Week 1 HW: Principles and Practices

1st Class Assignment

1. First, describe a biological engineering application or tool you want to develop and why. The Tool: “Neuro-Visor”: A Real-Time Bio-Computational Monitoring & Visualization Interface for Cerebral Assembloids. Currently, brain organoids are becoming more complex. We are moving from simple clusters to “Assembloids”, where different regions (like the cortex and thalamus) are fused together. While this is great for studying diseases like Alzheimer’s, it risks creating neural networks that exhibit “unintentional sentience” or organized electrical activity similar to a developing human fetus. As a UI/UX designer and Python developer, this tool would be a software-hardware interface that plugs into Multi-Electrode Arrays (MEAs), which are the chips these organoids grow on. It uses Python-based signal processing (Spike sorting, Fourier transforms) to translate raw electrical “noise” into a visual “Ethical Heatmap.” Without a standardized way to “see” the complexity of an organoid’s thoughts, we are flying blind. We need a tool that tells a researcher exactly when an organoid has crossed an ethical “Red Line.”

2. Next, describe one or more governance/policy goals related to ensuring that this application or tool contributes to an “ethical” future, like ensuring non-malfeasance (preventing harm). Break big goals down into two or more specific sub-goals. To prevent harm to potentially sentient tissue, the following goals are proposed: Goal: Standardizing “Neural Complexity Thresholds” for Ethical Research.

Sub-goal A: Automated Sentience Safeguards. Establishing “Software-Defined Red Lines” that automatically pause nutrient flow or bioprinting if the tool detects “Global Workspace” activity (a signature of consciousness).

Sub-goal B: Transparency and Public Visuals. Creating a “Community Bio-Audit” dashboard. This goal ensures that the public can “see” the status of high-complexity research, preventing a “black box” scenario where labs grow “brains” in secret.

3. Next, describe at least three different potential governance “actions” by considering the four aspects below (Purpose, Design, Assumptions, Risks of Failure & “Success”).

Action 1: Mandatory “Ethical Data Export” (Requirement/Rule)

Purpose: Currently, labs keep neural firing data private. I propose a rule where every funded organoid project must use a tool like Neuro-Visor to generate an “Ethical Compliance Report” (ECR) that logs neural complexity patterns. Design: Actor: Federal Regulators (NIH/NSF). They mandate the ECR for all grant renewals. I would build the Python scripts that generate these tamper-proof reports. Assumptions: We assume that “complexity” equals “moral status.” We might be wrong, since an organoid could be “silent” but still possess some form of awareness. Risks: Success creates a global standard for bioethics; Failure leads to “Data Faking,” where labs modify the Python scripts to hide high levels of activity.

Action 2: “Neural-Signature” Open-Source Database (Technical Strategy)

Purpose: Standardize the “language” of brain organoids across different hardware (C++ / Java backend). Design: Actor: Academic Researchers and Hardware Companies. They must opt-in to a standardized data format (like JSON for Bio-signals) so that all labs can compare their “Ethical Heatmaps” in real-time. Assumptions: That companies like Axion Biosystems will give up proprietary data formats for the sake of ethics. Risks: Success leads to a “Bio-Ethernet” for brains; Failure is that the data is too massive to process, leading to “lag” in ethical reporting.

Action 3: “Digital Twin” Donor Consent Vouchers (Incentive)

Purpose: Protect the privacy and autonomy of the person whose skin cells were used. Design: Actor: UI/UX Designers and Legal Tech Companies. Use SQL/PHP to create a decentralized ledger where donors can track how their “mini-brain” is being used and “withdraw” their cells if the research turns toward high-level sentience. Assumptions: We assume donors care about their “brain twin.” They might not care at all. Risks: Success empowers patients and builds trust; Failure creates a massive bureaucratic mess that slows down the cure for Alzheimer’s.

5. Last, drawing upon this scoring, describe which governance option, or combination of options, you would prioritize, and why. Outline any trade-offs you considered as well as assumptions and uncertainties. Based on the scoring and the technical nature of the Neuro-Visor, I recommend a Hybrid Governance Strategy that prioritizes a combination of Action 2 (The Open-Source Neural Database) and a streamlined version of Action 1 (The Mandatory Ethical Data Export).

The “Neural-Standard” Framework - I would present this recommendation to the National Institutes of Health (NIH) and the International Society for Stem Cell Research (ISSCR). These actors possess the funding and regulatory leverage to mandate standards across both academic and private sectors.

If you ask why, the primary goal is Safety through Transparency. By combining the standardized database with a mandatory reporting rule, we solve the “Black Box” problem. Action 2 provides the language (the standardized data format), and Action 1 provides the law (the requirement to speak it). Trade-offs

Implementing the Neuro-Visor as a mandatory tool will inevitably slow down research in the short term as labs adapt their hardware. However, the trade-off is avoiding a catastrophic ethical breach (like unintentionally creating a sentient entity) that could lead to a total federal ban on the field.

By requiring “Ethical Data Exports,” we force companies to reveal some of their neural firing data. The trade-off is designed to protect “proprietary logic” while exposing “safety-critical signals.”

