The definition of "species" varies between organisms. For many mammals the definition of species depends on whether members of one group can breed and produce fertile offspring with members of another. So, for example horses are considered a different species than donkeys because the offspring of cross-breeds (mules) are not fertile. However in other mammal lineages, we consider different groups to be different species even if they can cross. For example, we consider wolves and coyotes to be different species and most people recognize domestic dogs as a separate species from wolves, but wolves and domestic dogs can cross and so can coyotes and domestic dogs.

Many groups of mammals have been geographically isolated for various time periods and are now considered to be different species because they have not interbred for more than 15,000 or 100,000 or several million years but crossbreeding and testing fertility of the offspring either has not been done, or they are known to produce fertile offspring but are still considered to be different species anyway. So deciding how to define "species" is not equal in all types of mammals, and your definition should depend on what purposes you wish to use the designation for.

Hi Paras,

I think the answer to all of your questions is 'yes'. Assuming by 'significance value' you mean posterior probability, it sounds like the BI tree is telling you that your samples are falling into two strongly supported clades. On the other hand, the p-distance (and I would assume also the branch lengths on the tree) is indicating that there is very little genetic distance between the two clades. Taken together, this suggests that while there is clear genetic divergence between two groups within your species, it may be too low to warrant re-classification into two separate species. In addition, it would be interesting to further explore the levels of genetic differentiation using different loci. Is it possible that these two clades could be classified below the species level (i.e., subspecies or varieties)?

Good luck, sounds interesting!

Hi Paras,

There are also nice methods dealing with species delimitation using molecular data. Maybe it would be meaningful to try some of them to take decisions on whether you have one or two species in your dataset. Here I list some of them:

GMYC: You need a ultrametric tree as input (http://species.h-its.org/gmyc/)

PTP: as above, useful when you only have one genomic region (http://species.h-its.org/)

BPP: bayesian method, it can me used also with multiple loci. Use sequences as input (http://abacus.gene.ucl.ac.uk/software.html)

Dissect or Stacey: as for BPP. It is implemented in BEAST (available as a BEAST package in BEAST2)

You can also considered the possibility of using marginal likelihood to see unequivocaly which species delimitation model fits better your data (i.e., 1 species vs. 2-species).

I hope it helps. Good luck!

Salvatore