without pipette tips. A Bravo liquid handler then adds DNA and lysate to initiate the reactions. Plate readers collect measurements at defined time points, and an Inheco heat block performs the 60°C inactivation step used for PETase assays.

This automation allows thousands of precisely assembled reactions to be prepared and measured on a fixed schedule, significantly for efficiency and reproducibly than what can be done manually.

Because each step is programmable, AI systems can also propose new reaction compositions and evaluate them iteratively.

Fig 8. Nebula in Action. (1) A plate is retrieved from cold storage. (2) Plates travel between stations on the MagneMotion magnetic track. (3) A 6-axis robotic arm moves plates on and off instruments. (4) A plate is loaded at the Echo 525 acoustic dispenser, which fires nanoliter reagent droplets with sound instead of pipette tips.

Fig 9. Ginkgo’s Nebula System, an automated experiment platform

F. Study design and reaction constraints: the coopetition

The project on which this assignment is based compares human-led and AI-led approaches to CFPS reaction design. Human research teams and OpenAI AI agents independently design reaction compositions for the three target proteins.

Reaction compositions are submitted as JSON files and automatically validated before experiments are performed. In a related Ginkgo study, fewer than 1% of AI-designed plates contained an error that reached the laboratory (Smith et al. 2026). Each reaction has a total volume of 20 µL and contains both fixed and variable components. Fixed components:

• 5,000 nL lysate (25% of reaction)
• 2,000 nL DNA template
• 2,000 nL base buffer (Total 9,000 nL and are not modified.) Variable components: 11,000 nL must be selected from the available reagent menu. For any unused volume, nuclease-free water is added to bring the total reaction volume to exactly 20,000 nL. All reagent volumes must be specified in multiples of 25 nL.

The available reagents fall into several functional categories: Category Examples Function Salts and buffers potassium glutamate, magnesium glutamate, HEPES-KOH, potassium formate Control ionic strength and pH Building blocks ATP, CTP, GTP, UTP, nucleotide alternatives, amino acids Supply material for RNA and protein synthesis Energy regeneration phosphoenolpyruvate, pyruvate, maltose, maltodextrin, glucose Regenerate ATP Cofactors folinic acid, NAD, coenzyme A, spermidine, cAMP Support enzymatic reactions Detergents and solvents Brij-35, DMSO Support solubility and stability Oxygen-related components catalase, pyruvate oxidase Influence oxygen availability and metabolism Disulfide components DsbC, GSSG, GSH Promote and correct disulfide-bond formation

Because reaction compositions are evaluated partly by cost, both reagent price and protein yield matter. Lysate costs approximately $4.00/mL and constitutes 25% of each reaction, making it typically the largest single cost. Several other reagents are also relatively expensive, including CTP, GTP, and UTP at approximately $28–39/mL, NAD at approximately $33/mL, and coenzyme A at approximately $104/mL. Many salts and sugars cost only fractions of a cent, and water is treated as free.

Cost optimization therefore requires reducing expensive components without sacrificing protein yield. In prior work, improved pH buffering, partial replacement of nucleotide triphosphates with less expensive nucleotide monophosphates, and addition of small amounts of spermidine reduced cost per gram of protein produced (Smith et al. 2026). Reagent costs in this assignment follow Olsen et al. (2026).

3. ASSIGNMENT: CFPS REACTION DESIGN FOR SFGFP, PETASE, RETEPLASE

Copy and paste this section (section 3, A-F) into your website! Note, coopetition.htgaa.org is intended to support you as you answer the following questions.

Goal: using the collective intelligence of the HTBAA community and accessible AI systems, design novel CFPS reaction compositions for each of the three target proteins: sfGFP, PETase, and Reteplase to reduce the cost per gram of functional protein. Each design should be compatible with the experimental framework described above and should reflect the distinct expression and folding requirements of that protein. You can learn more about the details of the coopetition here.

A. COMPARE CELL-FREE REACTION COMPOSITIONS (20 minutes)

Why does the concentration of potassium glutamate and magnesium glutamate differ between compositions? Which is more sensitive?

Explain how a PEP+NTP reaction formulation is different from a Ribose+NMP reaction formulation.

Choose two cell-free reaction compositions from the literature, compare and contrast the choice of reagents and the concentration of reagents.

Note: A recommended starting point would be to review a least one published formulation, such as PANOx-SP (Jewett & Swartz 2004), a cost-optimized formulation from Olsen et al. (2026), and/or an AI-designed formulation from Smith et al. (2026).

B. DESIGN ONE REACTION COMPOSITION FOR EACH OF THE THREE PROTEINS (40 minutes)

Sign in to coopetition.htgaa.org and publish your first unique 20 µL CFPS reaction composition for each of the following (don’t worry, you can resubmit later): sfGFP PETase Reteplase

Note: if you’re unable to login with your HTGAA keycloak account, please post in Zulip so that you can participate in the coopetition!

For each composition, specify the reagents you would adjust from the relevant categories, including salts, energy sources, building blocks, cofactors, and redox components.