The “Proxy” Assumption: We are assuming that electrical complexity (measured by the Neuro-Visor) is a reliable proxy for consciousness. This is a massive uncertainty—we could be measuring “noise” that looks like “thought,” or missing “thought” that doesn’t produce typical electrical spikes. The Hardware Gap: We assume all labs can afford or integrate the Neuro-Visor interface. If they cannot, we risk centralizing brain research only in wealthy institutions, harming Global Equity.

6. Reflecting on what you learned and did in class this week, outline any ethical concerns that arose, especially any that were new to you. Then propose any governance actions you think might be appropriate to address those issues. This should be included on your class page for this week. Reflecting on the class discussions and the HTGAA curriculum, I was challenged by the idea of Moral Status. Now I have a new ethical concern: “The Silent Sufferer.” If a brain organoid lacks a body or a “mouth,” it cannot express pain. In a UI/UX context, we usually design for “user feedback.” In bioengineering, the “user” (the organoid) cannot give feedback. This creates a terrifying design gap where the system could be in a state of “distress” without any visible output. To address this, I propose the following action for our HTGAA lab community:

Action: A “Universal Assembloid Safety Checklist” for the HTGAA Open-Source Repository.

Purpose: To move beyond broad consent and implement “Technical Red Lines” for every student project.

Design: Every organoid project submitted to the class page must include a “Neural Complexity Estimate” using the shared Python script we developed. If the score exceeds a specific threshold, the project requires an immediate peer-review “Ethical Huddle.”

Risk of Success: This fosters a “Safety Culture” early in our careers, making ethical checking as second nature as checking for DNA contamination.

Assignment (Week 2 Lecture Prep)

🧬 Questions from Professor Jacobson

1. Polymerase Error Rates & The Human Genome

• What is the error rate of polymerase? The raw error rate of DNA polymerase is approximately $10^{-3}$ to $10^{-6}$ per base pair before any correction mechanisms.
• How does this compare to the length of the human genome? The human genome is roughly 3 billion base pairs ($3 \times 10^9$ bp). Without correction, a single cell division would result in thousands of mutations, which would be lethal for genomic stability over generations.
• How does biology deal with that discrepancy? Biology uses a multi-layered repair system:
  1. Proofreading: Replicative polymerases have 3’ $\rightarrow$ 5’ exonuclease activity to remove mismatched bases immediately.
  2. Mismatch Repair (MMR): Specialized enzymes scan the DNA post-replication to fix errors that escaped proofreading.
  • Result: These mechanisms lower the final error rate to approximately $10^{-9}$ to $10^{-10}$ per base.

2. Coding for Human Proteins

• How many different ways are there to code for an average human protein? Because the genetic code is redundant (most amino acids are coded by multiple codons), the number of possible DNA sequences for an average protein (~400 amino acids) is astronomical—often exceeding $10^{150}$ possible combinations.
• Why don’t all these codes work in practice? * Codon Usage Bias: Some codons are translated more slowly than others depending on the organism’s tRNA availability.
  • RNA Secondary Structure: Certain DNA sequences produce mRNA that folds into shapes (like hairpins) that block the ribosome.
  • GC Content: High GC content can make the DNA physically difficult to “unzip” or amplify.
  • Unintended Sites: A sequence might accidentally create “cryptic” splice sites or start/stop signals in the middle of the gene.

🧪 Questions from Dr. LeProust

1. Most Common Oligo Synthesis Method

The industry standard for oligonucleotide synthesis is the phosphoramidite method. Companies like Twist Bioscience have scaled this by using silicon-based microarray platforms to synthesize thousands of sequences in parallel.

2. Difficulty of Synthesis Beyond 200nt

The primary barrier is stepwise yield. Even with a 99.5% efficiency rate per base addition:

1. the yield of a correct 200-mer is roughly $(0.995)^{200} , which is around 36%.
2. as the chain grows longer, the accumulation of “n-1” (shortened) products and chemical side reactions (like depurination) makes it nearly impossible to isolate a pure, full-length product.

3. Why Can’t You Make a 2000bp Gene Directly?

Direct synthesis of 2000bp is impossible because the yield would be effectively zero ($(0.995)^{2000}$). Instead, we synthesize short oligos (under 200nt) and then use Assembly Methods (such as PCA - Polymerase Cycling Assembly) to “stitch” them together into a full 2000bp gene.

🧬 Question from George Church

The “Lysine Contingency” and Essential Amino Acids

• The 10 Essential Amino Acids: Phenylalanine, Valine, Threonine, Tryptophan, Isoleucine, Methionine, Histidine, Arginine, Leucine, and Lysine. (Mnemonic: PVT TIM HALL)
• View of the “Lysine Contingency”: The “Lysine Contingency” suggests that an organism can be contained by making it unable to produce an essential nutrient.
• The Critique: Professor Church argues that traditional auxotrophy is “leaky” because organisms can often scavenge nutrients from their environment or other dying cells. To achieve true biocontainment, he proposes Synthetic Auxotrophy, where an organism is recoded to require a non-canonical (synthetic) amino acid that does not exist in nature, making it impossible for the organism to survive outside the lab.

Week 1 Lab: Pipetting

Individual Final Project

Group Final Project