Impact of unequal sample sizes. Pairwise comparisons of perfect in silico replicate RNA-Seq datasets were made similarly to Figure 2, but keeping the total read count for each pair constant while systematically varying the relative size of both datasets, e.g., 1x106 plus 9x106, 3 x106 plus 7 x106, 5 x106 plus 5 x106. Pearson’s r and Kappa fell for unequal sample sizes demonstrating a maximum for equal sample sizes. SERE remained stable at 1.0 while the 99% CI of repeat measures was optimal (smallest) for equal sample sizes.
Schulze et al. BMC Genomics 2012 13:524 doi:10.1186/1471-2164-13-524