You can do real bioinformatics research in high school without access to a lab, because the sequences, protein structures and gene-expression datasets researchers use are publicly available. What you need is a focused question, the right dataset and someone who can check your methods as you work. If you’re comparing bioinformatics research programs for high school students, look at how much one-on-one technical feedback each program gives you, since that feedback is what turns a tutorial into an original project.
The six project ideas below show what that work can look like. Each one pairs a biological question with a public resource, a workable comparison and the skills you’ll need before you start, similar to the directions a research mentor might work through with a student.
Which parts of a protein remain similar across species, and how do those regions relate to its function? Use UniProt’s sequence-alignment tools to compare corresponding proteins from a small set of mammals. You might begin with a well-studied protein such as p53.
Select comparable, full-length sequences and confirm that you are comparing the same gene’s counterpart across species, rather than mixing related family members or different isoforms. Align the sequences, then compare conservation inside an annotated functional domain with conservation elsewhere.
You will need amino-acid vocabulary, FASTA-format familiarity and basic proportions. A short Python script can summarize matching positions, while browser tools support initial exploration. A useful extension tests whether the pattern persists when you change the species set. Conservation can motivate a functional hypothesis, but it does not by itself demonstrate what a residue does.
Which genes change expression consistently after a defined treatment? The Bioconductor airway dataset provides a practical starting point using human airway smooth-muscle cells exposed to dexamethasone and corresponding untreated samples. Work with the documented treatment/control subset and processed gene counts.
Compare treated and untreated samples while accounting for the matched cell lines in your statistical design. Examine effect sizes alongside adjusted p-values, then ask whether selected responses are consistent across cell lines. This makes biological variation part of the investigation instead of treating every sample as interchangeable.
This direction suits students ready to learn R, count-data normalization, statistical modeling and multiple-testing correction. Reproducing a published result establishes that your workflow behaves sensibly. An extension could examine a literature-motivated gene set or the robustness of a finding, with the original study and later literature defining what remains worth asking.
How do proteins sharing one domain differ in their other molecular components? InterPro identifies protein families and domains, giving you a way to investigate how domain combinations relate to annotated functions.
Choose a bounded group, such as human proteins containing a particular kinase domain. Record each protein’s domain arrangement and functional annotation. Compare proteins with and without an additional regulatory domain, using one defined sequence per protein.
Start with a spreadsheet, basic protein biology and careful identifier tracking. Python becomes useful when you automate a larger comparison. Your output could be a domain diagram and a summary of functional categories. Check whether differences remain after excluding incomplete sequences or poorly characterized entries. Treat annotations as evidence to examine, distinguishing experimentally supported functions from predictions rather than assuming every database label has equal support.
For a gene associated with an inherited condition, how do the classifications and supporting review levels differ across its reported variants? ClinVar provides public variant records, while its review-status definitions explain how supporting review is represented.
Choose one gene and one relevant condition. Build a clearly filtered set of germline variant-condition records, separating benign, pathogenic, uncertain and conflicting classifications. Compare category proportions across review-status groups, retaining likely benign and likely pathogenic labels separately or documenting any grouping.
This is an accessible descriptive project for students comfortable with spreadsheets, percentages and basic genetics. Record the download date and avoid counting repeated submissions as independent variants. A useful result explains how evidence filters change the picture of a gene’s reported variants. Those database proportions do not estimate disease prevalence or establish a new clinical classification.
How does hemoglobin’s structure differ when oxygen is bound? The RCSB PDB-101 hemoglobin guide introduces structures 1HHO and 2HHB, representing oxygenated and deoxygenated human hemoglobin. They provide concrete structures for connecting molecular shape with biological function.
Align corresponding chains, identify matching residues and compare selected distances near the heme group or between subunits. Keep atom definitions consistent and use the appropriate biological assembly when examining relationships among all four chains.
You will need protein-structure vocabulary, spatial reasoning and practice with a molecular viewer. Coding can help repeat measurements across additional structures. Begin by reproducing a known structural difference, then investigate whether that measurement is consistent across comparable entries. The initial pair is a learning baseline; a research extension needs a question beyond rediscovering this well-established oxygen-binding mechanism.
Which biological processes are represented among genes whose expression increases versus decreases in one experiment? Reactome’s pathway tools connect gene identifiers with curated biological pathways, helping you examine coordinated processes rather than isolated genes.
Start with an existing differential-expression analysis and define the same selection criteria for both directions. Map each list to pathways, compare the fraction of successfully mapped genes associated with selected processes, and inspect the actual genes behind each pattern. Record unmapped identifiers and overlapping pathway memberships.
This project requires gene-expression fundamentals, spreadsheet or R skills, and careful use of denominators when the lists differ in size. Test whether the descriptive pattern changes under a reasonable alternative threshold. Pathway coverage alone does not establish statistical enrichment or pathway activation. A mentor can help you develop an appropriate statistical extension if that is necessary for your question.
Write a short proposal explaining your question, why existing work leaves it open and how your comparison will answer it. Stanford’s proposal guidance offers a useful framework for connecting objectives, literature, methods and preparation. Our guide to choosing a research topic can help you narrow a broad interest.
A different dataset or extra chart becomes valuable when it tests a meaningful uncertainty. For example, you might investigate whether a published conclusion holds in a biologically justified comparison, or explain why a result changes under defensible analysis choices.
Programs like Echelon Scholars' research program for high school students are designed to help young researchers develop ambitious original research through individualized technical guidance, detailed feedback and sustained one-on-one mentorship.
Our unlimited-hour mentorship gives you sustained support for that technical iteration, from checking methods to developing a manuscript and responding to peer review. We provide accountability and detailed feedback while you build the understanding to explain your own decisions. Explore our admissions pathway if you want to develop a focused biological question into substantial research.
Yes. Small sequence, domain and variant-record projects can begin with browser tools and spreadsheets. Research mentorship or a research program designed for high school students can help you choose a useful scope and learn coding when repeated analyses or more demanding statistical methods make it necessary.
No. Public sequences, structures and processed datasets can support computational investigations. Expert guidance helps you select data that actually answer your biological question and interpret what the analysis can establish.
A small protein-domain comparison is a manageable starting point if you understand basic protein biology. You can build a clear comparison before adding statistical complexity, with a mentor helping you assess annotation quality and scientific relevance.
No. Originality depends on the question and contribution, not solely on who collected the data. A literature review and mentor feedback help distinguish a justified new analysis from repeating an established tutorial.
Produce a small, reproducible pilot with a documented dataset, clear comparison and interpretable figure. Explain how you obtained the result and what remains uncertain. That gives your mentor concrete work to evaluate before you scale up.
About Echelon Scholars:
Echelon Scholars is a research mentorship program that pairs high school students with graduate-level mentors from Harvard, Stanford, and UC Berkeley to conduct guided but independent research. Students interested in biomedical research, computer science, cancer research, environmental research, and more learn how to take their research question beyond a summer program, science fair, or summer research project and turn it into a full-scale PhD level research experience that results in publication to help them stand out on college applications. Our editorial team covers academic enrichment, research opportunities, and college preparation.
