Week 1 HW: Principles and Practices

A Living, Programmable Menstrual Health Diagnostic Cup

I’ve always been in awe of looking at the intricate patterns of the biological world through a mathematical lens. It makes me pause for a moment—wondering how evolution has managed to create systems that are so exquisitely ordered, yet profoundly chaotic at the same time. Finding patterns in this overwhelming biological diversity makes me think like an engineer: trying to understand the rules underneath, and—if I’m lucky—optimize life itself.

But this project did not begin purely as an intellectual curiosity. It began with pain.

I have endometriosis. And anyone who has lived with it knows that it is not just “bad periods.” It is pain that interrupts life, pain that is normalized, dismissed, and often endured in silence for years. Despite affecting millions, endometriosis still takes an average of 7–10 years to be diagnosed, often only after irreversible damage has already occurred. I found myself asking a question that was both deeply personal & scientific: Why do we wait for disease to become visible before we believe it exists?

This question became the foundation of the biological engineering tool I want to develop.

Why this project only?

Endometriosis, especially in its early stages, is not primarily a structural disease. Long before lesions are visible on imaging or during surgery, the body already exhibits immune dysregulation, altered inflammatory signaling, and estrogen-driven changes (and these are not commone to endometriosis alone and thus could be extrapolated to find out early signals of various of diseases). These early biological states are real—but they are invisible to our current diagnostic tools.

At the same time, one biological resource remains largely ignored: menstrual fluid.

Menstrual blood is not waste. It is a rich biological mixture containing immune cells, cytokines, prostaglandins, and hormone-responsive molecular signals. In many ways, it represents a monthly immune–endocrine snapshot of the uterus, and by extension, reproductive and systemic health. Yet because this information is complex, dynamic, and noisy, we have largely chosen to ignore it rather than learn how to interpret it.

As someone who experiences menstruation not just as a biological process but as a recurring reminder of disease, I began to see menstruation differently not as something to be endured, but as information waiting to be decoded.

The Core Idea

I propose to develop a menstrual-cup–based living diagnostic system that functions as a monthly health card for females. The system contains a sealed compartment within the menstrual cup that houses encapsulated engineered living cells.

These cells:

• Are not implanted in the body
• Never directly interact with human tissue
• Only respond to molecules diffusing from menstrual fluid

Rather than measuring single biomarkers, these engineered cells are programmed—using synthetic biology principles—to sense, integrate, and interpret biological signals present in menstrual fluid (synthetic biosensors).

In its initial form, the system is designed to detect early-stage endometriosis, by responding to a biologically meaningful combination of:

• inflammatory signals
• estrogen-responsive cues

Only when this specific immune–endocrine pattern is present does the system produce an interpretable output, such as a visible color change or fluorescence. This allows the detection of early disease states, not just late-stage anatomical damage.

Beyond Endometriosis: A Monthly Health Card

While endometriosis serves as the starting point, the broader vision is to create a programmable platform. By altering the genetic logic circuits inside the cells without changing the menstrual cup itself the system could be adapted to detect:

• other inflammatory reproductive disorders
• endocrine dysregulation
• chronic low-grade systemic inflammation
• immune-skewed health states

Over time, this transforms menstruation into something radically different: a longitudinal, non-invasive health report, generated monthly by the body itself.

Why This Matters

For me, this project is not just about solving a scientific problem it is about rewriting my story and the story of women’s health. A story where pain is believed earlier, where biology is interpreted rather than ignored, and where complexity is not simplified away, but thoughtfully engineered around.

I’ve always been fascinated by stories not just emotional ones, but scientific ones because stories move people. The hardest scientific problems, much like the hardest human ones, often become solvable once we understand the story correctly. Through this work, I’m trying to tell a story - my story, maybe your story and most definitely the story of women’s health.

(P.S. I’m still a novice—learning, questioning, and always open to being corrected. This project is as much a journey as it is an idea.)

2. Governance and Policy Goals for an Ethical Menstrual Health Diagnostic Tool

Goal 1: Ensuring Safety and Preventing Harm (Non-Malfeasance)

The first and most basic governance goal is to ensure that the tool does not physically or biologically harm users.

Sub-goal 1.1 : Strong Biological Containment Since the system uses engineered living cells, it is important that these cells are fully contained and never enter the body. The cells should be physically encapsulated within a sealed compartment of the menstrual cup and designed so that they cannot survive outside this environment. This reduces the risk of accidental exposure or environmental release.

