A rough guide that I have been told of is 90% for the same genus and 95% for the same species when applied to the 16S rDNA gene when analysing bacterial taxonomy, but it will inevitably depend on the life-form you are looking at and the reference gene/s you are using.

Just to be clear, if you are wishing to assign genus/species designations then you should be using 16s sequence and not just your gene of interest and its homology to reference genes. The reason for this is that different genes will change at very different rates, I've seen genes from two close species in the same genera have significantly diverged and other genes be essentially identical. Also you could have had horizontal gene transfer occurring which will also mask the relatedness (or lack) of the hosts. 

In order to identify species/genera similarities/designations you have to work with rRNA 16S or rRNA 18S.

Hi ..thank you very much  for the answers..yes..i am working with bacteria and archaea..i am using 16s rrna seq for taxonomy..

So it is 90% for same genus and 95% for same species..right?

For same species, some groups used to mention the percentage value as 97. 0.

So which value we can follow? 

16S rDNA sequence identity score is not sufficient for genus delineation. There are cases of bacterial species with >99% 16S rDNA sequence identity assigned to different genera. 16S rDNA sequence identity score should always be complemented by DNA-DNA hybridization score.

Besides, phylogeny is just one of the four aspects of the living thing; one should also consider the remaining three: morphology, physiology and ecology.

---

No single, objective value for 16S rDNA sequence identity score exists for the species or genus delineation.

I need a reliable source for this وthe classification of bacteria was based on the 16S rRNA gene, and therefore the similarity limit is currently accepted at