Week 1 HW: Principles and Practices

DNA Based Human Memory Storage and Governance

I want to develop a DNA based memory storage system that encodes key aspects of a person’s memories or life history into synthetic DNA, with the intent that these memories could be stored long term and potentially be made inheritable across generations. DNA is already being explored as a long term medium for digital information, with approaches using synthetic DNA and enzymatic DNA synthesis for data archiving. In my concept, instead of just archiving arbitrary digital files, the system would encode structured “memory blocks” (e.g., life events, medical history, personal preferences, etc) into DNA segments that are linked, similar to a blockchain, where each new “block” references the previous one for integrity and auditability.

In practice, the system would likely remain ex-vivo (e.g., DNA stored in biobanks or on chips), but it raises the possibility of integrating such synthetic “memory DNA” into germline cells or early embryos so that descendants inherit both genomic DNA and an attached “memory” of ancestral memory data. Existing debates around DNA banking, biobanks, and long term genomic data storage already highlight ethical questions about consent, control, and ownership of inherited DNA information. Extending this to intentional, inheritable memory storage amplifies those questions and forces us to think about intergenerational governance instead of just individual autonomy.

1. Governance and policy goals

High level goal: Ensure that DNA based memory storage and inheritance promotes autonomy, privacy, and constructive uses, while preventing harms related to discrimination, coercion, and injustice.

Specific sub goals

• G1: Protect individual and family autonomy over memory DNA.

  • G1a: Require explicit, informed consent for both creation and use of memory DNA, including clear options about whether any part is inheritable, reflecting current debates about consent and posthumous genomic data.
  • G1b: Recognize shared interests so that close relatives have some standing in decisions about inheritable memory DNA that could affect them, as argued in work on family interests in genomic data.

• G2: Prevent misuse, discrimination, and surveillance.

  • G2a: Prohibit use of memory DNA in law enforcement, insurance underwriting, or employment decisions, echoing current concerns about DNA databases and genomic EMR misuse.
  • G2b: Require strong technical safeguards (encryption, access control, logging) similar to emerging blockchain/genomic platforms so that memory DNA cannot be silently copied, de anonymized, or exploited.

• G3: Ensure safety and limit biological risks.

  • G3a: Restrict or ban integration of memory DNA into germline cells or embryos until there is ample evidence of biological safety and societal consensus, building on existing caution around germline modification and DNA banking.
  • G3b: Define environmental safeguards for any physical storage of DNA encoded memories so accidental release does not create ecological or biosafety concerns, paralleling governance discussions around biobanks and synthetic DNA storage.

• G4: Promote equitable, beneficial uses.

  • G4a: Avoid creating a “memory elite” where only wealthy groups can preserve detailed ancestral memory records, addressing concerns about inequities in genomic databases and data access.
  • G4b: Encourage uses that clearly benefit individuals and communities (e.g., family medical history, cultural archiving) under owner governed, privacy preserving frameworks inspired by emerging blockchain based genomic systems.

2. Governance actions

Option 1: Memory DNA consent and family governance framework

Purpose
Currently, genomic data governance mostly treats DNA as individual health or research data, with consent policies focused on the individual patient or research participant. For inheritable memory DNA, a new consent and governance framework that explicitly recognizes shared, intergenerational interests and defines when and how family level input is needed for creation, inheritance, and access.

Design

• Standardized multi layer consent forms specifying whether memory DNA can be created, whether it can be integrated into the germline or only somatically stored, and who can access it and under what conditions.
• A “family governance” procedure in which a minimal set of affected parties (e.g., co parents, existing children above a certain age) must be informed and given limited rights to object or request constraints.
• Implementation by clinical genetics services, IVF clinics, and certified DNA storage providers, overseen by national bioethics committees, data protection authorities, and IRBs.

Assumptions

• Families can practically be consulted and processes will not become unmanageably complex.
• A workable balance can be struck between individual autonomy and family interests, as suggested in arguments that individuals do not have unlimited rights over family relevant genomic information.
• Regulators and clinics have resources to implement nuanced consent and dispute resolution procedures

Risks of failure and “success”

• Failure: Process becomes too complex, leading to superficial consent or “checkbox” compliance; some family members may be excluded or feel coerced; conflicts over who counts as family.
• Success: Widespread family governance could unintentionally erode individual privacy by normalizing expectations that relatives can demand access to memory DNA.

