Key Biodiversity Areas (KBAs) are sites that contribute significantly to the global persistence of biodiversity. Distinct genetic diversity has been introduced as one of the metrics to estimate whether a site holds a threshold proportion of a species’ global genetic diversity during the KBA identification process. However, genetic data has so far not been used due to the lack of thoroughly tested methods and guidance. We tested the applicability of Analyses of Molecular Variance (AMOVA), allelic overlap, the diversity index Simpson's λ, Δ+, D_est, and effective population size (N_e) for identification of KBAs. We conclude that Δ+, a measure that has originally been developed to measure taxonomic distinctness of biotic communities, performs best in the context of KBA identification reflects the unique nature of a species’ genetic diversity, is based on simple allele frequencies, and can be easily applied and calculated. AMOVA, N_e, allelic overlap, and our modified version of λ, were difficult to apply, interpret, or both. D_est is easily applied for measuring genetic distinctiveness but not genetic diversity. For this reason, it may not be suitable for prioritizing areas for the long-term protection of the species.

Here, we deposited additional information on which methods, and how they were calculated on which data sets, including references. We included additional information on how these methods performed. Moreover we included raw reads, an already processed .str file and additional information of additional data set we created to add a case study to our publication. The code we used for our analyzes can be found on GitHub and Zenodo. For more information about the code and the detailed procedure, view the associated publication, the readme on GitHub, and the Supplemental_Information.pdf file we publish here.

As both SNP and microsatellite datasets are commonly used to analyze intraspecific genetic variance, we tested the performance of our chosen analytical approaches on 30 published diploid datasets, of which 15 used SNPs (with an average of 184 SNP loci) and 15 microsatellite datasets (with an average of 31 microsatellite loci). Each dataset was analyzed with six methods: AMOVA, allelic overlap , Δ⁺, D_est , λ corrected for sample size , and N_e. To apply all six methods, an R project was created that makes use of many packages that facilitate displaying results and working with genetic data and tables.

For better comparability between the six methods, all datasets were prepared in the same way. Sites with fewer than 30 individuals were removed from the analysis. Individuals with > 20% missing data were removed from the dataset. For loci with missing genotypes, the missing allele counts were replaced with the mean of the observed alleles at that locus across all individuals in the dataset.

To explore similarities between the different approaches, correlations between the results of all six methods, allelic overlap, AMOVA, Δ⁺, D_est, N_e, and λ_cor, were calculated in R. Correlations with allelic richness were additionally calculated. Two outliers were removed from AMOVA results. A Kendall correlation was chosen. Correlations between allelic overlap, allelic richness, AMOVA, Δ⁺, D_est, N_e, and λ_cor, are based on different sample sizes, since N_e could not be calculated for some areas.

We tested all results against two KBA criteria, A1b (> 1 % of the global distinct genetic diversity occurs at this site) and B1 (> 10 % of the global distinct genetic diversity occurs at this site). For each of the six methods, the proportion of distinct genetic diversity was calculated as the simple proportion of distinct genetic diversity at each location of the sum of all locations, as proposed by the KBA standard for AMOVA results. Areas lacking N_e were allocated the median of remaining areas to enable the application of KBA criteria without inflating the proportion of the remaining sites.

To illustrate the results and coverage of different genetic clusters, Structure analyses were conducted in addition to the calculation of the five metrics for two case studies: the Chinook salmon (Oncorhynchus tshawytscha; Gomez-Uchida et al. 2019) as well as a hitherto unpublished dataset of the Tenerife Short-winged Bush-cricket (Ariagona margaritae Kraus, 1892). The site selection was based upon KBA criterion B1. The results were processed in R using several packages. The maps were created using ArcGIS Pro.

To create the Tenerife Short-winged Bush-cricket dataset specimens were collected 2010–2023 on Tenerife and El Hierro. DNA was extracted using the Qiagen DNeasy® Blood & Tissue kit. ddRADseq libraries were prepared for paired-end sequencing on a High-Output Flow cell of an Illumina NextSeq platform (2 x 75bp). Stacks 2.6.6 was used to demultiplex, filter, and trim raw reads to 65bp, create an assembly and a catalogue of loci to finally identify SNPs (n= 64, -p 150, -r 1). Default settings were maintained. Individuals containing more than 20% of missing data were removed from the analysis. The resulting dataset comprised 108 individuals and 5198 loci. No area was excluded from the analysis as each had ≥20 sequenced individuals. The allelic overlap method was omitted for this dataset due to extensive calculation times. For N_e, the smallest possible natural number was added to transform all N_e into positive numbers. Apart from that, this dataset was analyzed in the same way as the previously used datasets.

Changes after Nov 7, 2025: one sentence in Supplemental_information changed, corrected code to be easier to understand, and updated abstract/methods for clarity