In this latter set, duplications with narrow expression patterns (i.e. tau-statistic > 0.85), often arose during the Dipteran split and are most prevalent in the male testis.
First, the root of the tree is given a creation time of 0. During DFS, just before calling DFS recursively on the left and right children of a node u, we set the duplication time t u d = max { max t { u, v } + 1, t u c + 1 }, where the second max is taken over all non-tree edges adjacent to u.
In addition, zebrafish differed noticeably from medaka, stickleback, and Tetraodon in average duplication set size, with 4.3 genes per duplication set compared to 5.4 genes per set in the three other species.
While the relationship between duplication set size and percentage of duplicated genes was similar among the four species., zebrafish again was the outlier, showing a pattern of more numerous small-scale duplications (set sizes 2 10).
The gene loss rate was kept constant, coinciding with the duplication rate of the data set with 10% duplications (0.003 per gene per PAM unit).
In terms of the duplicated gene expression, we observed that two genes in a duplication set show a strongly contrasting expression pattern.
In most sets, both members (genes) in a segmental duplication set were from same family.
We set a constant duplication rate λ = 0.2 for all simulations (Table 1).
Of the nine gene sets, only CSLA2 and CSLA4 in a segmental duplication set (CSLA2/ CSLA4) exhibited a relatively similar expression pattern.
Thus, any data set with a duplication on a minimum length branch will obviously have a lower maximum likelihood than one without.