Option 2: Technical owner governed memory DNA platform (blockchain + encryption)

Purpose
Genomic and DNA data are often stored in centralized databases, with access controlled by institutions, raising privacy, security, and trust concerns. An owner governed technical infrastructure for memory DNA, using encryption, secure computation, and blockchain based audit trails so individuals retain primary control and can monitor every access is needed.

Design

• Memory DNA stored in encrypted form, with decryption keys held by the individual (and possibly an emergency escrow under strict legal safeguards).
• A permissioned blockchain records all requests and uses, similar to blockchain based platforms for genomic data sharing that log owner controlled transactions.
• Smart contract–like rules define who can query, for what purposes, and what outputs are allowed.
• Use of privacy preserving computation (e.g., homomorphic encryption, zero knowledge proofs) to run computations on encrypted data without exposing raw memory DNA.
• Built by specialized companies and research consortia, with regulators setting minimum security and audit standards and professional bodies certifying providers.

Assumptions

• Cryptographic and blockchain-like technologies can scale to the volume and complexity of memory DNA queries without prohibitive cost or latency.
• Users and clinicians can manage keys and permissions effectively or can rely on usable interfaces and safe defaults.
• Permissioned blockchains and technical standards can be interoperable across borders and institutions.

Risks of failure and “success”

• Failure: Poor implementation could cause key loss, smart contract vulnerabilities, or overconfidence that “blockchain = perfectly secure.”
• Success: If widely adopted, this infrastructure could become critical and centralize power in a few vendors, and the ability to fine tune permissions might encourage commodification and market trading of memory DNA.

Option 3: Legal and regulatory restrictions on use and germline integration

Purpose
There is already concern that DNA databases and genomic data can be repurposed for law enforcement, immigration, or commercial profiling beyond their original intent. For memory DNA, explicit legal restrictions are needed: ban certain uses (e.g., subpoenas, underwriting, employment decisions) and set strict conditions or moratoria on germline integration.

Design

• National legislation that:
  • Treats memory DNA as highly sensitive personal data (PHI, PII), with protections at least as strong as medical records or genomic data in EMRs.
  • Prohibits compelled disclosure of memory DNA in criminal, civil, or immigration proceedings, similar to reforms debated for law enforcement DNA databases.
  • Bans use of memory DNA in insurance underwriting, employment screening, and credit scoring to reduce discrimination risks.
  • Establishes a moratorium (or narrow clinical exception regime) on germline incorporation of memory DNA, with periodic review by independent bioethics bodies.
• International guidelines (e.g., WHO/UNESCO like) to harmonize basic protections across borders, building on existing guidance for DNA banking and genomic research.

Assumptions

• Legislators are willing and able to understand the technology and pass targeted laws in time, rather than react only after abuses occur.
• International coordination is feasible and loopholes (e.g., jurisdiction shopping) can be limited.
• Enforcement agencies and courts will respect the special status of memory DNA, rather than gradually expanding exceptions.

Risks of failure and “success”

• Failure: Laws might be too vague or narrow, leaving room for secondary uses, or they could be poorly enforced, especially in settings with weaker rule of law.
• Success: Overly strict bans could chill beneficial research and clinical innovations that responsibly use memory DNA, echoing concerns that heavy handed regulation can impede constructive genomic applications.

3. Scoring the options (rubric)

Using 1–3 where 1 = best, 3 = worst, n/a = not applicable.

Option 1 scores well on autonomy and constructive use but adds procedural complexity and only indirectly affects biosecurity and environmental risks. Option 2 is strong on privacy, auditability, and enabling controlled constructive use, but it is technically demanding and costly, which hurts feasibility and burden scores. Option 3 is strongest at preventing harmful uses and some biosecurity/environmental risks but, if drafted too broadly, could hinder research and clinical innovation.

4. Recommended governance mix and audience

Prioritization of a hybrid approach that combines the technical owner governed infrastructure (Option 2) as the backbone, backed by clear legal restrictions (Option 3) on unacceptable uses and germline integration, and context sensitive consent/family governance (Option 1) for inheritable configurations. Option 2 is central because memory DNA will only remain ethically acceptable if individuals and, where appropriate, their families retain meaningful, technically enforced control over access and usage, including the ability to monitor every transaction and revoke permissions. However, technical controls alone are insufficient, so legal frameworks are needed to prohibit uses fundamentally incompatible with autonomy and justice, such as law enforcement or insurance exploitation and unregulated germline use.

