Fig. 2
From: Sensei: how many samples to tell a change in cell type abundance?
Results of simulation studies. a Comparison of false negative rate (y-axis) known from simulation against those estimated by Sensei and by the legacy approach, using datasets sampled from a beta-binomial distribution. Number of samples in the case group is indicated on the x-axis and in the control group by different colors. Markers correspond to result from different approaches. The average error is the mean absolute relative difference between the estimation and the simulation. b Comparison of false negative rates calculated by Sensei and the legacy approach, with those generated by simulation on the proportions of T cells in tumor and juxtatumoral samples in a breast cancer study