spider: an R package for DNA barcoding

Abstract

We present the R package \"spider\" (SPecies IDentity and Evolution in R) that provides customisable, user-friendly functions for calculating a diverse range of summary statistics essential for DNA barcoding, taxonomy, and analysis of species-level evolution. Until now, analyses of DNA barcode data have been carried out on multiple platforms, and often using piecemeal software applications and exploration of large datasets is limited by the availability of pre-programmed functions. The statistical programming language R, is a powerful, flexible and free environment for the analysis of a wide range of data, including nucleotide sequences. The spider package includes a novel sliding-window method for objectively evaluating discriminatory power across mini-barcode regions, an approach recommended for designing primers in degraded-DNA studies. In addition, spider also provides a suite of the more standard statistics for use in DNA barcoding, including: identification success rates using nearest neighbour, distance threshold, or neighbour-joining criteria; calculation of intraspecific and nearest non-conspecific neighbour distances; descriptive statistics (e.g. haplotype frequency); cumulative (false positive/negative) error analysis; inter-group permutation tests; character-based population aggregation analyses. We also include functions for producing publication-quality graphics of frequently used analyses, such as saturation plots and illustrations of the barcoding gap. Spider is available as a stable version from the Comprehensive R Archive Network (CRAN; http://cran.r-project.org) and as a development version from R-Forge (http://spider.r-forge.r-project.org). It is intended that further developments of the package will include additional methods for the analysis of DNA, morphological and spatial data.