Phylogenomic analysis of trichomycterid catfishes (Teleostei: Siluriformes) inferred from ultraconserved elements

Abstract

The family Trichomycteridae is one of the most diverse groups of freshwater catfishes in South and Central America with eight subfamilies, 41 genera and more than 300 valid species. Its members are widely distributed throughout South America, reaching Costa Rica in Central America and are recognized by extraordinary anatomical specializations and trophic diversity. In order to assess the phylogenetic relationships of Trichomycteridae, we collected sequence data from ultraconserved elements (UCEs) of the genome from 141 specimens of Trichomycteridae and 12 outgroup species. We used a concatenated matrix to assess the phylogenetic relationships by Bayesian inference (BI) and maximum likelihood (ML) searches and a coalescent analysis of species trees. The results show a highly resolved phylogeny with broad agreement among the three distinct analyses, providing overwhelming support for the monophyletic status of subfamily Trichomycterinae including Ituglanis and Scleronema. Previous relationship hypotheses among subfamilies are strongly corroborated, such as the sister relationship between Copionodontinae and Trichogeninae forming a sister clade to the remaining trichomycterids and the intrafamilial clade TSVSG (Tridentinae-Stegophilinae-Vandelliinae-Sarcoglanidinae-Glanapteryginae). Monophyly of Glanapteryginae and Sarcoglanidinae was not supported and the enigmatic Potamoglanis is placed outside tridentinae. Unraveling the relationships of major sections of the Tree of Life is one of the most daunting challenges of the evolutionary biology. Massively parallel DNA sequencing (so-called Next-gen sequencing) is a promising tool that is helping to resolve the interrelationships of longstanding problematic taxa 1-4. One of the most common classes of phylogenomic methods involves the sequence capture of nuclear regions in the flanks and cores of the ultraconserved elements (UCEs) 2. The more variable flanking UCE regions allow a better resolution of nodes across a broad range of evolutionary timescales in a given phylogeny 2. As variation in the flanks increases with distance from the core UCE, this combined approach displays a balance between having a high enough substitution rate while minimizing saturation, thus providing information for estimating phylogenies at multiple evolutionary timescales 2,3. Recent studies of actinopterygians 5 , flatfishes 6 , cichlids 7 , ostariophysan 8 , acanthomorphs 9 , Loricariidae 10 , knifefishes 11 , among other vertebrates groups 3,12 , have shown that UCEs are excellent markers for phylogenetic studies because of their ubiquity among taxonomic groups 13 , low degrees of paralogy 14 , and low saturation 3. According to Gilbert et al. 15 , the phylogenetic informativeness of the combined flank and core regions of UCEs outperfoms protein-coding genes used in multilocus studies. Additionally, phylogenomic approaches are characterized by their potential to collect data from at least one order of magnitude more loci than the traditional sequencing techniques applied to protein-coding le...

Repeated colonization of caves leads to phenotypic convergence in catfishes (Siluriformes:Trichomycterus) at a small geographical scale

“…1A). However, the vast geographical distances separating them and overall congruence between geography and phylogeny in the 205 group (Ochoa et al, 2017;Katz et al, 2018;Ochoa et al, 2020) suggests that species of Trichomycterus have independently colonized caves multiple times. In addition, it remains possible that cave species from Santander which we did not sample may have independently colonized these environments (Fig.…”

“…The importance of more extensive sampling to better understand patterns of cave colonization and evolution in our study system is illustrated by comparing our results to those implied by previous phylogenetic studies of trichomycterids. Based on work showing that two cave species lacking eyes and pigmentation (T. sandovali and T. rosablanca) occupied similar phylogenetic positions in analyses with partly overlapping -and incomplete-255 sampling (Mesa S. et al, 2018;Ochoa et al 2017;Ochoa et al 2020), it appeared likely that cave species from Santander had a single origin. Our phylogeny was generally congruent with trees based on thousands of loci including less taxa (Ochoa et al 2017;Ochoa et al 2020), yet only by sampling multiple additional populations we were able to determine (1) that T. sandovali and T. rosablanca belong to different clades, implying separate colonization of caves; and (2) that more than one event of cave colonization may have occurred among T. 260 sandovali and its close relatives.…”

“…2018;Castellanos-Morales, 2018;Mesa S. et al, 2018). Trichomycterus is considered a non-diagnosable, non-monophyletic taxon where species failing to meet diagnoses for other genera are often placed (Ochoa et al, 2017;Katz et al, 2018;Ochoa et al, 2020). One of the main issues complicating the taxonomy of Trichomycterus is their simplified morphology, which features relatively few 95 characters to work with.…”

“…All samples were deposited at IAvH-P. We also analyzed specimens collected in different drainages in Santander in 2018 during the "Santander Bio" expeditionary project (Torres & Quiñones, 2019), several of which likely correspond to undescribed species of 120 Trichomycterus (Table 1). Our sampling also included Nematogenys inermis (Nematogenyidae), Copionodon pecten (Copionodontinae), and Trichogenes longipinnis (Trichogeninae), which we included as outgroups for phylogenetic analyses based on previous studies (Ochoa et al, 2017;Ochoa et al, 2020).…”

“…Previous phylogenetic studies of the family Trichomycteridae have considered only two cave species, both from Santander: Trichomycterus sandovali Ardila-Rodríguez, 2006 was sampled by Ochoa et al (2017Ochoa et al ( , 2020 whereas 75 T. rosablanca Mesa, Lasso, Ochoa & DoNascimiento, 2018 was included in analyses by Mesa S. et al (2018). Because in those studies each cave species was independently found to be sister to the surface species Eremophilus mutisi Humboldt, 1805, it is possible that T. sandovali and T. rosablanca are close relatives, which would imply a single cave colonization event by Trichomycterus in Santander.…”