Sub-goal 1.2: Restricting Function to Sensing Only The engineered cells should be limited strictly to sensing and reporting biological states. They should not be capable of modifying menstrual fluid, producing therapeutic molecules, or interacting with human cells. Keeping the system purely interpretive helps prevent unintended biological effects and keeps the tool within a safer diagnostic boundary.

Sub-goal 1.3: Avoiding Overclaiming and Misuse From a policy perspective, it is important that this tool is not marketed or interpreted as a definitive diagnostic device. It should be clearly framed as an early-risk or health-state indicator that complements, rather than replaces, professional medical evaluation. This helps prevent both medical and psychological harm from misinterpretation.

Goal 2: Preserving User Autonomy

Sub-goal 2.1: Informed Consent Users should understand what information the system provides, its limitations, and how often data is generated, especially since the tool is used repeatedly over time.

Sub-goal 2.2: Data Ownership Any health information generated should belong to the user by default, with no automatic sharing with third parties without explicit consent.

Sub-goal 2.3: Right Not to Know Users should be able to choose which health states they wish to monitor, recognizing that receiving certain information can have psychological consequences.

Goal 3: Promoting Equity and Avoiding Harmful Norms

Sub-goal 3.1: Accessibility The system should be affordable and compatible with existing menstrual products to avoid becoming accessible only to a privileged population (Like all others this is very important to me coming from a developing country like India).

Sub-goal 3.2: Avoiding Pathologization Governance should ensure that normal menstrual variability is not automatically labeled as abnormal, which could reinforce stigma around menstruation.

3. Potential Governance Actions

Action 1: Regulatory Requirement for Containment and Functional Limits in Living Diagnostics

Purpose: Currently, most biosafety regulations focus on therapeutic or environmental uses of engineered cells. I propose a specific governance requirement for living diagnostic devices that mandates strict physical containment and limits engineered cells to sensing and reporting functions only.

Design: Regulators and institutional review boards would require proof of encapsulation, biological kill-switches, and non-therapeutic function before approval. Academic researchers and companies developing such systems would need to comply during biosafety and ethics review processes.

Assumptions: This assumes that containment strategies are reliable and that living diagnostic tools can be clearly distinguished from therapeutic applications.

Risks of Failure & “Success”: Overregulation could slow early research. However, if successful, this action could normalize safe, contained use of living cells in diagnostics, potentially accelerating public acceptance without sufficient understanding of system limitations.

Action 2: User-Owned Data as a Default Standard for Menstrual Health Technologies

Purpose: Menstrual and reproductive health data is highly sensitive, yet current digital health tools often retain ownership or control over user data. I propose a policy standard where all data generated by menstrual health diagnostics is owned by the user by default.

Design: Companies would be required to design systems that store data locally or share it only through explicit opt-in consent. Policymakers could support this through consumer data protection rules, similar to frameworks used in financial or digital privacy systems.

Assumptions: This assumes users value data ownership and that companies can still innovate without extensive data aggregation.

Risks of Failure & “Success”: Restricting data access could limit large-scale research and population-level insights. If successful, however, this approach could set an ethical precedent for handling intimate biological data across women’s health technologies.

4. Governance actions against rubric of Policy Goals.

Option 1: Regulatory requirement for containment & functional limits in living diagnostics

Option 2: User-owned data as the default standard for menstrual health technologies

Option 3: Clinical framing & education requirement

5. Prioritization and Governance Recommendation

I would prioritize a layered governance strategy, implementing different actions at different stages of technological development.

Primary priority: Regulatory requirements for containment and functional limits (Option 1). At early stages, the greatest risk comes from misuse, accidental exposure, or public mistrust of living diagnostics. Strict containment, kill-switches, and limits to sensing-only functions establish a clear safety boundary and help normalize the ethical use of engineered cells in diagnostics. Although this may slow early research, predictable and principled rules can ultimately accelerate adoption by increasing trust and regulatory clarity.

Secondary priority: User-owned data as a default standard (Option 2). As the technology scales, governance must shift toward protecting autonomy and preventing data exploitation. Default user ownership of menstrual health data preserves agency and dignity, even if it limits large centralized datasets. Voluntary, consent-based data sharing can still support research without extractive practices.

Trade-offs and uncertainties include distinguishing diagnostic from therapeutic systems, potential overregulation, and reduced access to population-level data. Despite these, prioritizing safety first and autonomy second offers the strongest path toward an ethical, trusted, and socially legitimate menstrual health diagnostic platform.

