Fig. 8
From: The performance of coalescent-based species tree estimation methods under models of missing data
Comparison of species tree estimation methods given estimated gene trees for the Mclade model of missing data. Species tree error (defined as the RF error rate) is shown for species trees estimated by giving increasing numbers of genes as inputs to ASTRAL (blue), ASTRID (orange), MP-EST (green), and SVDquartets (red). All three summary methods were given estimated gene trees, and SVDquartets was given the true multiple sequence alignment. Each column represents a different level of incomplete lineage sorting (ILS): panels (a), (d), and (g) show datasets with low ILS, panels (b), (e), and (h) show datasets with high ILS, and panels (c), (f), and (i) show datasets with very high ILS. Each row represents a different percentage of genes with missing data: panels (a)-(c) show datasets with 0% of genes missing data, panels (d)-(f) show datasets with 55% of genes missing data, and panels (g)-(i) show datasets with 95% of genes missing data. For example, if 55% of the 1000 genes are missing data, then 450 genes are complete and 550 genes are incomplete (due to clade-based missing data). Lines represent the average over 20 replicate datasets, and filled regions indicate the standard error. Datasets shown here have deep speciation events; results for datasets with recent speciation are shown in Additional file 1