Fast and accurate bootstrap confidence limits on genome-scale phylogenies using little bootstraps

Abstract

Felsensteins bootstrap resampling approach, applied in thousands of research articles, imposes a high computational burden for very long sequence alignments. We show that the bootstrapping of a collection of little subsamples, coupled with median bagging of subsample confidence limits, produces accurate bootstrap confidence for phylogenetic relationships in a fraction of time and memory. The little bootstraps approach will enhance rigor, efficiency, and parallelization of big data phylogenomic analyses.