Fig. 1

Performance comparison for data sets with a varying number of cells. SiFit’s tree reconstruction accuracy is compared against that of SCITE, OncoNEM, and MrBayes. The y-axis denotes the tree reconstruction error, which measures the distance of the inferred tree from the ground truth. Three points for n=200, n=400, and n=600 are plotted on the x-axis. In each case, the mean tree reconstruction error over ten data sets is plotted. The vertical error bar indicates the standard deviation of the tree reconstruction error over ten data sets. a Performance comparison for data sets with 50 cells and δ=0. b Performance comparison for data sets with 100 cells and δ=0. c Performance comparison for data sets with 200 cells and δ=0. d Performance comparison for data sets with 50 cells and δ=0.05. e Performance comparison for data sets with 100 cells and δ=0.05. f Performance comparison for data sets with 200 cells and δ=0.05. g Performance comparison for data sets with 50 cells and δ=0.1. h Performance comparison for data sets with 100 cells and δ=0.1. i Performance comparison for data sets with 200 cells and δ=0.1. For d–i, the tree reconstruction error is measured without considering the doublets in both true and inferred trees. FN false negative, FP false positive