How much is enough? Sampling intensity influences estimates of reproductive variance in an introduced population
Data files
Jun 05, 2021 version files 21.90 GB
-
EAP21-0107_PedigreeSubsamplingCode.Rmd
-
fastqs.zip
-
McLennan_et_al._Ecological_Applications_metdata.xlsx
Abstract
Conservation introductions to islands and fenced enclosures are increasing as in situ mitigations fail to keep pace with population declines. Few studies consider the potential loss of genetic diversity and increased inbreeding if released individuals breed disproportionately. As funding is limited and post-release monitoring expensive for conservation programs, understanding how sampling effort influences estimates of reproductive variance is useful. To investigate this relationship, we used a well-studied population of Tasmanian devils (Sarcophilus harrisii) introduced to Maria Island, Tasmania. Pedigree reconstruction based on molecular data revealed high variance in number of offspring per breeder and high proportions of unsuccessful individuals. Computational subsampling of 20, 40, 60, and 80% of observed offspring resulted in inaccurate estimates of reproductive variance compared to the pedigree reconstructed with all sampled individuals. With decreased sampling effort, the proportion of inferred unsuccessful individuals was overestimated and the variance in number of offspring per breeder was underestimated. To accurately estimate reproductive variance, we recommend sampling as many individuals as logistically possible during the early stages of population establishment. Further, we recommend careful selection of colonising individuals as they may be disproportionately represented in subsequent generations. Within the conservation management context, our results highlight important considerations for sample collection and post-release monitoring during population establishment.
Methods
This upload contains the fastq files geneated via DArT-seqTM, scripts and metadata used to perform pedigree reconstruction and subsampling analysis presented in the Ecological Applications article "How much is enough? Sampling intensity influences estimates of reproductive variance in an introduced population" (EAP21-0107).
SNP calls were first made using the Stacks/2.0b pipeline from fastq files geneated via DArT-seqTM (uploaded here). The output from Stacks 2.0b (vcf. file) was further filtered in R (script available https://doi.org/10.1186/s12864-019-5806-y). All filtering parameters are presented in the Materials and Methods section of the article.