search

arvkevi / tgviz

1000 Genomes Project population visualizations using dimensionality reduction
http://tgviz.herokuapp.com/
MIT License
11 stars 4 forks source link
data-visualization dimensionality-reduction genomics genomics-visualization pca tsne umap

readme

tgviz

One Thousand Genomes Project -- Population Visualizations
http://tgviz.herokuapp.com/

How well do published ancestry informative SNPs discriminate between populations in the One Thousand Genomes Project?

This app and supporting jupyter notebooks (in notebooks/) provide Python functions and visualizations that can help answer this question.

tgviz is an interactive visualization app built in Python with Dash and deployed on Heroku. Developing this app was an opportunity to familiarize myself with the Dash framework. Be sure to read the README below the visualization on the app page to see more details.

gif

App Performance

After landing on the site, please allow a few seconds for the application to start.

Switching between PCA and UMAP dimensionality reduction algorithms are relatively quick, however selecting TSNE takes 15-20 seconds to load.

Project Personal Genotypes

The plot_walkthrough.ipynb in the notebooks directory outlines how to project genotype data from your own .vcf into the 1000 Genomes Project population space.

mev1kgp

Contributing

Contributions are welcome, if you find a bug or an inaccuracy, submit an issue, if you would like work on an open issue, or improve the app, submit a Pull Request.