Assessing the Darwinian shortfall in the knowledge of freshwater fishes

Background/Question/Methods

Multiple approaches have increasingly integrated phylogenetic information into ecology and conservation science, such as defining conservation priorities, estimating ecosystem services and processes, understanding ecological community patterns and processes, and inputting functional traits when basic information is unavailable. Nonetheless, the Darwinian shortfall of biodiversity knowledge (i.e., lack of fully resolved phylogenies, poor resolution of branch lengths and time calibration, and unknown evolutionary models for ecological traits) hampers a widespread application of phylogenetic methods in ecology and conservation.Herein, we present the PD deficit, a new measure for estimating the Darwinian shortfall to prioritize future research efforts. To calculate PD deficit, one should first assemble on a fully resolved phylogeny using a supertree approach, in which taxa are added on a backbone tree based on the available information (published phylogenies and taxonomic hierarchy). Then, PD deficit is calculated as the sum of the branch lengths of the inserted species into the backbone tree divided by the sum of the branch lengths of the supertree. We exemplify the PD deficit using an open global database of 14,886 freshwater fishes inhabiting 3,119 basins. All analyses were performed in the R environment using the FishPhyloMaker package.

Results/Conclusions

The backbone phylogeny comprised 6,418 species from the global pool of species. The FishPhyloMaker algorithm inserted 11,569 species in the backbone phylogeny and generated a supertree encompassing 14,705 tips. The highest number of performed insertions was for the Afrotropical (1,942) and Neotropical (3,804) regions, while the lowest was for the Nearctic (146) and Palearctic (547) regions. The PD deficit ranged from 0 (all species included in the backbone phylogeny) to 1 (all species inserted). The lowest values were concentrated in the Nearctic, Palearctic and Australasian regions, while the highest were concentrated in the Afrotropics. Globally distributed lineages such as Cyprinodontiformes and Siluriformes exhibited lower PD deficit values than Neotropical and Afrotropical ones (e.g., Characiformes). Observed patterns are explained by a myriad of factors, but primarily by sampling and scientific investment biases towards the Nearctic and Palearctic and the high diversity in tropical systems. Beyond general patterns, our tool provides an alternative on how biodiversity science could address future efforts to fill knowledge gaps. Nonetheless, the tool should further integrate proportional values for clade richness, map all the freshwater and marine fishes, and even expand for other clades.