What is the minimum percentage of identity (blastp) to predict the orthologous genes in different species?

Question

[Kumar]

I would like to predict the core genes/orthologous genes from different species of viruses. I have used roary to do the same, in which I could get least number of core genes while I set the minimum percentage identity of blastp is 30 to 40 (-i 30-40). Because, when I set the percentage of identity above 40 I could not get any core genes. Please some one clarify, is it the right way to do the analysis.

[Joe]

It sounds like you are trying to do core genome analysis among more distantly related bacteria. roary is only appropriate to approximately within genus comparisons (see their documentation).

A blastp identity of 30% is basically nothing. I would be seriously worried if you aren't getting any core genes with an identity over 40%. A typical value for roary would be 90-95% depending on how strict you want to be.

[Kumar]

@Joe Thank you for your valuable information and making me to aware the facts. I will not proceed the analysis with this low percentage of identity value.