HTGAA 2026 - Week 1 Assignment Principles & Practices in Biotechnology Adrian Guzman Ph.D Centro de Rehabilitación Neuropsicológica Due: February 10, 2026

1. Biological Engineering Application

Application: Synthetic Gene-Based Neurotransmitter Optimization System

I propose developing a personalized synthetic biology platform for optimizing neurotransmitter production in patients with neuropsychological disorders. This system would use engineered bacterial chassis (E. coli or yeast) containing synthetic gene circuits that produce specific neurotransmitter precursors or modulators tailored to individual patient neurochemistry profiles.

Why This Application?

At Centro de Rehabilitación Neuropsicológica, we work with patients suffering from traumatic brain injury, stroke, neurodegenerative diseases, and psychiatric conditions. Current pharmaceutical interventions often have limited efficacy due to: Individual variability in neurotransmitter metabolism Side effects from systemic drug delivery Inability to produce sustained, regulated release of therapeutic compounds Limited options for combination therapies Leveraging next-generation gene synthesis technologies described in Professor Jacobson's lecture (chip-based synthesis enabling ~50 Mbp for $1K), we could rapidly prototype and test thousands of genetic circuit variants. This would allow us to create personalized neurotransmitter-modulating organisms that could be administered orally or through gut microbiome engineering.

Technical Approach

The system would utilize: Chip-based oligonucleotide synthesis (up to 1M oligos/chip) to create gene libraries encoding neurotransmitter biosynthesis pathways Error-correcting gene synthesis with MutS repair systems to achieve <1:10⁴ error rates Gibson assembly for constructing metabolic circuits (3-10 kbp pathways) Programmable biosensors using CRISPR/Cas9 guides with embedded strand-displacement programs for real-time monitoring Field Programmable Gene Arrays (bioFPGA concept) allowing rapid reprogramming of therapeutic production based on patient response

2. Governance and Policy Goals

Primary Goal: Ensuring Non-Malfeasance and Patient Safety

Sub-goal 2.1: Prevent Unintended Ecological Impacts

Ensure that engineered organisms cannot persist in the environment or transfer genetic material to wild-type bacteria. This includes preventing horizontal gene transfer of synthetic neurotransmitter pathways that could disrupt natural microbial ecosystems.

Sub-goal 2.2: Prevent Misuse for Neuroenhancement or Coercion

Establish safeguards against the use of neurotransmitter-modulating organisms for non-therapeutic purposes, including cognitive enhancement in healthy individuals, behavioral control, or as agents of chemical/biological weapons.

Sub-goal 2.3: Ensure Equitable Access

Prevent this technology from becoming available only to wealthy patients or institutions, thereby exacerbating healthcare disparities in neuropsychological treatment.

Sub-goal 2.4: Protect Patient Autonomy and Informed Consent

Ensure patients fully understand the novel nature of synthetic biology therapeutics, including uncertainties about long-term effects and the ability to discontinue treatment.

3. Proposed Governance Actions

Action 1: Genetic Biocontainment Requirements

Purpose

Currently, therapeutic bacteria used in microbiome research are not required to have kill-switches or auxotrophic dependencies. I propose mandating that all neurotransmitter-producing organisms must incorporate at least two orthogonal biocontainment systems: (1) synthetic auxotrophy requiring non-natural amino acids, and (2) conditional lethality circuits that activate outside specific pH/temperature ranges found in the human gut.

Design

This would require: FDA/COFEPRIS regulation mandating dual biocontainment for live therapeutic organisms Academic researchers and biotech companies to design and validate biocontainment before IND (Investigational New Drug) applications Third-party testing laboratories to verify biocontainment efficacy Post-market surveillance requiring patient stool samples to confirm no environmental persistence

Assumptions

Key uncertainties: Assumption that biocontainment systems won't evolve escape mutants over time Assumption that dual systems provide sufficient redundancy (may need triple) Uncertainty about long-term evolutionary stability in complex gut environments

Risks of Failure & Success

Failure risks: If biocontainment systems fail, engineered organisms could establish in wastewater systems or natural environments, potentially producing neurotransmitter precursors that affect wildlife behavior. Success risks: Overly stringent biocontainment requirements might make organisms non-viable in therapeutic contexts, or add significant development costs that limit commercialization and patient access.

Action 2: Neuro-Synthetic Biology Screening Framework

Purpose

Currently, gene synthesis companies screen orders against pathogen databases (IGSC guidelines) but don't specifically evaluate neuropharmacological risks. I propose creating a new screening framework where companies must flag sequences encoding enzymes in neurotransmitter biosynthesis pathways (serotonin, dopamine, GABA, acetylcholine) and require end-use verification for large orders.

