How is possible to establish sequence identity cut-off values for species approximation is a global problem for molecular ecologist and taxonomists. In the new molecular era, where high-throughput data is the rule rather than the exception, and genomes seem replace the classic fingerprinting based on single genes, how we can solve this dilemma?
I think the best way to get identification is to consult a taxonomist, if the expertise in a particular taxon is available. There is no universal cut-off for all taxa. Inter- and intraspecific genetic distances vary greatly among taxa. Hope this paper may help: "The use of mean instead of smallest interspecific distances exaggerates the size of the barcoding gap and leads to mis-identification".
http://www.academia.edu/1104441/The_use_of_mean_instead_of_smallest_interspecific_distances_exaggerates_the_size_of_the_barcoding_gap_and_leads_to_misidentification_Systematic_Biology_
Guanyang,
thank you for your reply and the very interesting article.
If you have a very small mean intra-specific variability for some species and a very high inter-specific variability between those species you will have a huge gap among this case of congeneric species. The two distances then are well separated and species, as separated taxa, are strongly supported. As you argued, if I use the smallest inter-specific distance instead of the mean it would be surely lower than the mean intra-specific and then the overlap should obviously result smaller…
My doubt concerned how meaningful is the smallest distance if this smallest distance is not generalizable, because the sequences produced until now represent only a small portions of the existing species? In this case do a mean value of what is already sequenced better explain the limit for identification than an hypothetic smallest difference between 2 congeneric species even represented by few sequences?
For certain organisms, such microscopic or macroscopic fungi, is impossible to find morphological species boundaries also if I will consult the most expert taxonomist. I think that molecular tools are then an obvious solution to this problem, it's just fundamental to use the correct way to use and develop it...
I’d love this brainstorming continue,
Gian