Week 2 Lecture Prep

Homework Questions from Professor Jacobson

Nature’s machinery for copying DNA is called polymerase. What is the error rate of polymerase? How does this compare to the length of the human genome. How does biology deal with that discrepancy?

Solution: Error rate of DNA polymerase in DNA Polymerase is ~1 error per 10^5 nucleotides during raw DNA synthesis. When compared to human genome (3 × 10⁹ base pairs) this accounts to 30,000 mutations per replication. This discrepancy is handled by

1. Proofreading activity : This is done by exonuclease (3′→5′). This improves fidelity by 100x
2. Mismatch repair (post-replication) : Detects mismatches that escape proofreading. It uses strand discrimination to fix the new strand and improves fidelity by another ~100–1000×.

How many different ways are there to code (DNA nucleotide code) for an average human protein? In practice what are some of the reasons that all of these different codes don’t work to code for the protein of interest?

How many different DNA sequences can encode an average human protein?

An average human protein is ~400 amino acids long. Because the genetic code is degenerate, each amino acid is encoded by ~3 synonymous codons on average. Thus, the number of possible DNA sequences encoding the same protein is approximately:

[ 3^{400} ]

Although many sequences are theoretically possible, biological constraints limit functionality:

• Codon usage bias: preferred codons match abundant tRNAs; rare codons reduce translation efficiency
• Translation speed & folding: codon choice affects co-translational protein folding
• mRNA structure: some sequences form inhibitory secondary structures
• Regulatory signals in coding DNA: codon changes can disrupt splicing or RNA-binding motifs
• GC content & cryptic signals: extreme GC or hidden stop/splice signals impair expression

Homework Questions from Dr. LeProust

1. What is the most commonly used method for oligonucleotide synthesis currently?

The most commonly used method for oligonucleotide synthesis is phosphoramidite-based solid-phase DNA synthesis. In this method, DNA is synthesized stepwise on a solid support by the sequential addition of protected nucleotide phosphoramidites. Each synthesis cycle consists of deprotection, coupling, capping of unreacted chains, and oxidation. This method is highly automated, reproducible, and provides high coupling efficiencies (~99–99.5% per nucleotide), making it suitable for routine synthesis of short DNA oligonucleotides.

2. Why is it difficult to make oligonucleotides longer than ~200 nucleotides by direct synthesis?

Direct synthesis of long oligonucleotides is limited due to cumulative error and yield loss during sequential nucleotide addition. Although each coupling step is highly efficient, the probability of obtaining a full-length correct product decreases exponentially with increasing length. For example, with a coupling efficiency of 99.5%, only about 37% of molecules remain full-length after 200 synthesis cycles. In addition, side reactions such as incomplete coupling, depurination, and truncation accumulate, making purification increasingly difficult and reducing the overall fidelity and yield of long oligos.

3. Why can’t a 2000 bp gene be synthesized directly using oligo synthesis?

Synthesizing a 2000 bp gene would require approximately 2000 sequential chemical coupling steps. Even with high coupling efficiency, the cumulative probability of producing a full-length, error-free product becomes negligibly small (on the order of 10⁻⁵). Furthermore, the high frequency of insertions, deletions, and truncations makes purification impractical and the error rate biologically unacceptable. Therefore, long genes are not synthesized directly but are instead constructed by assembling shorter, chemically synthesized oligonucleotides using enzymatic methods such as PCR-based assembly or Gibson assembly.

Homework Question from George Church

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

Animals cannot synthesize the following amino acids in sufficient amounts and must obtain them from the diet:

1. Histidine
2. Isoleucine
3. Leucine
4. Lysine
5. Methionine
6. Phenylalanine
7. Threonine
8. Tryptophan
9. Valine
10. Arginine

Implications for the “Lysine Contingency”

The fact that lysine is essential in all animals implies that animals have irreversibly lost the biosynthetic pathways required to produce it. This supports the idea of the lysine contingency: once early animals evolved in lysine-rich environments (e.g., feeding on plants or microbes that could synthesize lysine), there was no selective pressure to retain the complex and energetically costly lysine biosynthesis pathway. Over evolutionary time, this loss became locked in.

As a result, lysine availability became a nutritional dependency rather than a metabolic choice, shaping animal diets and trophic relationships. The lysine contingency highlights how evolution can constrain future possibilities by eliminating biosynthetic options that are no longer immediately necessary.

What code would you suggest for AA:AA interactions? could not understand this question