Genetic legacies of mega-landslides: Cycles of isolation and contact across flank collapses in an oceanic island

Catastrophic flank collapses are recognised as important drivers of insular biodiversity dynamics, through the disruption of species ranges and subsequent allopatric divergence. However, little empirical data supports this conjecture, with their evolutionary consequences remaining poorly understood. Using genome-wide data within a population genomics and phylogenomics framework, we evaluate how mega-landslides have impacted evolutionary and demographic history within a species complex of weevils (Curculionidae) within the Canary Island of Tenerife. We reveal a complex genomic landscape, within which individuals of single ancestry were sampled in areas characterised by long-term geological stability, relative to the timing of flank collapses. By contrast, individuals of admixed ancestry were almost exclusively sampled within the boundaries of flank collapses. Estimated divergence times among ancestral populations aligned with the timings of mega-landslide events. Our results provide the first evidence for a cyclical dynamic of range fragmentation and secondary contact across flank collapse landscapes, with support for a model where this dynamic is mediated by Quaternary climate oscillations. The context within which we reveal climate and topography to interact cyclically through time to shape the geographic structure of genetic variation, together with related recent work, highlights the importance of topoclimatic phenomena as an agent of diversification within insular invertebrates.

Methods
Representative geographical sampling from the Tenerife species of the L. tessellatus complex was achieved by complementing previous sampling from Faria et al. (2016) and García-Olivares et al. (2017) with 74 specimens from 61 new localities. This additional sampling effort gave rise to a total of 126 individuals from 102 sites in Tenerife (Table S1). We also included 5 individuals from the L. tessellatus complex, belonging to a sister clade from the nearby island of Gran Canaria (García-Olivares et al., 2019), as an outgroup.

We extracted DNA using the Qiagen DNeasy Blood & Tissue kit following the manufacturer’s instructions. DNA was processed using the double-digestion restriction-site associated DNA sequencing protocol (ddRADseq, Peterson et al., 2012) as described in Mastretta-Yanes et al. (2015) and García-Olivares et al. (2019). In brief, DNA was digested with the restriction enzymes MseI and EcoRI (New England Biolabs, Ipswich, MA, USA). Genomic libraries were pooled at equimolar ratios and size selected for fragments between 200-250 base pairs (bp) and, then, sequenced in a single-end 100-bp lane on an Illumina HiSeq2500 platform (Lausanne Genomic Technologies Facility, University of Lausanne, Switzerland).

We first used FASTQC version 0.11.7 (Andrews, 2010) to quality check raw reads. Then, raw sequences were demultiplexed, quality filtered, and de novo assembled using IPYRAD version 0.9.81 (Eaton & Overcast, 2020). Only reads with unambiguous barcodes were retained (max_barcode_mismatch) and a stricter filter was applied to remove Illumina adapter contamination (filter_adapters). We converted base calls with a Phred score <20 into Ns and discarded reads with >5 Ns (max_low_qual_bases). Afterward, we clustered the retained reads within- and across samples considering a threshold of sequence similarity of 85% (clust_threshold) and discarded those clusters with a minimum coverage depth of less than 5 (mindepth_majrule). Resulting in loci shorter than 35 bp (filter_min_trim_len), containing one or more heterozygous sites across more than 50% of individuals (max_shared_Hs_locus), and showing more than 20% polymorphic sites (max_SNPs_locus) were discarded. In a final filtering step, we only retained loci that were present in at least 80% of the samples (min_samples_locus), which yielded a total of 4987 and 6018 unlinked SNPs, when including and excluding the outgroup, respectively. Optimal parameter tuning in ipyrad is performed following results from the sensitivity analyses conducted by García-Olivares et al. (2019) for the Laparocerus tessellatus species complex.