Design

Implementation requires: Gene synthesis providers (Twist, IDT, GenScript) to expand screening databases Industry consortium to develop reference database of concerning neuropharmacological sequences Researchers ordering flagged sequences to provide IRB approval or institutional biosafety committee clearance Annual audits by industry association or government agency

Assumptions

This assumes: That neuro-synthetic sequences pose meaningful dual-use risks (vs. existing pharmaceutical manufacturing) That screening can effectively identify concerning sequences despite sequence diversity in enzyme families That bad actors wouldn't simply synthesize sequences in fragments to avoid detection

Risks of Failure & Success

Failure: Screening could be circumvented through distributed synthesis, offshore providers, or in-house chip-based synthesis (which Professor Jacobson's work makes increasingly accessible at ~$1K per 50 Mbp). Success: Could create friction for legitimate neuroscience research, particularly for groups working on brain organoids, neural circuit mapping, or psychiatric disorder models. May disproportionately impact researchers in countries with less developed IRB infrastructure.

Action 3: Patient-Centered Neuro-Synthetic Therapeutics Registry

Purpose

No current system tracks long-term outcomes of patients receiving engineered microbiome therapies for neuropsychological conditions. I propose creating a mandatory international registry where clinicians must report patient outcomes, adverse events, and discontinuation rates for all neuro-synthetic treatments. This would be modeled on cancer registries but specifically designed for living therapeutic products.

Design

Implementation requires: WHO or regional health authorities (PAHO, EMA, FDA) to establish registry infrastructure Healthcare providers and rehabilitation centers to report data Patient consent for long-term monitoring with privacy protections Funding mechanism (potentially fee from biotech companies per patient enrolled) Public data access for researchers while protecting patient privacy

Assumptions

This assumes: Clinicians will comply with reporting requirements (experience with drug registries shows variable compliance) Patients will consent to long-term monitoring The data collected will be sufficient to detect rare adverse events or unexpected long-term effects

Risks of Failure & Success

Failure: Poor compliance or inadequate funding could result in incomplete data that provides false reassurance about safety. Success: A robust registry could actually accelerate appropriate innovation by providing clear evidence of safety and efficacy, potentially reducing regulatory barriers. However, it might also reveal unexpected risks that trigger public backlash against all synthetic biology therapeutics.

4. Governance Options Scoring

Scoring: 1 = Excellent, 2 = Moderate, 3 = Poor, N/A = Not Applicable

5. Recommended Governance Strategy

Primary Recommendation: Layered Governance Approach

Based on the scoring analysis, I recommend implementing all three governance options in a coordinated fashion, prioritized as follows:

Phase 1 (Immediate): Genetic Biocontainment Requirements

Biocontainment scored best overall (1s in lab safety prevention, environmental protection, feasibility, and promoting constructive applications). This should be the foundation requirement as it directly addresses the highest-consequence risks (environmental release, horizontal gene transfer). The technology exists today and is already practiced voluntarily by leading research groups. Target audience: Federal regulators (FDA/COFEPRIS) to establish mandatory biocontainment standards Academic institutions to require biocontainment in IRB approvals for neuro-synthetic projects

Phase 2 (1-2 years): Patient Registry Establishment

The registry scored excellently in biosecurity response (1), lab safety response (1), environmental response (1), and not impeding research (1). It provides essential long-term safety monitoring without preventing innovation. This is critical for maintaining public trust and detecting unexpected issues early. Target audience: WHO/PAHO to coordinate international registry standards National health authorities to mandate reporting Professional organizations (rehabilitation medicine, neurology) to encourage participation

Phase 3 (3-5 years): Enhanced Screening Framework

Screening scored best in biosecurity prevention (1) and minimizing costs (1), but poorly in impeding research (3). I recommend delaying implementation until we have better data from the registry about actual misuse patterns. The field is too nascent to know which sequences truly merit screening, and premature implementation could stifle legitimate neuroscience research. Target audience: Gene synthesis industry consortium to develop screening criteria Academic biosecurity researchers to validate screening effectiveness

Trade-offs and Assumptions

Key Trade-off: Innovation vs. Precaution

The most significant tension is between enabling therapeutic innovation for patients who desperately need better treatments (many neuropsychological patients have tried all conventional therapies) versus preventing potential long-term harms that we may not fully understand yet. I prioritize biocontainment first because it addresses concrete, known risks without requiring us to predict what sequences might be misused.

Critical Uncertainty: Horizontal Gene Transfer Rates

The entire biocontainment strategy assumes we can engineer organisms with sufficiently low escape rates. Recent studies suggest gut environments may promote higher-than-expected horizontal gene transfer. If biocontainment proves inadequate, we may need to pivot toward abiotic delivery systems (synthetic nanoparticles, RNA therapeutics) rather than living organisms.

