On the inference of complex phylogenetic networks by Markov Chain Monte-Carlo

Abstract

For various species, high quality sequences and complete genomes are nowadays available for many individuals. This makes data analysis challenging, as methods need not only to be accurate, but also time efficient given the tremendous amount of data to process. In this article, we introduce an efficient method to infer the evolutionary history of individuals under the multispecies coalescent model in networks (MSNC). Phylogenetic networks are an extension of phylogenetic trees that can contain reticulate nodes, which allow to model complex biological events such as horizontal gene transfer, hybridization, introgression and recombination. We present a novel way to compute the likelihood of biallelic markers sampled along genomes whose evolution involved such events. This likelihood computation is at the heart of a Bayesian network inference method called SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW, as it extends the SO_SCPLOWNAPPC_SCPLOW method [1] inferring evolutionary trees under the multispecies coalescent model, to networks. SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW is available as a package of the well-known beast 2 software. Recently, the MCMCBiMarkers method [2] also extended SO_SCPLOWNAPPC_SCPLOW to networks. Both methods take biallelic markers as input, rely on the same model of evolution and sample networks in a Bayesian framework, though using different methods for computing priors. However, SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW relies on algorithms that are exponentially more time-efficient on non-trivial networks. Using extensive simulations, we compare performances of SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW and MCMCBiMarkers. We show that both methods enjoy similar abilities to recover simple networks, but SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW is more accurate than MCMCBiMarkers on more complex network scenarios. Also, on complex networks, SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW is found to be extremely faster than MCMCBiMarkers in terms of time required for the likelihood computation. We finally illustrate SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW performances on a rice data set. SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW infers a scenario that is compatible with simpler schemes proposed so far and provides additional understanding of rice evolution. Author summaryNowadays, to make the best use of the vast amount of genomic data at our disposal, there is a real need for methods able to model complex biological mechanisms such as hybridization and introgression. Understanding such mechanisms can help geneticists to elaborate strategies in crop improvement that may help reducing poverty and dealing with climate change. However, reconstructing such evolution scenarios is challenging. Indeed, the inference of phylogenetic networks, which explicitly model reticulation events such as hybridization and introgression, requires high computational resources. Then, on large data sets, biologists generally deduce reticulation events indirectly using species tree inference tools. In this context, we present a new Bayesian method, called SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW, dedicated to phylogenetic network inference. Our method is competitive in terms of execution speed with respect to its competitors. This speed gain enables us to consider more complex evolution scenarios during Bayesian analyses. When applied to rice genomic data, SO_SCPLOWNAPPC_SCPLOWNO_SCPLOWETC_SCPLOW suggested a new evolution scenario, compatible with the existing ones: it posits cAus as the result of an early combination between the Indica and Japonica lineages, followed by a later combination between the cAus and Japonica lineages to derive cBasmati. This accounts for the well-documented wide hybrid compatibility of cAus.