SARS-CoV-2 Spike Variant Binding Study
ISEF Category: Biochemistry
Ready to Turn This Idea Into a Real Project?
This guide was put together with the help of AI research tools to give you a solid starting point. But a competitive science fair project lives in the details: refining your research question, fine-tuning your variables, analyzing your data, and presenting your findings like a seasoned scientist.
For next steps tailored to your interests, skill level, and timeline, work one-on-one with a MehtA+ mentor. Learn more about MehtA+ Science & Engineering Research Mentorship →
Subcategory: Structural Biochemistry · Difficulty: Advanced · Setup: Home Setup · Time: 1 to 2 Months
The Hook
A few amino acid swaps can change how tightly a virus grabs a human cell. You can study that without growing virus by comparing predicted spike structures across variants. Think of the RBD like a key and ACE2 like a lock, then ask which key version fits best.
What Is It?
Spike is the protein SARS-CoV-2 uses to start infection. The receptor-binding domain, or RBD, is the part that touches ACE2, a protein on many human cells. Think of the RBD like the teeth of a key, and ACE2 like the lock that the virus tries to open.
AlphaFold-Multimer predicts how two proteins may fit together. Interface energetics means you estimate how stable that contact surface looks, based on things like hydrogen bonds, salt bridges, and shape fit. You are not measuring live-virus behavior, but you are asking a clean structural question about how the Wuhan, JN.1, and KP.3 RBDs differ at the binding edge.
Why This Is a Good Topic
This is a strong science fair topic because you can test it with public data and clear measurements. It connects directly to viral evolution and how mutations change protein-protein binding, which matters for tracking new variants. You can learn sequence alignment, structure comparison, and simple statistics without needing a wet lab. The question is narrow enough for a student, but deep enough to support real analysis.
Research Questions
- How does the predicted ACE2 interface contact count differ between Wuhan, JN.1, and KP.3?
- What is the effect of each variant on predicted interface energy with ACE2?
- Does the number of RBD mutations near known ACE2 contact residues change across the three variants?
- To what extent do predicted hydrogen bonds and salt bridges shift at the ACE2 binding edge?
- Which variant shows the biggest change in model confidence at the interface?
- How does the overall RBD fold compare with the interface region across variants?
Basic Materials
- Laptop or desktop computer with at least 16 GB RAM.
- Reliable internet access.
- Spreadsheet software such as Google Sheets or Excel.
- Python with Jupyter Notebook.
- Sequence and structure records from NCBI Virus, UniProt, and RCSB PDB.
Advanced Materials
- GPU-capable workstation or cloud compute instance.
- ColabFold or a local AlphaFold-Multimer installation.
- UCSF ChimeraX for interface inspection.
- FoldX for interface energetics estimates.
- GROMACS for optional molecular dynamics checks.
Software & Tools
- Python: Cleans sequence data, tracks mutation tables, and plots interface scores.
- Jupyter Notebook: Keeps your workflow reproducible and easy to rerun.
- Biopython: Reads FASTA and PDB files and maps mutations to residues.
- UCSF ChimeraX: Visualizes the spike-ACE2 interface and measures contacts.
- ColabFold: Predicts multimer structures when you do not have a local GPU.
Experiment Steps
- Define the exact RBD sequence window and the lineage set you will compare.
- Pick one prediction workflow and keep it fixed across every variant.
- Set ACE2 as the same partner in each run, so your interface comparison stays fair.
- Choose the structural readouts you will score, such as contact count, hydrogen bonds, salt bridges, and interface energy.
- Build a comparison table that links each mutation to its predicted effect and flags where model confidence is low.
Common Pitfalls
- Comparing variant models that were built from different RBD residue ranges, which makes the interface numbers incompatible.
- Using different prediction settings for each variant, which turns method changes into fake biological differences.
- Ignoring residue-number shifts after sequence alignment, which hides the exact contact sites that changed.
- Trusting one predicted pose without checking model confidence, which can overstate weak structural signals.
- Reading interface energy as direct proof of transmission advantage, which goes beyond what the model can show.
What Makes This Competitive
A stronger version of this project does more than rank variants by one score. It maps each mutation to a specific contact change, checks more than one prediction run, and compares the same result across models. You can raise the level again by testing whether interface shifts agree with known experimental literature on ACE2 binding. Clear limits, careful controls, and honest uncertainty make the work look much more like real structural research.
Project Variations
- Compare JN.1 and KP.3 against an earlier Omicron RBD to see whether the binding edge keeps shifting in the same direction.
- Swap the scoring method, then see whether AlphaFold-Multimer and a second interface tool point to the same mutation hotspots.
- Focus on residue-level contact maps instead of global energy, then test which contacts best explain the predicted binding shift.
Learn More
- RCSB PDB: Search the Protein Data Bank for spike, RBD, and ACE2 structures, then read the linked experimental papers.
- UniProt: Find reviewed spike protein records and mutation annotations for sequence mapping.
- NCBI Virus: Download SARS-CoV-2 lineage sequences and compare variant mutations.
- NIH PubMed: Search review articles on SARS-CoV-2 spike structure and ACE2 binding.
- MIT OpenCourseWare: Look for free structural biology and bioinformatics lectures for background on folds and interfaces.