Factors affecting the efficiency of molecular species delimitation in a species-rich insect family

In the context of global biodiversity loss, molecular species delimitation approaches can be very useful for accelerating species discovery through DNA taxonomy and inventory through DNA metabarcoding. In this study, the effect of some intrinsic factors on the efficiency of various single-marker species delimitation methods (fixed and variable nucleotide distance thresholds, ABGD, ASAP, GMYC, mPTP) was tested on more than 90 empirical data sets, derived from a set of 7,237 COI sequences attributed to 542 leaf beetles species (Coleoptera: Chrysomelidae). The considered factors were: (i) the number of haplotypes per species (as a proxy for genetic diversity), (ii) the geographic distance among conspecific collection localities (as a proxy of sampling width), (iii) the difficulty related to morphological identification of species, and (iv) the taxonomic rank. Distance-based methods, with on average more than 70% of match with morphological identification, outperformed those relying on phylogenetic trees, with less than 59%. A high number of haplotypes per species was found to have a negative effect on delimitation efficiency, whereas large geographic distances within species had a positive effect. All methods delimitations (except for GMYC) were significantly affected by the presence of species that are difficult to be identified, decreasing their efficiency. Finally, the only method influenced by the taxonomic rank of the data set was GMYC, showing lower efficiency in data sets at the genus than at higher levels. The observed biases we highlighted affecting efficiency could be accounted for when developing input data sets for species delimitation analyses to obtain a more reliable representation of biological diversity.

Automated summary

This study tested the efficiency of various single-marker species delimitation methods on leaf beetles and found that factors such as number of haplotypes, geographic distance, difficulty in species identification, and taxonomic rank can affect efficiency. Distance-based methods outperformed phylogenetic tree methods, and a higher number of haplotypes per species had a negative effect on delimitation efficiency. The presence of species that are difficult to identify also significantly decreased the efficiency of delimitation methods.