by Martin Vingron, Max Planck Institute for Molecular Genetics
Visualizing very large transcriptome data sets is still a major challenge. In particular, it is hard to visualize how sets of genes are associated to clusters of conditions in high-dimensional data, in terms of their expression. Even correspondence analysis, a method intended for studying this “gene-condition association question”, fails due to the shortcomings of a 2D- or 3D-embedding vis-a-vis the dimensionality of the data. We introduce Association Plots for visualizing condition-specific genes in complex data. Association Plots are two-dimensional, independent of the dimension of the data, and show the genes associated to a cluster of conditions. We demonstrate our method on GTEx RNA-seq data and PBMC single-cell RNA-seq data, with Association Plots depicting clearly those genes which characterize tissues or cell clusters.
Hosted by Donate Weghom
More information here