Expanding anchored hybrid enrichment to resolve both deep and shallow relationships within the spider tree of life

Abstract

Background: Despite considerable effort, progress in spider molecular systematics has lagged behind many other comparable arthropod groups, thereby hindering family-level resolution, classification, and testing of important macroevolutionary hypotheses. Recently, alternative targeted sequence capture techniques have provided molecular systematics a powerful tool for resolving relationships across the Tree of Life. One of these approaches, Anchored Hybrid Enrichment (AHE), is designed to recover hundreds of unique orthologous loci from across the genome, for resolving both shallow and deep-scale evolutionary relationships within non-model systems. Herein we present a modification of the AHE approach that expands its use for application in spiders, with a particular emphasis on the infraorder Mygalomorphae.

Results: Our aim was to design a set of probes that effectively capture loci informative at a diversity of phylogenetic timescales. Following identification of putative arthropod-wide loci, we utilized homologous transcriptome sequences from 17 species across all spiders to identify exon boundaries. Conserved regions with variable flanking regions were then sought across the tick genome, three published araneomorph spider genomes, and raw genomic reads of two mygalomorph taxa. Following development of the 585 target loci in the Spider Probe Kit, we applied AHE across three taxonomic depths to evaluate performance: deep-level spider family relationships (33 taxa, 327 loci); family and generic relationships within the mygalomorph family Euctenizidae (25 taxa, 403 loci); and species relationships in the North American tarantula genus Aphonopelma (83 taxa, 581 loci). At the deepest level, all three major spider lineages (the Mesothelae, Mygalomorphae, and Araneomorphae) were supported with high bootstrap support. Strong support was also found throughout the Euctenizidae, including generic relationships within the family and species relationships within the genus Aptostichus. As in the Euctenizidae, virtually identical topologies were inferred with high support throughout Aphonopelma.

Conclusions: The Spider Probe Kit, the first implementation of AHE methodology in Class Arachnida, holds great promise for gathering the types and quantities of molecular data needed to accelerate an understanding of the spider Tree of Life by providing a mechanism whereby different researchers can confidently and effectively use the same loci for independent projects, yet allowing synthesis of data across independent research groups.

Keywords: Anchored Hybrid Enrichment; Anchored phylogenomics; Arachnida; Araneae; Conserved regions; Mygalomorphae; Targeted sequence capture; Ultraconserved elements.

Fig. 1

Assemblies for each taxon group and specimen (a = Araneae, b = Euctenizidae, c = Aphonopelma), used to produce the homologous sequence sets. Each point represents one consensus sequence, with the length (bp) on the y-axis and the sequences grouped by sample on the x-axis

Fig. 2

Parsimony Informative Characters (PICs) for each taxon dataset and each locus. a represents the number of PICs in each locus. b represents the correlation between the length of a locus and the number of PICs in that locus

Fig. 3

The distribution of Parsimony Informative Characters (PICs) for the three taxon groupings (a = Araneae, b = Euctenizidae, c = Aphonopelma) and their Anchored Hybrid Enrichment (AHE) loci. The zero position is the center of the anchor sequence for all loci. Colored dots indicate how many total loci had at least one PIC at that nucleotide position (left y-axis). Grey dots indicate the distribution of locus length by plotting the number of loci (right y-axis) with any nucleotide at that position. The figures are plotted on the same y-axis scale to more easily compare the levels of variation across the taxonomic scales

Fig. 4

a - Maximum Likelihood analysis of the 33 Araneae taxa concatenated supermatrix. The dataset comprises 327 loci and 67,870 bp. Black circles denote 100 % bootstrap support; black squares denote bootstrap support between 99-80 %; white squares denote bootstrap support less than 80 %. b - ASTRAL species tree inference based on the Maximum Likelihood inferred individual gene trees from the 327 loci Araneae dataset. ASTRAL analyzes unrooted gene trees; tree was subsequently rooted at the branch leading to the Liphistiidae outgroup. ASTRAL node support values = support based on the RAxML bootstrap support from all trees and all loci. Our preference of the supermatrix inferred phylogeny is likely the more appropriate evaluation of the data at this timescale. Lineages with a different placement between the two outcomes are indicated in yellow. Inset key and colors denote certain historical taxonomic groupings recovered with our sampling. Photographs illustrate a generalized spider lineage corresponding to that region of the phylogeny

Fig. 5

a - ASTRAL species tree inference based on the Maximum Likelihood inferred individual gene trees from the 25 taxa Euctenizidae family/genus level dataset, comprising 403 loci and 133,614 bp. ASTRAL analyzes unrooted gene trees; tree was subsequently rooted at the branch leading to the Idiopidae outgroup. ASTRAL node support values = support based on the RAxML bootstrap support from all trees and all loci. b - Maximum Likelihood analysis of the Euctenizidae concatenated supermatrix. Black circles denote 100 % bootstrap support; black squares denote bootstrap support between 99-80 %; white squares denote bootstrap support less than 80 %. The species tree inference is likely the more appropriate evaluation of the data at this timescale. Inset key and colors denote certain historical taxonomic groupings recovered with our sampling. The non-monophyly of the morphological Atomarius and Hesperus species groups is identified. Novel taxa are denoted by an asterisk. Photographs illustrate a generalized spider lineage corresponding to that region of the phylogeny

Fig. 6

a - ASTRAL species tree inference based on the Maximum Likelihood inferred individual gene trees from the 83 taxa Euctenizidae family/genus level dataset, comprising 581 loci and 334,436 bp. ASTRAL analyzes unrooted gene trees; tree was subsequently rooted at the branch leading to the non-Aphonopelma outgroup taxa. ASTRAL node support values = support based on the RAxML bootstrap support from all trees and all loci. b - Maximum Likelihood analysis of the Aphonopelma concatenated supermatrix. Black circles denote 100 % bootstrap support; black squares denote bootstrap support between 99-80 %; white squares denote bootstrap support less than 80 %. The species tree inference is likely the more appropriate evaluation of the data at this timescale. Inset key and colors denote taxonomic groupings recovered with our sampling. All genealogically exclusive species [79] are identified with a grey bar; A. iodius, a paraphyletic species as presently defined, is identified by the black boxes. Novel taxa are denoted by an asterisk. Lineages with a different placement between the two outcomes are indicated in yellow. Photographs illustrate the Aphonopelma species corresponding to that region of the phylogeny