Fig. 3
From: GEMINI: a computationally-efficient search engine for large gene expression datasets
Differential gene expression for ovarian (OV) and breast (BRCA) cancer samples from TCGA. (left) Principal component analysis is used to project the 17,813 dimension gene expression data to two dimensions for visualization. The ovarian samples and breast samples clearly cluster. One ovarian sample (C) has an expression pattern similar to breast cancer samples and one (D) shows an expression pattern outside of both the ovarian and breast clusters. Representative breast (A) and ovarian (B) samples are circled. (right) A boxplot of all non-zero pairwise distances in the joint breast and ovarian cancer data sets. The nearest neighbors for the four queries are shown as symbols in the legend. We find that the nearest neighbors all fall closer than the lower quartile of all of the distances