Mycoplasma Protein Allocation Under Nutrient Limits

ISEF Category: Cellular and Molecular Biology

Subcategory: Other  ·  Difficulty: Advanced  ·  Setup: University Lab  ·  Time: Full Year

The Hook

A cell has a tiny budget, just like you do. If nutrients get scarce, the cell cannot make everything it wants, so it has to choose. That tradeoff can change growth, stress response, and survival. Your project can model those choices and compare them with real proteomics data.

What Is It?

This project asks how a very small cell, Mycoplasma genitalium, changes its protein-making plan when nutrients run low. Think of the cell like a factory with limited money, limited workers, and too many orders. Protein allocation means how the cell splits its resources across making enzymes, ribosomes, transporters, and repair proteins. A whole-cell model tries to track many of those moving parts at once.

Agent-based modeling means each part of the system follows rules, then the full behavior emerges from those rules. In this case, you would use a model that estimates how the cell rebalances resources under nutrient limitation. Public proteomics data help you check whether the model matches real protein abundance patterns. That gives you a way to test whether the model is just pretty math, or a useful prediction tool.

Why This Is a Good Topic

This is a strong science fair topic because you can make real predictions, test them against public data, and improve the model when it fails. The question connects to cell survival, metabolism, and synthetic biology, which matter in medicine and basic biology. You can learn model building, data cleaning, calibration, and statistical validation without needing to grow the organism yourself. The project also leaves room for a novel angle, since you can compare different nutrient limits or different allocation rules.

Research Questions

• How does nutrient limitation change predicted allocation to ribosomes versus stress-response proteins?
• What is the effect of changing one nutrient pool on simulated growth rate and protein distribution?
• Does adding a protein-budget constraint improve fit to public proteomics data?
• To what extent do different allocation rules change the model's predicted response to carbon versus nitrogen limitation?
• Which model parameter most strongly affects agreement between simulated and observed protein abundance?
• How does the calibrated model perform on proteomics datasets from different growth conditions?

Basic Materials

• Laptop or desktop computer with enough memory to run simulations.
• Spreadsheet software for organizing proteomics summaries and model outputs.
• Python installed with scientific libraries.
• Access to public proteomics datasets from PubMed-linked papers or PRIDE.
• Notes app or lab notebook for tracking assumptions, parameter choices, and model versions.
• Reference papers on Mycoplasma genitalium metabolism and whole-cell modeling.

Advanced Materials

• High-performance workstation or university cluster access.
• Python environment for simulation, data processing, and plotting.
• MATLAB or Mathematica if the model source code requires it.
• Public proteomics datasets and any available metabolic reconstructions.
• Statistical software for model comparison and sensitivity analysis.
• Version control system for tracking code changes and parameter sets.

Software & Tools

• Python: Runs simulation code, data cleaning scripts, and custom analysis.
• Jupyter Notebook: Helps you document model choices, run code, and show results in one place.
• pandas: Organizes proteomics tables and model output into searchable data frames.
• NumPy: Handles numerical calculations for simulation and parameter testing.
• Matplotlib: Makes plots that compare predicted and observed protein allocation.

Experiment Steps

1. Define the biological question you want the model to answer, and pick one nutrient limitation to test first.
2. Map the model inputs, outputs, and assumptions so you know which parts of the cell you are treating as fixed, and which parts can change.
3. Gather public proteomics datasets that match your chosen growth conditions, and decide how you will clean and standardize them.
4. Build a calibration plan that separates training data from validation data, so you can test the model on conditions it has not seen.
5. Choose comparison metrics that measure both protein-level fit and system-level behavior, such as growth predictions or allocation shifts.
6. Plan a sensitivity analysis to find which parameters most affect the model's predictions, then decide how you will report uncertainty.

Common Pitfalls

• Trying to fit every protein at once, which makes the model impossible to interpret.
• Mixing proteomics datasets with different normalization methods, which can create fake differences between conditions.
• Treating missing proteins as true zeros, which can distort calibration and validation.
• Changing too many parameters at the same time, which hides which assumption caused the result.
• Comparing model output to the wrong biological condition, which makes the validation look better or worse than it really is.

What Makes This Competitive

A competitive version of this project would do more than copy a published model. You would test a clear hypothesis, compare multiple allocation rules, and show which one predicts public data best. Strong entries also quantify uncertainty, run sensitivity analysis, and validate on a dataset the model did not use for calibration. If you can connect the model to a biologically meaningful tradeoff, your project feels like real research instead of a simulation demo.

Project Variations

• Compare nutrient limitation effects in two different public proteomics datasets from separate labs.
• Test whether a simpler allocation rule can match the full model under one stress condition.
• Rebuild the analysis around a different output, such as predicted growth rate, stress proteins, or transporter allocation.

Learn More

• PubMed: Search for review articles and original papers on Mycoplasma genitalium, whole-cell models, and proteomics calibration.
• PRIDE Archive: Find public proteomics datasets that you can compare against model predictions.
• NCBI Bookshelf: Read accessible background chapters on molecular biology, metabolism, and systems biology.
• MIT OpenCourseWare: Look for free systems biology or computational biology course materials with modeling examples.
• NIH and NCBI resources: Search for gene, protein, and pathway background information tied to your organism and nutrients of interest.
• Cell Systems: Search for peer-reviewed papers on whole-cell modeling, protein allocation, and constraint-based analysis.