For any form of inheritable memory DNA, Option 1’s governance model helps resolve the tension between individual and family interests that already appears in debates on posthumous genomic data. Tthis recommendation is directed primarily to national regulators and bioethics councils (e.g., national health departments and data protection authorities) and secondarily to large academic medical centers and biotech consortia likely to build early memory DNA platforms. Regulators can set baseline legal rules and mandate core technical protections, while clinical and research institutions develop and test practical consent and family governance procedures.

5. Ethical concerns and additional governance ideas

Intergenerational autonomy vs. inherited data is a central concern, since existing work already shows genomic data challenges purely individual consent models and memory DNA would encode not just risk but ancestors’ personal narratives. One additional governance idea is to create “sunset” rules so that certain categories of memory DNA automatically expire or require re affirmation by each new generation before use, echoing periodic review of stored genomic samples.

Privacy and surveillance creep are also major worries, since controversies around DNA databases show that once large repositories exist, there is pressure to repurpose them for law enforcement, immigration control, or commercial profiling. A governance response would be to classify memory DNA as a special category of data that governments and corporations cannot collect or compel without extremely narrow, transparent exceptions, and to require public reporting and independent oversight for any large scale repositories.

Finally, psychological and social impacts may be profound because high fidelity records of ancestors’ memories could change family dynamics, expectations, and identity, similar to how genomic information in EMRs can alter self perception and relationships. A reasonable governance action would be to require integrated psychosocial counseling and ethics review for any clinical or research deployment of inheritable memory DNA, following genetic counseling standards for high stakes genomic tests.

Overall, governance for DNA memory “blockchains” must explicitly address intergenerational justice, psychological well being, and the politics of whose memories are preserved in biological form, not just privacy and safety checklists.

Jacobson – Question 1

What is the error rate of polymerase? How does this compare to the human genome, and how does biology handle it?

DNA polymerase (with proofreading and repair) makes about 1 error in 10^{6 bases.
The human genome is ~3×10}9 bp, so biology relies on proofreading and mismatch‑repair systems to keep the actual number of permanent mutations very low.

Jacobson – Question 2

How many ways to code an average human protein, and why don’t they all work?

There are astronomically many different DNA sequences that could encode a typical human protein (because of synonymous codons).
Many of these don’t work well because of bad codon usage, extreme GC content, strong mRNA secondary structures, repeats/homopolymers, or unwanted regulatory motifs that hurt synthesis or expression.

LeProust – Question 1

Most commonly used method for oligo synthesis?

Solid‑phase phosphoramidite chemical synthesis on a solid support is the standard method.

LeProust – Question 2

Why is it hard to make oligos >200 nt by direct synthesis?

Each chemical addition isn’t perfect, so errors and truncations accumulate. Beyond ~150–200 nt, the fraction of correct full‑length product becomes very small.

LeProust – Question 3

Why can’t you make a 2000 bp gene directly as one oligo?

A 2000 nt oligo would go through so many imperfect cycles that essentially no full‑length, error‑free strands would be produced. Instead, long genes are assembled from many shorter oligos or fragments.

Church – Question 1

What are the 10 essential amino acids in all animals and how does this affect your view of the “Lysine Contingency”?

The 10 essential amino acids for most animals are histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, valine, and arginine (though arginine is only conditionally essential in adult humans, making the strict count 9). In Jurassic Park, Dr Wu engineered the dinosaurs to be unable to produce lysine — a failsafe called the “Lysine Contingency.” Without supplemental lysine in their feed, the animals would die. This is based on real science since no animal can synthesize lysine. The flaw? Lysine is abundant in animal protein, so carnivores like T. rex and Velociraptors would easily get it from prey. Herbivores would have a harder time since lysine is the most limiting amino acid in grains and many plants, though legumes and certain vegetation do contain it. Either way, as Malcolm put it, “life finds a way.”

Google: “What are the 10 essential amino acids in all animals” and “can animals synthesize lysine”. I also recently started watching Jurrasic Park

Week 2 HW: DNA Read, Write, & Edit

Working on catching up after being sick this past week. Apologies for the delay.