Equity Considerations

Compliance costs for biocontainment and registry reporting will disproportionately burden smaller rehabilitation centers and clinics in lower-resource settings. I recommend establishing funding mechanisms (perhaps through biotech company fees or government grants) to subsidize compliance for clinics serving underserved populations.

Assumption: Regulatory Harmonization

This strategy assumes reasonable harmonization between FDA, EMA, COFEPRIS, and other regulators. If regulatory fragmentation makes multinational research impractical, we may need stronger international coordination through WHO or a dedicated treaty mechanism (similar to the Biological Weapons Convention).

6. Ethical Reflections on Week 1 Lab Work

New Ethical Concerns from Pipetting Lab

During the basic pipetting exercises, several ethical considerations emerged that I hadn't fully appreciated before:

1. Accessibility and Technical Skill Barriers

Precise pipetting requires significant fine motor control and practice. For a rehabilitation center, this raises questions about who can actually implement these techniques. Patients with tremors, visual impairments, or cognitive processing difficulties would be excluded from participating in their own therapeutic development—a troubling form of exclusion. This suggests governance frameworks should include requirements for accessible protocol development and community involvement that doesn't require lab expertise.

2. Error Propagation in Chip-Based Synthesis

Professor Jacobson's slides show chip-based synthesis enabling 1M oligos per chip with error rates of 1:10⁴ even after correction. When synthesizing therapeutic genes, these errors could create off-target proteins with unknown pharmacological effects. The scale of synthesis (50 Mbp for $1K) makes it economically feasible for smaller groups but also means errors affect vastly more sequences than traditional synthesis. This raises questions about who bears responsibility for sequence validation—the synthesis provider, the researcher, or the clinic?

3. Dual-Use Implications of Democratized Synthesis

The dramatic cost reduction in gene synthesis ($1K for 50 Mbp vs. previous ~$1M) means small groups—including rehabilitation centers, community labs, or even individuals—could synthesize complete metabolic pathways for neurotransmitter production. While this democratization enables innovation, it also means we can't rely on expensive synthesis as a barrier to misuse. The bioFPGA concept is particularly concerning: the ability to 'reprogram' organisms with 1,500 different genes raises questions about whether we need new frameworks for tracking therapeutic organisms that can be altered post-deployment.

Proposed Additional Governance Actions

Action 4: Accessible Protocol Development Requirements

Require that all protocols for patient-directed therapeutic development include accommodations for individuals with disabilities. This could involve: Automated pipetting systems for patients with motor impairments Visual/auditory interface design for protocol documentation Community review processes that include patient representatives

Action 5: Sequence Validation Standards for Low-Cost Synthesis

Establish mandatory validation protocols for genes synthesized via chip-based methods when intended for therapeutic use: 100% sequencing confirmation for any construct intended for human or animal use Protein expression validation to confirm no unexpected products Clear liability frameworks for synthesis errors that cause patient harm Insurance requirements for groups using low-cost synthesis for clinical applications

7. Week 2 Lecture Preparation: DNA Read, Write, and Edit

Professor Jacobson's Questions

Question 1: Polymerase Error Rate vs. Human Genome Length

Error rate of polymerase: Nature's DNA polymerase has an error rate of approximately 1:10⁶ (one error per million base pairs) after proofreading mechanisms. From the lecture slides, chemical synthesis has a much higher error rate of 1:10² before correction. Comparison to human genome: The human genome is approximately 3.2 billion base pairs (3.2 Gbp). With a polymerase error rate of 1:10⁶, we would expect ~3,200 errors per replication if there were no additional correction mechanisms. How biology deals with this discrepancy: Biology employs multiple layers of error correction: 3'-5' proofreading exonuclease activity (shown in lecture slides) catches errors during replication MutS mismatch repair system (illustrated in lecture) scans for base-pairing errors post-replication These mechanisms reduce the final error rate to approximately 1:10⁹ to 1:10¹⁰, meaning only ~0.3-3 errors per genome replication Additional systems like base excision repair and nucleotide excision repair handle spontaneous damage

