Building a Robust, Densely-Sampled Spider Tree of Life for Ecosystem Research

Abstract

Phylogenetic relatedness is a key diversity measure for the analysis and understanding of how species and communities evolve across time and space. Understanding the nonrandom loss of species with respect to phylogeny is also essential for better-informed conservation decisions. However, several factors are known to influence phylogenetic reconstruction and, ultimately, phylogenetic diversity metrics. In this study, we empirically tested how some of these factors (topological constraint, taxon sampling, genetic markers and calibration) affect phylogenetic resolution and uncertainty. We built a densely sampled, species-level phylogenetic tree for spiders, combining Sanger sequencing of species from local communities of two biogeographical regions (Iberian Peninsula and Macaronesia) with a taxon-rich backbone matrix of Genbank sequences and a topological constraint derived from recent phylogenomic studies. The resulting tree constitutes the most complete spider phylogeny to date, both in terms of terminals and background information, and may serve as a standard reference for the analysis of phylogenetic diversity patterns at the community level. We then used this tree to investigate how partial data affect phylogenetic reconstruction, phylogenetic diversity estimates and their rankings, and, ultimately, the ecological processes inferred for each community. We found that the incorporation of a single slowly evolving marker (28S) to the DNA barcode sequences from local communities, had the highest impact on tree topology, closely followed by the use of a backbone matrix. The increase in missing data resulting from combining partial sequences from local communities only had a moderate impact on the resulting trees, similar to the difference observed when using topological constraints. Our study further revealed substantial differences in both the phylogenetic structure and diversity rankings of the analyzed communities estimated from the different phylogenetic treatments, especially when using non-ultrametric trees (phylograms) instead of time-stamped trees (chronograms). Finally, we provide some recommendations on reconstructing phylogenetic trees to infer phylogenetic diversity within ecological studies.

What DNA barcodes reveal: microhabitat preference, hunting strategy and dispersal ability drive genetic variation across Iberian spider species

“…We used the R package 'ade4' (Dray & Dufour, 2007) to generate a Gower dissimilarity matrix to which we applied a principal coordinates analysis (PCoA) to represent the overall functional space including all species. We controlled for the effect of phylogenetic relatedness in statistical models using the time-calibrated phylogeny in the study by Macías-Hern andez et al (2020a). This phylogeny was inferred from six genes and topological constraints based on transcriptome information (Fernandez et al, 2018) for 1450 species of spiders, including those in the present study.…”

Biogeographic position and body size jointly set lower thermal limits of wandering spiders

“…The former could have benefited from higher cold resistance to expand and increase survival of populations in coldest areas, and this might explain the more limited distribution range of D. plantarius , abilities to tolerate cold being an important factor to explain past colonization (Sunday et al., 2011). Nonetheless, a better knowledge of the phylogeny of European fishing spider that is unfortunately still poorly documented (see Macías‐Hernández et al, 2020; Piacentini & Ramírez, 2019; Tanikawa & Miyashita, 2008) would allow us to conclude on the difference in species adaptation to climate by using a phylogenetic comparative method (Blomberg & Garland, 2002; Garland & Adolph, 1994).…”

Dispersal ability and niche breadth influence interspecific variation in spider abundance and occupancy

“…A phylogeny was constructed using a backbone matrix comprising 4570 sequences for 1450 species retrieved from Macías-Hernández et al . [ 55 ], which includes (i) the concatenated gene matrix of Wheeler et al . [ 56 ], (ii) sequences bioinformatically extracted from the original transcriptome data in Fernández et al .…”

“…A region of the 28S rRNA gene was amplified using the primers 28S-O and 28S-B [ 58 , 59 ], using conditions described in Macías-Hernández et al . [ 55 ]. The phylogenetic tree was inferred using maximum-likelihood implemented in the program RAxML-HPC2, remotely run on XSEDE at the CIPRES Science Gateway ( ), with the topology of Fernández et al .…”