Week 3 HW: Lab Automation

1. Published Paper Using Opentrons for Novel Biological Applications

Paper Reference: “Semiautomated Production of Cell-Free Biosensors” (ACS Synthetic Biology, 2024)

Summary: This study demonstrates the use of the Opentrons OT-2 liquid handling platform to manufacture cell-free biosensor reactions with improved consistency and throughput compared to manual assembly. The researchers compared manual vs. semiautomated approaches for assembling fluoride-sensing biosensors, constructing an entire 384-well plate of reactions. The automation significantly reduced quality control issues and performance variability that typically plague manual biosensor assembly. The study validated that automated liquid handling could achieve detection outcomes close to expected performance while dramatically improving reproducibility and scaling potential.

Key Innovation: The paper demonstrates that automation can overcome major bottlenecks in biosensor development - specifically the variability and time constraints of manual liquid handling that limit both throughput and consistency in biosensor characterization.

2. Final Project Description: Automated PFAS Biosensor Development and Optimization

Project Goal

Develop and optimize engineered bacterial biosensors for detecting PFAS (per- and polyfluoroalkyl substances) in environmental water samples, with specific focus on Michigan’s contaminated waterways. The project will use automation to systematically optimize biosensor performance and establish reliable detection protocols.

Automation Workflow Overview

Phase 1: Biosensor Library Construction

• Automated transformation of PFAS-responsive genetic circuits into E. coli chassis
• Systematic construction of promoter-reporter combinations
• Automated culture propagation and glycerol stock preparation

Phase 2: Screening and Optimization

# Pseudocode for PFAS concentration-response screening
def pfas_screening_protocol():
    # Prepare biosensor cultures
    for strain in biosensor_library:
        protocol.pick_colony(strain, growth_plate)
        protocol.inoculate(LB_media, overnight=True)
    
    # Set up concentration series
    pfas_concentrations = [0, 0.1, 1, 10, 100, 1000]  # ng/L
    
    # Automated screening
    for strain in biosensor_library:
        for pfas_conc in pfas_concentrations:
            well = protocol.get_next_well()
            protocol.transfer(strain_culture, well, 100_ul)
            protocol.transfer(pfas_solution[pfas_conc], well, 10_ul)
            protocol.mix(well, cycles=3)
    
    # Automated monitoring
    for timepoint in [1, 2, 4, 8, 12, 24]:  # hours
        protocol.read_plate(absorbance_595nm)
        protocol.read_plate(fluorescence_gfp)
        protocol.log_data(timepoint, well_id, strain, pfas_conc, signal)

    return optimization_data

Phase 3: Environmental Sample Testing

• Automated serial dilution of water samples
• Spiked recovery experiments for validation
• Cross-reactivity testing with other environmental contaminants

Hardware Configuration

• Primary Platform: Opentrons Flex with enhanced deck space for multiple plate types
• Modules:
  • Heater-shaker for bacterial culture incubation
  • Plate reader for automated absorbance/fluorescence measurement
  • Magnetic module for cell washing steps
  • Temperature module for reagent stability
• Custom Labware: 3D-printed holders for environmental sample vials

Expected Automation Benefits

1. Reproducibility: Eliminate manual pipetting variability in multi-step protocols
2. Throughput: Screen 10-20 biosensor variants against 6-8 PFAS concentrations simultaneously
3. Time Efficiency: Reduce hands-on time from ~8 hours to ~1 hour per screening round
4. Data Quality: Automated timing ensures consistent incubation periods and measurement intervals
5. Scalability: Protocol easily adaptable for testing additional PFAS compounds or environmental matrices

Integration with Cloud Laboratory (Future Extension)

• Use Ginkgo Nebula for automated design-build-test cycles
• Automated DNA synthesis and cloning of optimized biosensor variants
• High-throughput screening of next-generation biosensor designs
• Machine learning integration for predictive biosensor optimization

Validation Metrics

• Detection limit (target: <10 ng/L PFOA/PFOS)
• Response time (target: <4 hours for field deployment)
• Specificity (minimal cross-reactivity with structural analogs)
• Stability (consistent response over 48-72 hours)

Real-World Application

The optimized biosensors would ultimately be incorporated into portable detection systems for field deployment across Michigan’s water monitoring network, enabling rapid screening that complements traditional analytical chemistry methods.

This automation-centered approach addresses the key challenge in biosensor development: the need for systematic, reproducible optimization across multiple variables (strain, conditions, analyte concentrations) while maintaining the throughput necessary for meaningful statistical analysis.

Week 1 Lab: Pipetting

Individual Final Project

Group Final Project