Fetal Blood Cell Switching With RNA Velocity

ISEF Category: Cellular and Molecular Biology

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

The Hook

Your blood starts life in a very different mode than it uses after birth. A tiny cell population can appear, change fast, and then vanish as development moves on. If you can track that shift, you can study one of the key handoffs in human biology. That makes this a strong data science project, not just a biology one.

What Is It?

This project uses single-cell RNA sequencing data, which lets you see which genes are active in each cell one at a time. Think of it like listening to every instrument in an orchestra separately instead of hearing the whole song at once. You are looking for a short-lived progenitor population, which means a stem-like cell that can still become different blood cells.

RNA velocity is a method that estimates where a cell is headed next based on the balance of newly made RNA and older RNA. In plain language, it gives you a direction arrow for each cell. You can use that arrow to test whether one cell state disappears as another blood program, including hemoglobin switching from fetal to adult forms, begins to rise.

Why This Is a Good Topic

This is a strong science fair topic because the data already exist, the question is narrow, and the analysis can still be original. You are not just copying a chart, you are testing whether a rare cell state lines up with a developmental transition. The project connects to real problems in blood disorders, newborn biology, and the way researchers study cell fate. You can learn single-cell analysis, differential expression, trajectory mapping, and careful interpretation of noisy biological data.

Research Questions

• How does the abundance of the transient progenitor population change across fetal developmental stages?
• What is the effect of removing low-quality cells on the detection of the progenitor cluster?
• Does RNA velocity point from the transient progenitor state toward erythroid maturation?
• To what extent does the timing of progenitor disappearance match the rise of adult hemoglobin gene expression?
• Which marker genes best separate the transient progenitor population from nearby hematopoietic cells?
• How does the inferred cell trajectory change when you compare scVelo with a second trajectory method?

Basic Materials

• Computer with enough memory and storage to handle single-cell datasets.
• Internet access for downloading public HCA data.
• Python installed with Jupyter Notebook or JupyterLab.
• scVelo and Scanpy Python packages.
• Basic spreadsheet software for tracking samples and results.
• PubMed and GEO access for finding related papers and datasets.

Advanced Materials

• Access to a university or high-memory workstation for larger single-cell datasets.
• Python environment with scVelo, Scanpy, anndata, matplotlib, seaborn, and pandas.
• R with Seurat for cross-checking clustering and marker analysis.
• Reference genome and annotation files matched to the dataset.
• Cloud computing account or campus server for repeat runs and parameter testing.
• ImageJ or Napari if you want to inspect published figures or custom cell plots.

Software & Tools

• scVelo: Estimates RNA velocity and helps you test cell-state direction in developing blood cells.
• Scanpy: Handles single-cell quality control, clustering, and marker gene analysis.
• JupyterLab: Lets you document each analysis step and keep code, notes, and plots together.
• Python: Runs the core single-cell analysis workflow and custom plotting.
• PubMed: Helps you find review articles and primary papers on fetal hematopoiesis and hemoglobin switching.

Experiment Steps

1. Define the biological question by choosing one developmental window and one cell lineage to track.
2. Select a public dataset with enough cells, metadata, and sequencing depth to support trajectory analysis.
3. Plan quality filters and clustering rules before you inspect the final result, so you do not tune the analysis to match your guess.
4. Build marker-based cell labels and a standard expression comparison for the progenitor and mature states.
5. Add RNA velocity and trajectory tools to test whether the cell arrows support a transient intermediate state.
6. Compare your findings across at least one alternate dataset or analysis setting to see whether the pattern stays stable.

Common Pitfalls

• Treating every early blood cell as the same population, which hides the rare progenitor state you want to study.
• Using a dataset with weak developmental coverage, which makes the timing of the switch impossible to resolve.
• Over-filtering cells, which can remove the transient progenitor cluster before you analyze it.
• Reading RNA velocity arrows as direct proof of lineage, instead of as a model that needs validation.
• Ignoring batch effects or sample differences, which can make the progenitor seem real when it only reflects dataset structure.

What Makes This Competitive

A stronger project will not just report one cluster. It will test whether the signal survives strict quality control, a second dataset, and at least one alternate trajectory method. You can raise the level by comparing fetal stages, marker sets, and hemoglobin gene programs with careful statistics. If you explain where the model is confident, and where it is weak, your analysis will read like real research.

Project Variations

• Use cord blood or neonatal blood datasets to see whether the same transient progenitor signal appears outside fetal tissue.
• Compare scVelo with pseudotime methods such as Monocle to test whether both approaches recover the same developmental order.
• Focus on globin gene families and test whether progenitor loss predicts the fetal-to-adult hemoglobin transition more closely than overall erythroid maturation.

Learn More

• HCA Data Portal: Search for fetal tissue single-cell datasets and metadata from the Human Cell Atlas project.
• PubMed: Search for review articles on fetal hematopoiesis, hemoglobin switching, and RNA velocity.
• NIH Genomic Data Sharing resources: Read about responsible use of public genomic datasets and metadata.
• scVelo documentation: Learn the analysis workflow, theory, and plotting options from the official project docs.
• MIT OpenCourseWare: Look for free lectures on genetics, genomics, and bioinformatics foundations.