The Utility of Genome Skimming for Phylogenomic Analyses as Demonstrated for Glycerid Relationships (Annelida Glyceridae)

摘要

Glyceridae (Annelida) are a group of venomous annelids distributed worldwide from intertidal to abyssal depths. To trace the evolutionary history and complexity of glycerid venom cocktails, a solid backbone phylogeny of this group is essential. We therefore aimed to reconstruct the phylogenetic relationships of these annelids using Illumina sequencing technology. We constructed whole-genome shotgun libraries for 19 glycerid specimens and 1 outgroup species (Glycinde armigera). The chosen target genes comprise 13 mitochondrial proteins, 2 ribosomal mitochondrial genes, and 4 nuclear loci (18SrRNA, 28SrRNA, ITS1, and ITS2). Based on partitioned maximum likelihood as well as Bayesian analyses of the resulting supermatrix, we were finally able to resolve a robust glycerid phylogeny and identified three clades comprising the majority of taxa. Furthermore, we detected group II introns inside the cox1 gene of two analyzed glycerid specimens, with two different insertions in one of these species. Moreover, we generated reduced data sets comprising 10 million, 4 million, and 1 million reads from the original data sets to test the influence of the sequencing depth on assembling complete mitochondrial genomes from low coverage genome data. We estimated the coverage of mitochondrial genome sequences in each data set size by mapping the filtered Illumina reads against the respective mitochondrial contigs. By comparing the contig coverage calculated in all data set sizes, we got a hint for the scalability of our genome skimming approach. This allows estimating more precisely the number of reads that are at least necessary to reconstruct complete mitochondrial genomes in Glyceridae and probably non-model organisms in general.