1kgp_dimred's Introduction

1KGP_dimred

Interactive demonstration of how to use PCA, t-SNE, and UMAP on genotype data from the Thousand Genome Project.

For the related manuscript, see Diaz-Papkovich, Alex, Luke Anderson-Trocmé, Chief Ben-Eghan, and Simon Gravel. "UMAP reveals cryptic population structure and phenotype heterogeneity in large genomic cohorts." PLoS genetics 15.11 (2019): e1008432. https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1008432

1kgp_dimred's People

Contributors

Stargazers

Watchers

1kgp_dimred's Issues

Number of samples in 1000 genomes

Your preprint says

The 1KGP contains genotype data of 3,450 individuals from 26 relatively distinct labeled populations

and when I run your script that's what I get too. At the same time, https://www.nature.com/articles/nature15393 says

Here we report completion of the project, having reconstructed the genomes of 2,504 individuals from 26 populations

Why 2504 vs 3450?

Adding a personal genome omni-chip individual

Hi,

I read the tweet about 1KGP_dimred https://twitter.com/adp_diaz/status/1044235174718951425 and I would like to know how difficult would it be to do the following:

(1) Take the omni file, since the personal genome I want to add is from MyHeritageDNA which is Illumina's 700K Omni chip. I presume it's the file below:
http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/supporting/hd_genotype_chip/ALL.chip.omni_broad_sanger_combined.20140818.snps.genotypes.vcf.gz
Would that file load if substituted to the one you point at in the ipynb?
(2) Add more individuals to the .vcf.gz file:
Given the format needed for the ipynb, what would be the best way to add extra individuals to the file, re-index, append the file to the .panel and tabular info files, and run the ipynb again?

What are the parameters for Figure S2

I tried to reproduce the results in Figure S2 (# princial components = 15). However, the results are different from what are shown in the supporting materials. I was just wondering if you could provide me the specific parameters for UMAP.

Below is my experiment configuration:
I first applied PCA to the transposed_genotype_matrix to get the proj_pca. Then I applied UMAP to proj_pca[:,:15] with n_components=2, n_neighbors=15, min_dist=0.5.