Technical Comments

Response to Comment on “Statistical binning enables an accurate coalescent-based estimation of the avian tree”

See allHide authors and affiliations

Science  09 Oct 2015:
Vol. 350, Issue 6257, pp. 171
DOI: 10.1126/science.aaa7719


  • Fig. 1 Phylogenomic pipelines: unbinned (top), weighted statistical binning (middle), and unweighted statistical binning (bottom).

    Statistical binning divides the genes into bins that have no highly supported conflicts, estimates supergene trees on each bin, and then combines the supergene trees using the selected summary method. The WSB method differs from unweighted binning by replicating each supergene tree by the number of genes within its bin.

  • Fig. 2 Simulation studies evaluating the impact of WSB on MP-EST analyses.

    (A) Tree error rates (percentage of missing branches) for species trees estimated with WSB (MLBS gene trees, fully partitioned ML analyses) and unbinned analyses on the five-species data sets studied by Liu and Edwards, and similar model conditions with 10 and 15 species, all with 1000 genes. Symbols: ↑ indicates that using WSB increases species tree estimation error; ↓ indicates that using WSB decreases error. (B) Number of replicates for which the true species tree is recovered on the five-species data sets studied by Liu and Edwards. (C) Results on simulated avian data sets with 1X branch lengths and 500-bp sequences per locus, based on the avian tree from (12). MP-EST analyses are based on multilocus bootstrapping, and unpartitioned ML analyses are used to compute supergene trees.

Navigate This Article