Predicting genetic biodiversity in salamanders using geographic, climatic, and life history traits

pattern one is attempting to examine is especially important in meta-analyses, as different pat- terns emerge at different scales [89]. Conclusions Here, we chose to utilize biodiversity data from phylogatR (i.e., genetic data for which directly associated specimen locality information is available) to avoid potential discrepancies between the distribution of the genetic and geographic data analyzed. By doing so we hoped to gain a more fine-grain understanding of how species genetic diversity is influenced by geographic and environmental factors [23]. However, making this choice significantly decreased the amount of data available and led to a greatly reduced dataset. Our study included 1676 DNA barcoding sequences from the genes COI and cytb (768 and 908 sequences each, respectively). However, a 3/31/23 search of GenBank for salamander barcoding sequences from the genes COI and cytb returned a total of 17097 sequences (4468 and 12629 sequences each, respec- tively; see S6 Data). Similarly, while we were able to obtain 1676 occurrence records tied to the genetic sequences used in this study, a GBIF search for geographic occurrences tied to sala- mander preserved specimens and material samples returned 675243 records (see S6 Data). This study highlights the lack of genetic data with easily-associated geographic information. Despite limitations in dataset size and geographic coverage, our framework effectively iden- tified salamander species likely to contain undescribed genetic diversity with agreement across

multiple delimitation methods. These species likely represent good candidates for further taxo- nomic evaluation. While we were unable to pinpoint a specific predictor variable as the most important for predicting undescribed diversity, our findings suggest that hidden diversity in salamanders is likely higher in species with broad geographic ranges characterized by signifi- cant climatic variability. This insight serves as a starting point for future integrative taxonomic work and underscores that much diversity remains undiscovered. Although our study was constrained by data availability, the framework we used could further elucidate these relation- ships with access to more comprehensive genetic and geographic data, highlighting the crucial importance of such data in biodiversity studies. The numerous benefits of making biological data more broadly available have been repeat- edly demonstrated [90]. And recent years have seen a significant increase in the amount of available specimen and biodiversity data. The utility of these data to address large scale pat- terns of biodiversity, such as those examined in this study, is enhanced by our ability to inte- grate and synthesize data across different data sources, types, and taxonomic groups [91]. Our study highlights the importance of not just making these data available but making them avail- able in a way that is standardized and will facilitate integration and re-use for future genera- tions to come (e.g., [92,93]). Supporting information

S1 File. (Contains Figure S1: Distribution of missing data in the salamander trait database; Figure S2: Distribution of imputed trait data; Figure S3: Correlation matrix of predictor variables; Figure S4: Variable importance for predictive models). (DOCX) S1 Table. Comparison of salamander data available for each loci on the phylogatR data- base. (DOCX) PLOS ONE Predicting hidden lineages in salamanders PLOS ONE | https://doi.org/10.1371/journal.pone.0310932 October 18, 2024 15 / 20