During the summer and fall of 2023 I worked with Professors David Laidlaw and Sohini Ramachandran to figure out novel ways to display genetic data. We were working with datasets taken from the 1000 Genome Project that had genetic samples from across the human population. The issue we were trying to solve was that the way genetic data is usually displayed makes it seem like there is more genetic variation between population groups than within a given population, which is not true and is prone to being used by people acting in bad faith.

We researched a few different ways to display this data. Our first step was to use Principle Component Analysis on the datasets to give us data we could graph in 2D space. After that, we tried limiting the allele frequencies of the data we used, changing the amount of total gene samples, and using Geovar to try to isolate individual genes. These were useful experiments, but ultimately our research led us into creating something more interactive.

The program that we design and then implemented in Javascript allows multiple datasets to be loaded and tracks samples used between datasets. This helps show where a sample lies within its own population and within the populations. The user can also display the data as a heatmap, which is useful for seeing where there are high concentrations in samples.