To test this command, first read the best tree that you could find from all the searches (e.g. after fusing and ratcheting). We will first generate a good tree for the bremer supports:
read ("course.fasta") search (max_time:0:0:3) select () (*Store the tree in a file for the next tutorial *) report ("tree_for_bremer.tre", trees) (*We have the tree from which bremer is to be calculated*) select (best:1) calculate_support (bremer)
Notice that this command is slow, as it performs a very intense search. Ten trees are built and swapped on each branch. If your original search was not powerful enough, this command may find shorter trees than your initial search.
When this command is finished, we are ready to report the support values for this tree.
Another way to calculate Bremer supports is to not constrain the search at all, and from the trees visited, collect the information for the clades not present. To do this we can use the visited argument within swap.
build (100) swap (cisited:"for_bremer.txt") select (best:1) read ("tree_for_bremer.tre") report (supports:bremer:"for_bremer.txt")
This strategy has several advantages. First and foremost, it tends to yield tighter support values (lower). Secondly, it is more efficient in the sense that the default technique has to repeat the search for every clade to collect the necessary information, while this strategy collects all the clades from the same strategy. Finally, you can calculate Bremer support with your search strategy of choice. You may cancel this Bremer search at any time and the necessary information will remain stored in the file and the best tree found in your overall search, which means that you will never recover negative Bremer support values.