PAICE: A new R package to estimate the number of inter‐island colonizations considering haplotype data and sample size

Abstract

Aim: Colonization is a central topic in ecology and one of the cornerstones of island biogeography. Although the evolutionary history of island species is widely studied, the quantification of colonization is particularly challenging because the same area may be colonized multiple times by the same species, whereas initially successful colonization events may eventually be followed by extinction. Nevertheless, an estimation of the actual number of within-archipelago colonization events can be achieved when using adequate sample size and genetic data, which are essential parameters in the inference of colonization success of any species.Location: Canary Islands, Azores and Galápagos Islands.Taxon: Buteo galapagoensis, Croton scouleri, Setophaga petechia aureola and Xylocopa darwini (Galápagos); Canarina canariensis, Cistus monspeliensis, Juniperus cedrus and Olea europaea subsp. guanchica (Canary Islands); and Juniperus brevifolia and Picconia azorica (Azores). Methods:The new R package PAICE uses haplotype (from organelle DNA) sharing and haplotype relationships, and controls for sampling effort to estimate the number of within-archipelago colonization events in island-like systems. PAICE applies a sampling-effort correction based on rarefaction curves of field sampling (number of individuals or populations) and genetic sampling (number of DNA variable positions).The number of colonization events for the 10 insular species were estimated with PAICE and results compared with previous methods.Results: PAICE estimates a number of inter-island colonization events up to an order of magnitude greater than previous methods. Furthermore, PAICE can quantify the colonization events of any study species, in multiple biogeographic contexts and considering sampling size, thus providing a standardized estimate of colonization success. Main conclusions:The new package PAICE provides an estimation of the number of inter-island colonization events (regardless of dispersal routes or rates) based on haplotype data across islands. This new tool will allow gaining new insights on the intensity of long-distance-dispersal events, their drivers and consequences for the assembly of insular faunas and floras.

Phylogenetics and phylogeography ofEuphorbia canariensisreveal an extreme Canarian-Asian disjunction and limited inter-island colonization

“…Finally, we used the argument method = 1 to allow the algorithm to fit the accumulation curve of colonization events with fewer parameters if curves could not be fitted initially. As the outgroup was connected to E. canariensis at two points (two missing haplotypes between haplotypes A and B; see Results), we repeated the analysis considering each of these possible ancestral haplotypes (see Coello et al, 2022 for additional information).…”

“…Given the occurrence of this species in the six major islands of the Canarian archipelago (considering Lanzarote and Fuerteventura as the single palaeoisland of Mahan; Troll and Carracedo, 2016), E. canariensis has experienced at least five inter-islands colonization events based on chorology alone (Arjona et al, 2018;Vargas et al, 2015). The geographic distribution of plastid DNA diversity (Figure 3) indicates a minimum of eight colonization events, and estimates obtained in PAICE considering sampling effort (Coello et al, 2022) suggested c. 20 -50 colonization events (see Results, Figure 4). Results of this study corroborate results of Coello et al (2022) regarding the fact that DPA analysis (c. 13 colonization events with a 95% confidence interval of 9 -19 colonization events) estimates a smaller number of inter-island colonization events than PAICE.…”

“…We estimated the number of inter-island colonization events based on phylogeographic data (see above) using the recently developed method PAICE (Phylogeographic Analysis of Island Colonization Events; Coello et al, 2022). We considered the palaeoisland of Mahan as a single island unit combining Lanzarote and Fuerteventura to avoid confounding within-island range expansion with inter-island colonization events, given that these two islands were connected in the Last Glacial Maximum (Troll and Carracedo, 2016) (see Results).…”

“…In any case, we only included individuals from Fuerteventura, as the species does not seem to be naturally present on Lanzarote. First, we inferred the minimum number of colonization events required to explain the geographic distribution of E. canariensis haplotypes using the colonization function of the PAICE package (Coello et al, 2022). We also calculated asymptotic estimators of colonization events.…”

“…Despite the LDD process is considered critical to allow oceanic island colonization (Nathan, 2006; Traveset et al, 2014), E. canariensis does not display LDD syndromes. Although general patterns of island colonization can be derived from floristic studies and distribution ranges (Arjona et al, 2018; Heleno and Vargas, 2015; Vargas et al, 2012), more detailed inference of colonization patterns for a particular species can be derived from phylogeographic analyses (Coello et al, 2022, 2021). Therefore, a detailed study of inter-island colonization events is required to understand the phylogeographic history of E. canariensis .…”