AliSim: A Fast and Versatile Phylogenetic Sequence Simulator for the Genomic Era

Abstract

Sequence simulators play an important role in phylogenetics. Simulated data has many applications, such as evaluating the performance of different methods, hypothesis testing with parametric bootstraps, and, more recently, generating data for training machine-learning applications. Many sequence simulation programmes exist, but the most feature-rich programmes tend to be rather slow, and the fastest programmes tend to be feature-poor. Here, we introduce AliSim, a new tool that can efficiently simulate biologically realistic alignments under a large range of complex evolutionary models. To achieve high performance across a wide range of simulation conditions, AliSim implements an adaptive approach that combines the commonly used rate matrix and probability matrix approaches. AliSim takes 1.4 h and 1.3 GB RAM to simulate alignments with one million sequences or sites, whereas popular software Seq-Gen, Dawg, and INDELible require 2-5 h and 50-500 GB of RAM. We provide AliSim as an extension of the IQ-TREE software version 2.2, freely available at www.iqtree.org, and a comprehensive user tutorial at http://www.iqtree.org/doc/AliSim.

Keywords: molecular evolution; phylogenetics; sequence simulation.

Fig. 1.

Sequence simulation process with two scenarios: (A) Simulating an MSA from a phylogenetic tree and a Markov substitution model, and (B) Simulating an MSA that mimics the underlying evolutionary process of a user-provided MSA. Here, the phylogenetic tree and the substitution model parameters are internally inferred from the user-provided MSA, which are used to simulate a new MSA.

Fig. 2.

Runtimes and peak memory consumptions of five software AliSim, Seq-Gen, Dawg, INDELible, and phastSim for deep-data (varying number of sequences and 30K sites; sub-panels A and B) simulations without indels, long-data (varying number of sites and 30K sequences; sub-panels C and D) simulations without indels, and varied #sequences (varying number of sequences and setting root sequence length at 30K sites; sub-panels E and F) simulations with indels.