Dear Hyphy team,

I am using the Relative Rate Test implementation in Hyphy. To represent my statistically significant results I am making bar graphs to compare the estimates for the synonymous rates from the two species (threeTaxaTree.FirstSpecies.SynRate threeTaxaTree.SecondSpecies.SynRate). Similarly I am making a graph for the non synonymous substitution. Until here all makes sense and it is quite clear which species have relatively higher or lower rates of Syn or Non Syn substitutions. My question is that for a few comparisons in one of the species the estimate of Synonymous rate is 0.

I know that this could happen, but it does not make a lot of sense to me because if I look for the synonymous substitutions between the sequences compared, I find that there are indeed some of them. I know that the results from Hyphy are estimates based on a model, not actual counts, but still I find difficult to understand this 0 in my results.

Here is what I get on the screen for a comparison with these kind of outcome:

(A:1.03047,(B:0.205465,C:0.0640284))         |    6.5395  | 0.0380151 (*)

Here is what I get on the output file for the unconstrained estimates corresponding to this comparison. These are numbers that I am actually using for my graphs and where I found this puzzling 0.

Outgroup, Taxon 1, Taxon 2, threeTaxaTree.FirstSpecies.nonSynRate, threeTaxaTree.FirstSpecies.synRate, threeTaxaTree.SecondSpecies.nonSynRate, threeTaxaTree.SecondSpecies.synRate

A, B, C, 0.865952, 18.0469, 1.97615, 0, 

Why do I have an estimate of 0 when there are synonymous substitutions between the sequences compared?

BTW are the units of these results also substitutions per branch?

I hope you can shed some light on this.

Thanks in advance!

m&m

Thanks Sergei for the prompt answer!

Concerning the units, would they be expected substitutions per nucleotide site even if I used a codon-based model for the estimation?

Cheers,

m&m