Question 2: Protein Coding Redundancy

Number of different ways to code for a protein: Due to codon degeneracy, there are astronomically many ways to code for an average human protein. For a protein of ~345 amino acids (average human protein is 1,036 bp ÷ 3 bp/codon), and considering that most amino acids have 2-6 coding options, the number of possible DNA sequences is approximately 10^200+ different nucleotide sequences that could code for the same protein. Why all these codes don't work: Codon usage bias: Different organisms preferentially use certain synonymous codons based on tRNA availability mRNA secondary structure: As shown in the lecture slides demonstrating minimum free energy structures at different GC content (10%, 50%, 90%), sequence composition dramatically affects RNA folding. Excessive secondary structure can block ribosome access or cause premature termination Ribosome pausing: Rare codons cause ribosomal stalling, affecting co-translational folding RNA stability: Some sequences create RNase cleavage sites, reducing transcript half-life Cryptic regulatory elements: Certain sequences inadvertently create splice sites, polyA signals, or transcription factor binding sites GC content extremes: The lecture demonstrates that 10% GC creates unstable transcripts while 90% GC creates excessive secondary structure—both problematic for expression

Dr. LeProust's Questions

Question 1: Most Common Oligo Synthesis Method

The most commonly used method for oligonucleotide synthesis is phosphoramidite chemistry, specifically the acid-based deprotection cycle illustrated in the lecture slides. This method uses: 5-minute cycles per base addition ~300 seconds per base throughput Error rate of 1:10² before correction Machine capacity of ~0.5 Mbp/year for traditional synthesizers The lecture also shows an alternative light-based deprotection method that offers potential advantages for chip-based array synthesis.

Question 2: Difficulty Making Oligos Longer Than 200nt

Direct synthesis beyond 200 nucleotides becomes difficult because: Error accumulation: With a per-base error rate of 1:10², errors accumulate multiplicatively. After 200 cycles, the probability of a completely correct sequence becomes very low: (0.99)²⁰⁰ ≈ 13.4% yield of perfect product Stepwise yield: Each coupling reaction is not 100% efficient (typically ~98-99%), so longer sequences have exponentially decreasing yields Chemical side reactions: Depurination, oxidation, and other side reactions accumulate over multiple cycles Purification challenges: Separating full-length products from truncated failure sequences becomes increasingly difficult

Question 3: Why Can't You Make a 2000bp Gene via Direct Synthesis?

Direct synthesis of 2,000 bp genes is impractical because: With a 1:10² error rate, a 2,000 nt synthesis would have (0.99)²⁰⁰⁰ ≈ 0.00000002% yield of error-free product—essentially zero Even with improved chemistry achieving 99.5% per-step yield, you'd get (0.995)²⁰⁰⁰ ≈ 0.004% perfect sequences This is why the lecture emphasizes assembly methods like Gibson assembly, where shorter oligos (150-200 bp) are synthesized separately and then assembled together, leveraging biological error-correction mechanisms The lecture shows that for genes and metabolic pathways (3-10 kbp), the practical approach is: (1) chip-based synthesis of overlapping oligos up to ~200 bp each, (2) error correction using MutS system, (3) assembly via Gibson assembly or similar homology-based methods.

Professor Church's Question (Selected)

Question Selected: AA:AA Interaction Code Proposal

Given that Professor Church's slides show: DNA nucleotides (4 types) code for amino acids (20 types) via triplet codons This gives us the NA:AA (nucleic acid to amino acid) code Proposed AA:AA interaction code: For encoding amino acid-amino acid interactions (such as in protein-protein binding sites, antibody-antigen interfaces, or enzyme-substrate specificity), I would propose a position-dependent triplet code where each interaction is encoded by three parameters: Interaction type: Encoded in first position (6 bits): Electrostatic (charge-charge): 00 Hydrophobic: 01 Hydrogen bonding: 10 Van der Waals: 11 Interaction strength: Encoded in second position (4 bits): Weak (<2 kJ/mol): 0 Medium (2-10 kJ/mol): 1 Strong (>10 kJ/mol): 2 Spatial geometry: Encoded in third position (5 bits): Face-to-face (β-sheet-like): 000 Helix packing: 001 Edge-to-edge: 010 Buried: 011 Rationale for this encoding: Draws on the Holliday Junction and DNA origami concepts from the lecture—if base pairing energies can create precise 3D structures (G/C ~-2.0 kcal/mol, A/T ~-1.2 kcal/mol), we can encode interaction energetics The 15-bit code (4+4+5+2 for specificity) gives us 32,768 possible interaction types, sufficient to capture the diversity of protein-protein interfaces This could be implemented using the molecular beacon/strand displacement logic shown in lecture, where DNA sequences report on specific AA:AA interactions For neuropsychological applications, this code could help design receptor-ligand pairs with precise affinity tuning Implementation using chip-based synthesis: Using the 1M oligos/chip capacity described in lecture, we could synthesize a library encoding all possible AA:AA interaction variants for a given protein binding domain. This would enable high-throughput screening of engineered protein interactions for therapeutic development.