For each of the three WGD events, two files are included:

Tree Files: A tree in Newick format, where the branch lengths are the values of alpha-t given in the manuscript and the other parameter estimates are included as well.

Conditional Probabilities File: A conditional probabilities file for each WGD.  This file, gives, for each pillar and each branch, the conditional probabilities of all state transitions. 

Fields in the conditional probabilities file:
First row: Field #1 is the Pillar ID.  Then, for each branch in the tree, there are 7 fields. The first is the branch name. The second is the branch length and the remaining 5 are padding to match the next line with the branch number. For internal branches, the name is "NONE," but the branch can be identified by its length in the corresponding tree.

Second row: Field #1 is the Pillar ID. Then, for each branch are all valid state transitions (7 total), including a state transitioning to itself. See Figure 2 of the main text for state definitions, where U == Dupl, F == DuplFix, S1 == Copy1, S2 == Copy2.

All following lines give the corresponding probabilities for each pillar. Thus, Dupl-->Copy2 is the probability that a ohnolog pair was duplicated at the start of that branch and transitioned to the Copy2 state by the end of that branch.

The Grass and At-alpha datasets were inferred under the WGD-bf model and the yeast dataset under the WGD-btf model (see Figure 2 for details).