Shared Cytokine Storm Hub Target Discovery

ISEF Category: Cellular and Molecular Biology

Subcategory: Cellular Immunology  ·  Difficulty: Advanced  ·  Setup: University Lab  ·  Time: Full Year

The Hook

A cytokine storm can turn your immune system from protector to threat. One virus can trigger it, and another can trigger a similar crash. That makes the pattern worth tracking, because the same control point might matter across more than one infection. You can use public data to look for that shared switch.

What Is It?

A cytokine storm is an overactive immune response. Cytokines are small signaling proteins that cells use to talk to each other. When that signaling gets too loud, the body can damage its own tissues. Think of it like a fire alarm that keeps ringing after the fire is out.

This project asks whether different infections, like influenza, SARS-CoV-2, and dengue, leave behind a shared gene expression pattern. Bulk RNA-seq gives you a snapshot of which genes are turned up or down in a sample. Network analysis helps you find genes that sit near the center of that pattern, often called hub genes. A hub transcription factor is a gene regulator that can influence many other genes, which makes it a strong candidate for further study.

AlphaFold can help you look at the protein shape that the transcription factor might fold into. Druggability scoring asks whether a protein has pockets or features that could bind a drug-like molecule. You are not proving a treatment works. You are building a data-driven shortlist of targets that might be worth deeper study.

Why This Is a Good Topic

This is a strong science fair topic because you can ask a clear question, use public data, and compare multiple pathogens in a structured way. The project connects to a real medical problem, severe immune overreaction, and to the search for broad antiviral or anti-inflammatory targets. You can learn RNA-seq analysis, gene network thinking, and basic protein structure reasoning without needing to run a wet lab from scratch.

Research Questions

• How does the gene expression network for cytokine-storm samples differ across influenza, SARS-CoV-2, and dengue infections?
• What is the effect of pathogen type on the identity of the top hub transcription factors in cytokine-storm signatures?
• Does combining multiple public RNA-seq datasets increase the overlap of shared immune-response genes across pathogens?
• To what extent do candidate hub transcription factors show predicted druggable pockets in AlphaFold-based structural models?
• Which network centrality measure best identifies shared hub genes across the three pathogen groups?
• How does excluding low-quality or small datasets change the list of shared candidate targets?

Basic Materials

• Laptop or desktop computer with enough memory for large data files.
• Reliable internet access for downloading public RNA-seq datasets and annotations.
• Spreadsheet software for sample tracking and cleanup.
• R or Python installed for data analysis.
• Bioconductor packages or Python libraries for RNA-seq normalization and visualization.
• Access to PubMed for background reading and gene validation.
• NIH GEO database or ArrayExpress for public bulk-RNA-seq datasets.

Advanced Materials

• High-memory workstation or university cluster account for large-scale transcriptomic analysis.
• RStudio or Jupyter Notebook for reproducible analysis workflows.
• Bioconductor packages for differential expression and enrichment analysis.
• Cytoscape for gene network visualization and hub ranking.
• AlphaFold structure files from the Protein Data Bank or AlphaFold DB.
• Structure analysis tools for pocket inspection and druggability scoring.
• Pathway databases such as Reactome, KEGG, or MSigDB for enrichment comparison.

Software & Tools

• R: Runs differential expression, correlation, and network analysis on public RNA-seq data.
• Python: Handles data cleaning, plotting, and automated comparison across pathogen groups.
• Cytoscape: Visualizes gene interaction networks and highlights hub transcription factors.
• ImageJ: Measures protein structure images or exported figure annotations if you need image-based comparisons.
• AlphaFold DB: Provides predicted protein structures that you can inspect for druggable features.

Experiment Steps

1. Define the exact immune phenotype you will study, then decide which public datasets count as cytokine-storm cases and which count as controls.
2. Choose one consistent RNA-seq processing path, then plan how you will normalize datasets from different studies before comparing them.
3. Build the gene lists that represent each pathogen group, then decide which overlap test or enrichment test will answer your question best.
4. Construct a network model, then select the centrality measure that will rank possible hub transcription factors in a fair way.
5. Screen the top hub candidates with structural evidence, then decide what counts as enough support for a protein to look druggable.
6. Pre-plan validation checks, then test whether your result stays the same after removing one dataset at a time or changing the ranking method.

Common Pitfalls

• Mixing datasets from different tissues or cell types, which can make pathogen effects look like immune-storm effects.
• Comparing raw counts across studies without normalization, which can create fake differences.
• Calling the highest-expression gene a hub without checking network centrality, which confuses abundance with influence.
• Using too many candidate genes at once, which makes the project hard to interpret and weakens the final target list.
• Skipping quality checks on public RNA-seq metadata, which can leave you with mislabeled severity groups or controls.

What Makes This Competitive

A strong version of this project does more than list shared genes. You would test whether the same hub survives different datasets, different centrality methods, and different filtering rules. You could also compare transcript factors against a few non-hub genes to show why your top pick stands out. If you pair that with a clear structure-based argument for druggability, your project looks much closer to real target discovery.

Project Variations

• Compare cytokine-storm signatures only in blood samples, then see whether the shared hub changes by tissue source.
• Focus on one pathogen pair, such as influenza and SARS-CoV-2, to test whether the overlap is stronger than the full three-pathogen comparison.
• Replace AlphaFold druggability scoring with pathway enrichment and upstream regulator analysis to see whether the same hub appears from a different method.

Learn More

• NCBI Gene Expression Omnibus (GEO): Search for public bulk-RNA-seq studies on influenza, SARS-CoV-2, and dengue, then download sample metadata and count data.
• PubMed: Search review articles on cytokine storm, transcriptomic signatures, and immune dysregulation to build your background section.
• NIH National Library of Medicine Bookshelf: Find free textbook chapters on immunology, gene regulation, and RNA-seq basics.
• Reactome: Use the pathway browser to check whether your candidate genes sit in interferon, inflammatory, or cytokine signaling pathways.
• Cytoscape Documentation: Read the free tutorials to learn how to build and rank gene interaction networks.
• AlphaFold Protein Structure Database: Look up predicted protein structures for your hub candidates and inspect possible binding pockets.