Plant Virus Mapping With Citizen Science
ISEF Category: Microbiology
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: Virology · Difficulty: Advanced · Setup: University Lab · Time: Full Year
The Hook
Plant viruses do not spread the way human viruses do, but they can still move fast through a garden. Aphids often act like tiny hitchhikers that carry them from plant to plant. If you can map symptoms and vector sightings, you can test a real transmission model. That turns a neighborhood garden into a live epidemiology dataset.
What Is It?
This project studies how plant viruses spread in real spaces. You look for mosaic-like symptoms, which are patchy light-and-dark patterns on leaves. Then you compare those symptom maps with aphid observations from citizen-science records. The goal is to see whether places with more likely vectors also show more plant disease.
You also train a CNN, which stands for convolutional neural network, a type of image model that learns patterns from photos. Think of it like teaching a student to spot the visual clues that point to disease. PlantVillage gives you labeled plant images to build that skill. After training, you can test whether the model helps sort real field photos into likely healthy or likely infected groups.
Why This Is a Good Topic
This is a strong science fair topic because you can connect three real skills, image classification, ecology, and disease spread. You have a clear question, a public dataset, and a way to collect your own local observations. You can test whether aphid density, garden type, or plant spacing predicts symptom clusters. That gives you a project with data, models, and a story about plant health that matters to growers and community gardens.
Research Questions
- How does aphid observation density relate to the number of mosaic-symptom plants in community gardens?
- What is the effect of garden plant spacing on the local clustering of suspected virus symptoms?
- Does a CNN trained on PlantVillage images classify field photos of symptomatic leaves better than a simple color-based rule?
- To what extent do symptom hotspots overlap with areas that have more aphid sightings in iNaturalist records?
- Which host plant species show the highest rate of likely virus symptoms in the same neighborhood?
- How does adding location and date metadata change the strength of the transmission model?
Basic Materials
- Smartphone camera with manual focus and exposure control.
- Notebook or spreadsheet for field notes.
- Access to PlantVillage image dataset.
- Access to iNaturalist observation records.
- Free mapping tool such as Google My Maps or QGIS.
- Computer with internet access.
- Measuring tape or phone GPS for rough garden distances.
Advanced Materials
- Digital microscope camera or clip-on macro lens.
- GPS-enabled tablet or phone for geotagged field sampling.
- University or school server access for CNN training.
- Python environment with TensorFlow or PyTorch.
- Image annotation tool such as LabelImg or CVAT.
- ArcGIS or QGIS for spatial analysis.
- R or Python libraries for spatial statistics and model testing.
- Access to plant pathology reference images or herbarium records.
Software & Tools
- Python: Runs image classification, data cleaning, and statistical analysis.
- Google Colab: Lets you train a CNN without installing heavy software.
- QGIS: Maps symptom locations and aphid observations.
- ImageJ: Measures leaf color patterns and symptom area in photos.
- R: Tests whether vector density predicts disease clustering.
Experiment Steps
- Define the disease signal you will score, such as mosaic pattern, chlorosis, or leaf distortion.
- Choose your sampling unit, such as one plant, one bed, or one garden plot, so your maps stay consistent.
- Build a training set from PlantVillage and separate out a clean test set from your own field photos.
- Plan a scoring system that links each photo to a location, a host species, and a symptom label.
- Design a comparison between symptom hotspots and aphid observation density, then pick the spatial test you will use.
- Decide how you will check model accuracy, false positives, and whether the pattern holds across sites or dates.
Common Pitfalls
- Using only PlantVillage-style lab photos, which makes the CNN fail on messy real garden images.
- Mixing different leaf angles and lighting conditions, which confuses symptom severity with photo quality.
- Treating every mosaic pattern as a virus without checking for nutrient stress, herbicide injury, or mechanical damage.
- Comparing aphid sightings from one app date range to plant symptoms from a different season, which breaks the transmission link.
- Ignoring spatial autocorrelation, which makes nearby plants look like independent evidence when they are not.
What Makes This Competitive
A stronger project goes beyond simple symptom counts. You would test multiple host plants, compare more than one classifier, and use a real spatial model instead of a basic graph. You can also separate aphid abundance from garden structure, weather, and plant species so your result says something specific about transmission. That kind of analysis feels much closer to research than a classroom demo.
Project Variations
- Compare virus-like symptom maps in school gardens versus home gardens to see whether management style changes clustering.
- Replace aphid data with whitefly or thrips observations to test whether a different vector better matches symptom spread.
- Train one model on leaf photos alone, then test a second model that also uses location and date metadata.
Learn More
- PlantVillage dataset: A public plant disease image collection, found by searching the PlantVillage site and related research papers.
- iNaturalist: A nonprofit citizen-science database for aphid and other insect observations, searchable by location and date.
- USDA Agricultural Research Service: Plant virus fact sheets and crop disease background, available through USDA pages.
- PubMed: Search review articles on plant virus transmission, aphid vectors, and mosaic symptoms.
- QGIS Documentation: Free guides for mapping and spatial analysis, found on the QGIS website.
- Google Colab Help: Free notebooks for running Python models in the browser, with tutorials on model training and file upload.