Exploring the impact of morphology, multiple sequence alignment and choice of optimality criteria in phylogenetic inference: a case study with the Neotropical orb-weaving spider genus Wagneriana (Araneae: Araneidae)

Abstract

We present a total evidence phylogenetic analysis of the Neotropical orb-weaving spider genus Wagneriana and discuss the phylogenetic impacts of methodological choices. We analysed 167 phenotypic characters and nine loci scored for 115 Wagneriana and outgroups, including 46 newly sequenced species. We compared total evidence analyses and molecular-only analyses to evaluate the impact of phenotypic evidence, and we performed analyses using the programs POY, TNT, RAxML, GARLI, IQ-TREE and MrBayes to evaluate the effects of multiple sequence alignment and optimality criteria. In all analyses, Wagneriana carimagua and Wagneriana uropygialis were nested in the genera Parawixia and Alpaida, respectively, and the remaining species of Wagneriana fell into three main clades, none of which formed a pair of sister taxa. However, sister-group relationships among the main clades and their internal relationships were strongly influenced by methodological choices. Alignment methods had comparable topological effects to those of optimality criteria in terms of ‘subtree pruning and regrafting’ moves. The inclusion of phenotypic evidence, 2.80–3.05% of the total evidence matrices, increased support irrespective of the optimality criterion used. The monophyly of some groups was recovered only after the addition of morphological characters. A new araneid genus, Popperaneus gen. nov., is erected, and Paraverrucosa is resurrected. Four new synonymies and seven new combinations are proposed.

Total evidence and sensitivity phylogenetic analyses of egg‐brooding frogs (Anura: Hemiphractidae)

“…Regarding the resolution of optimal trees, we found a positive or neutral effect in five out of six analyses (two and three, respectively), while it decreased tree resolution only in SAPg analysis. This generally positive or neutral effect is commonly reported in the literature (Nylander et al, 2004;de S a et al, 2014;Chakrabarty et al, 2017;Mirande, 2017;Koch and Gauthier, 2018;Martin et al, 2018;S anchez-Pacheco et al, 2018;Cabra-Garc ıa and Hormiga, 2020). Another positive result (at least in our view) is that the addition of the phenotypic characters increased congruence up to 12% among the optimal topologies of the different analyses in 80% of the comparisons.…”

“…Arguably, most attention has been directed to differences caused by choice of optimality criterion, on both simulated and empirical studies (e.g., Felsenstein, 1978; Huelsenbeck and Hillis, 1993; Huelsenbeck, 1995, 1997; Yang, 1996a; Siddall and Kluge, 1997, 1999; Siddall, 1998; Wiens and Servedio, 1998; Siddall and Whiting, 1999; Huelsenbeck et al, 2001; Pol and Siddall, 2001; Swofford et al, 2001; Leaché and Reeder, 2002; Huelsenbeck and Lander, 2003; Kolaczkowski and Thorton, 2004; Yu et al, 2008; Puttick et al, 2017). However, a number of other factors could be equally influential on the inferred phylogenetic hypotheses, such as alignment of sequence data, indel coding, character weighting, model selection, heuristic tree‐search strategies, and representation of results such as optimal trees versus a variety of consensus trees (Chippindale and Wiens, 1994; Simon et al, 1994; Yang et al, 1994; Milinkovitch et al, 1996; Yang, 1996b; Morrison and Ellis, 1997; Simmons and Ochoterena, 2000; Simmons et al, 2001; Brandley et al, 2005; Ogden and Rosenberg, 2006; Zwickl, 2006; Brown and Lemmon, 2007; Kumar and Filipski, 2007; McGuire et al, 2007; Simmons et al, 2007; Li et al, 2008; Wong et al, 2008; Ward et al, 2010; Sanderson et al, 2011; Goloboff, 2014; Cabra‐García and Hormiga, 2020). Furthermore, in the context of total evidence analyses, combined matrices of molecular and non‐molecular data have attracted discussion, with emphasis on the role of phenotypic characters in large molecular datasets (Giribet, 2015; Lee and Palci, 2015; Chakrabarty et al, 2017; Koch and Gauthier, 2018; Martin et al, 2018).…”

Ocrepeira klamt sp. n. (Araneae: Araneidae), a novel spider species from an Andean páramo in Colombia

“…dissections of O . darlingtoni in [ 27 ]) with our cleared view ( Fig 3D–3F ) suggests that the morphology of the copulatory ducts could be highly informative to distinguish species.…”

“…n. thus provides a valuable character to be used in conjunction with morphology for species identification. Being the third ever reported sequence from the genus Ocrepeira [26,27], it adds to the growing reference database of COI sequences for spiders [72,76].…”

“…To date, this orb-weaving genus includes 67 recognized species [13], which are found exclusively in the Americas and prevailingly inhabit high mountain environments, where the geographic isolation causes a high degree of endemism [25]. In providing the so far third ever sequence of the barcoding gene cytochrome c oxidase subunit I (COI) from the genus Ocrepeira [26,27], and exploring the phylogenetic position of the new species, we complement the morphological description with molecular data.…”