We use this type of map to cluster phenotypes by genetic relatedness. Each point is a genome-wide association study (GWAS) for a specific phenotype, with a radius proportional to the GWAS sample size. The map is based on genetic correlations (computed using LD Score regression). The genetic correlation matrix was mapped into a 2D space using different algorithms: UMAP, GTM, and t-SNE, using the python packages umap-learn 0.3.7, ugtm 2.0.0, and scikit-learn 0.20.0 with the following hyperparameters: {n_neibhbors = 30, min_dist = 0.1} for UMAP, {perplexity = 30, early exaggeration = 12, learning rate = 200} for t-SNE, {map resolution = 12*12, radial basis function grid size = 3*3, regularization = 0.1, rbf width factor = 0.3} for GTM. Maps with different hyperparameter settings can be selected in the navigation bar. The interactive visualization was generated using the python package altair, which is built on top of the Vega-Lite visualization grammar. Tablets and smartphones will not provide a full user experience (no click or interval selection available).