This tree filtering procedure uses a reference tree (the estimated ML tree in this case) as a topological constraint against which other trees (e.g., bootstrap trees or gene trees) can be matched.
Through systematic resampling of the data by bootstrapping, a wider range of trees can be constructed.
Although phylogenetic trees can be constructed, subsequent confidence metrics such as bootstrap values are low, suggesting that these trees may be unreliable predictions of gene relationships [ 37, 38].
The estimation error in gene trees can be incorporated into the analysis using bootstrapping techniques [ 47, 48].
The reliability of the tree can be evaluated with the bootstrap method with replications (default value 100).
In this way, an average bootstrap score of internal branches of each pseudo-replicated tree can be used to assign a weight to this tree.
The tree containing all bootstrap values can be found in Additional file 4. The phylotypes are colored as in the core-gene tree.
The tree containing all bootstrap values can be found in Additional file 3. The four main phylotypes A, B1, B2 and D are marked by the colors blue, red, purple and green, respectively.
As in the case of the LS-based bootstrapping, the normalized secondary bootstrap scores can be computed and used to estimate the tree robustness.
Alternatively, secondary bootstrapping can be performed to obtain a robustness estimate for each of the trees built from pseudo-replicates.
