Fig. 1
From: Size matters: how sample size affects the reproducibility and specificity of gene set analysis
Pine plots for GSEA-G and GSEA-S (dataset GSE53757) Pine plots for dataset GSE53757 showing reproducibility of the results from GSEA-S (left) and GSEA-G (right) across sample sizes. Reproducibility is quantified by overlap score (Eq. 1). Each layer of the pine plot illustrates the overlap score of the results of a method for 10 replicate datasets with the same sample size. From top to bottom, the pine plots show replicates with sample size 2×20, 2×15, 2×10, 2×5, and 2×3. The overlap score ranges from 0 to 1 represented by a gradient from blue to red, respectively, separated by yellow in the middle (overlap of 0.5). The overlap score increases as sample size increases; however, the rate of increase for GSEA-S is lower than that of GSEA-G