As a HTBAA research team, we will have the opportunity to suggest new reagents to add to the cell-free reagent inventory. Contribute (or upvote) at least one reagent to the “Community Reagents” section of the website and provide a short rationale. Include in your homework documentation how you decided on that specific reagent and provide links to any relevant research.

Visit the cfps-coopetition forum page to ensure you will be able to take part in the discussion. Bonus: send a note in the channel about a CFPS optimization idea that you have! If you would like clarification on aspects of the project, please message in this dedicated channel.

C. ASIMOV KERNAL (OPTIONAL)

Build T7 compatible plasmids for cell-free expression using the amino acid coding sequences below:

sfGFP MSKGEELFTGVVPILVELDGDVNGHKFSVRGEGEGDATIGKLTLKFICTTGKLPVPWPTLVTTLTYGVQCFSRYPDHMKRHDFFKSAMPEGYVQERTISFKDDGKYKTRAVVKFEGDTLVNRIELKGTDFKEDGNILGHKLEYNFNSHNVYITADKQKNGIKANFTVRHNVEDGSVQLADHYQQNTPIGDGPVLLPDNHYLSTQTVLSKDPNEKGTRDHMVLHEYVNAAGITWSHPQFEK

PETase MEKKIWSHPQFEKGGSGENLYFQSGGASPAGGGANPYERGPNPTDALLEASSGPFSVSEENVSRLSASGFGGGTIYYPRENNTYGAVAISPGYTGTEASIAWLGERIASHGFVVITIDTITTLDQPDSRAEQLNAALNHMINRASSTVRSRIDSSRLAVMGHSMGGGGTLRLASQRPDLKAAIPLTPWHLNKNWSSVTVPTLIIGADLDTIAPVATHAKPFYNSLPSSISKAYLELDGATHFAPNIPNKIIGKYSVAWLKRFVDNDTRYTQFLCPGPRDGLFGEVEEYRSTCPFGGSGVSGWRLFKKIS

Reteplase MEKKIWSHPQFEKGGSGENLYFQSGGASPAGGGNSDCYFGNGSAYRGTHSLTESGASCLPWNSMILIGKVYTAQNPSAQALGLGKHNYCRNPDGDAKPWCHVLKNRRLTWEYCDVPSCSTCGLRQYSQPQFRIKGGLFADIASHPWQAAIFAKHRRSPGERFLCGGILISSCWILSAAHCFQERFPPHHLTVILGRTYRVVPGEEEQKFEVEKYIVHKEFDDDTYDNDIALLQLKSDSSRCAQESSVVRTVCLPPADLQLPDWTECELSGYGKHEALSPFYSERLKEAHVRLYPSSRCTSQHLLNRTVTDNMLCAGDTRSGGPQANLHDACQGDSGGPLVCLNDGRMTLVGIISWGLGCGQKDVPGVYTKVTNYLDWIRDNMRPGGSGVSGWRLFKKIS

Bonus question: design a T7 helper plasmid that you could add to the cell-free reaction to improve the yield of functional reteplase (general chaperones, disulfide-bond machinery, etc).

4. Use of AI

You are encouraged to use AI and online tools for this assignment to assist your reagent composition development process. As you will present your work during recitation, you must be able to explain and justify the reasoning behind your design.

If you use AI, ensure that you cite the AI client and your prompts used.

5. REFERENCES AND RESOURCES

• Smith, A. A., et al. (2026). Using a GPT-5-driven autonomous lab to optimize the cost and titer of cell-free protein synthesis (Ginkgo Bioworks × OpenAI). bioRxiv preprint. DOI: 10.64898/2026.02.05.703998. https://www.biorxiv.org/content/10.64898/2026.02.05.703998v1

• Olsen, M. L., Copeland, C. E., Sundberg, C. A., Aw, R., Shaver, Z. M., Rao, G., Swartz, J. R., Karim, A. S., &

• Jewett, M. C. (2026). Design-driven optimization of low-cost reagent formulations for reproducible and high-yielding cell-free gene expression. Nature Communications. DOI: 10.1038/s41467-026-69605-8.

• Jewett, M. C., & Swartz, J. R. (2004). Mimicking the Escherichia coli cytoplasmic environment activates long-lived and efficient cell-free protein synthesis. Biotechnology and Bioengineering, 86(1), 19–26. DOI: 10.1002/bit.20026.

• Silverman, A. D., Karim, A. S., & Jewett, M. C. (2020). Cell-free gene expression: an expanded repertoire of applications. Nature Reviews Genetics, 21, 151–170. DOI: 10.1038/s41576-019-0186-3.

• Rasor, B. J., Vögeli, B., Landwehr, G. M., Bogart, J. W., Karim, A. S., & Jewett, M. C. (2021). Toward sustainable, cell-free biomanufacturing. Current Opinion in Biotechnology, 69, 136–144. DOI: 10.1016/j.copbio.2020.12.012.

Final Project