CRISPR-Cas13 Guide RNA Design

ISEF Category: Biomedical and Health Sciences

Subcategory: Genetics and Molecular Biology of Disease  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: 1 to 2 Months

The Hook

Imagine pointing a pair of molecular scissors at a virus the moment it shows up in a wastewater sample. CRISPR-Cas13 cuts single-stranded RNA, which is exactly what pandemic viruses like flu, SARS-CoV-2, and Nipah are made of. The hardest part is not the chemistry. It is picking the right short target so the enzyme hits the virus and ignores your own cells.

What Is It?

Cas13 is a CRISPR enzyme that binds and cuts single-stranded RNA. You program it with a short guide RNA, called a gRNA, that matches the virus sequence you want to destroy. Think of the gRNA as a printed address label and Cas13 as a delivery truck that only stops at that address.

A guide RNA design pipeline scores every possible target window across a viral genome and asks three questions about each one. Is this region conserved across many strains, so the design still works next year? Does the human transcriptome contain a near-match where Cas13 might cause off-target cutting? Is the RNA in this region actually accessible, or is it folded into a tight hairpin that Cas13 cannot reach?

The whole project lives on a laptop. You download public sequences from NCBI, score them with open-source tools, and rank candidate guides. No wet lab. No live virus. Just sequence files, scripts, and a clean writeup.

Why This Is a Good Topic

This is a strong science fair topic because it ties a real public health problem (the next pandemic) to a measurable, fully computational pipeline you can defend in front of judges. Every design choice produces a number you can compare: conservation score, off-target hit count, accessibility energy. You can build, break, and rebuild the pipeline without leaving your desk, and a high school student can realistically learn alignment, scoring functions, and basic RNA structure prediction in a few weeks.

Research Questions

• How does the number of viable Cas13 guide RNAs change as you raise the conservation threshold across known strains?
• What is the effect of including predicted RNA secondary structure on the final guide ranking compared to sequence alone?
• Does scoring off-targets against the full human transcriptome change which guides reach the top 10 versus scoring against coding sequences only?
• To what extent do top-ranked guides differ between two related viruses in the same family, such as influenza A and influenza B?
• Which conserved regions of a pandemic-potential virus genome consistently produce high-quality guides across multiple scoring schemes?

Basic Materials

• Laptop with at least 8 GB of RAM and 50 GB of free disk space.
• Stable internet connection for downloading reference sequences.
• Notebook or digital lab journal to record every parameter you change.
• Free GitHub account to version-control your scripts and results.

Advanced Materials

• Workstation or cloud VM with 32 GB of RAM for larger transcriptome scans.
• Access to a university high-performance computing cluster for full-genome multiple sequence alignments.
• Local install of BLAST+ command-line tools for fast off-target searches.
• Reference human transcriptome FASTA downloaded from Ensembl or GENCODE.

Software & Tools

• Python: Writes the scoring pipeline, parses FASTA files, and ranks candidate guide RNAs.
• Biopython: Reads sequence files and runs basic alignment operations from inside your script.
• MAFFT: Produces multiple sequence alignments across viral strains so you can score conservation.
• ViennaRNA (RNAfold): Predicts secondary structure and accessibility of each candidate target site.
• BLAST+: Scans the human transcriptome for near-match off-targets of every guide candidate.

Experiment Steps

1. Choose one pandemic-potential RNA virus family and download a representative strain set from NCBI Virus.
2. Decide your guide length and your minimum conservation threshold before you look at any results.
3. Build a multiple sequence alignment and write code that scores every window for conservation across strains.
4. Plan an off-target step that defines what counts as a near-match in the human transcriptome and why.
5. Add an accessibility filter so deeply folded regions of the viral RNA are downranked.
6. Compare your top guides against your starting assumptions and document every parameter you changed and why.

Common Pitfalls

• Picking only one virus strain, which makes your conservation score meaningless because there is nothing to compare against.
• Defining off-targets as exact matches only, which hides the real risk because Cas13 also cuts at sequences with one or two mismatches.
• Using the human genome instead of the transcriptome for off-target scoring, which misses intronic and noncoding RNAs that Cas13 can still bind.
• Ignoring secondary structure entirely, so your top guides target regions that are folded into stable hairpins and not actually accessible.
• Changing several pipeline parameters at once during tuning, which makes it impossible to tell which choice produced the new ranking.

What Makes This Competitive

A class-level version of this project ranks guides by conservation and stops there. A competitive ISEF-level version layers in a real off-target model (mismatch-tolerant search against the human transcriptome) and an accessibility model (per-base unpaired probability from ViennaRNA), then shows how the top guides shift under each layer. Add a sensitivity analysis that varies one parameter at a time and reports how the top 10 changes. Finish with a side-by-side comparison across two related viruses, so judges see the pipeline generalizes.

Project Variations

• Run the same pipeline on a different virus family, such as flaviviruses, to test whether your scoring scheme transfers without tuning.
• Replace the off-target step with a tissue-specific transcriptome (e.g., lung) and ask whether top guides change when the at-risk tissue changes.
• Add a fairness check that asks how well guides chosen on early-pandemic sequences score against later, more diverged variants pulled from public databases.

Learn More

• NCBI Virus: Free portal at NCBI for downloading curated viral genome sequences by species, host, and country.
• Ensembl and GENCODE: Free reference human transcriptome FASTA and annotation files you can use for off-target scoring.
• ViennaRNA Web Services: Free RNAfold and RNAplfold tools with documentation on accessibility and structure prediction.
• PubMed: Search review articles on Cas13 diagnostics and guide-RNA design for background and citations.
• MIT OpenCourseWare: Free introductory computational biology lectures that cover sequence alignment and scoring functions.
• Rosalind: Free online platform for practicing bioinformatics programming problems in Python.