The columnar medulla intrinsic (Mi) cells send branches to a minimum of two different layers of the medulla.
Indeed, 166 of 1,000 simulations in which at least one duplication occurred on a minimum length branch had a maximum likelihood lower than our data set.
To avoid numerical instability and focus on branches of topological relevance, we computed DFs only for branches with a minimum length (>0.0004) in the source tree, which included 91 96% of branches in the complete gene tree for each gene family.
Thus, any data set with a duplication on a minimum length branch will obviously have a lower maximum likelihood than one without.
However, one of the reconstructed genome duplications from our actual data appears to have occurred on a minimum length branch.
In order to test the effects that a duplication on a minimum length branch would have on the data set, we also simulated 1,000 data sets using maximum likelihood values for the root count, λ, μ, and δ on Tree-10 and placed one duplication at random on a minimum length branch.
Because the vast majority of our simulations did not have a duplication on a minimum length branch, we would expect them to have higher likelihoods than our actual data.
Given this estimated value of δ, we would expect WGDs to occur on a minimum length branch in 0.943% of our simulations, and indeed, they actually did occur in only nine of our simulations.
The Sigmurethra Orthurethra branch has a minimum branch length, and as a consequence, the likelihood of a duplication on this branch was less on trees with smaller minimum branch lengths.
And the Bavarian branch of the trade union confederation plans to start collecting signatures on May 1st for a referendum on a minimum wage.
