Dear Miguel,

Set

Code:

ACCEPT_ROOTED_TREES=1;
 

flag before you call Tree id = ... to tell HyPhy not to unroot the tree.
This is safer than trying to figure out where your node went after rerooting.

Cheers,
Sergei

Dear Miguel,

If you need to have access to node names in HyPhy batch language, you can try using the following example. HyPhy can convert a tree object into a list of nodes in post-order traversal (using the tree^0 operation). For example:

Code:

ACCEPT_ROOTED_TREES = 1;
Tree 				myTree = ((0,(1,3)),outgroup);
flatTree = 			myTree^0;

fprintf (stdout, flatTree);_hyphyAssociativeArray={};

 

produces the following output

Code:

(_hyphyAssociativeArray["1"])={};
(_hyphyAssociativeArray["1"])["Name"] = "0";
(_hyphyAssociativeArray["1"])["Length"] = -1;
(_hyphyAssociativeArray["1"])["Depth"] = 2;
(_hyphyAssociativeArray["1"])["Parent"] = 5;
(_hyphyAssociativeArray["2"])={};
(_hyphyAssociativeArray["2"])["Name"] = "1";
(_hyphyAssociativeArray["2"])["Length"] = -1;
(_hyphyAssociativeArray["2"])["Depth"] = 3;
(_hyphyAssociativeArray["2"])["Parent"] = 4;
(_hyphyAssociativeArray["3"])={};
(_hyphyAssociativeArray["3"])["Name"] = "3";
(_hyphyAssociativeArray["3"])["Length"] = -1;
(_hyphyAssociativeArray["3"])["Depth"] = 3;
(_hyphyAssociativeArray["3"])["Parent"] = 4;
(_hyphyAssociativeArray["4"])={};
(_hyphyAssociativeArray["4"])["Name"] = "Node3";
(_hyphyAssociativeArray["4"])["Length"] = -1;
(_hyphyAssociativeArray["4"])["Depth"] = 2;
(_hyphyAssociativeArray["4"])["Parent"] = 5;
((_hyphyAssociativeArray["4"])["Children"])={};
((_hyphyAssociativeArray["4"])["Children"])["0"] = 2;
((_hyphyAssociativeArray["4"])["Children"])["1"] = 3;
(_hyphyAssociativeArray["5"])={};
(_hyphyAssociativeArray["5"])["Name"] = "Node1";
(_hyphyAssociativeArray["5"])["Length"] = -1;
(_hyphyAssociativeArray["5"])["Depth"] = 1;
(_hyphyAssociativeArray["5"])["Parent"] = 7;
((_hyphyAssociativeArray["5"])["Children"])={};
((_hyphyAssociativeArray["5"])["Children"])["0"] = 1;
((_hyphyAssociativeArray["5"])["Children"])["1"] = 4;
(_hyphyAssociativeArray["6"])={};
(_hyphyAssociativeArray["6"])["Name"] = "outgroup";
(_hyphyAssociativeArray["6"])["Length"] = -1;
(_hyphyAssociativeArray["6"])["Depth"] = 1;
(_hyphyAssociativeArray["6"])["Parent"] = 7;
(_hyphyAssociativeArray["7"])={};
(_hyphyAssociativeArray["7"])["Name"] = "Node0";
(_hyphyAssociativeArray["7"])["Length"] = -1;
(_hyphyAssociativeArray["7"])["Depth"] = 0;
((_hyphyAssociativeArray["7"])["Children"])={};
((_hyphyAssociativeArray["7"])["Children"])["0"] = 5;
((_hyphyAssociativeArray["7"])["Children"])["1"] = 6;
(_hyphyAssociativeArray["0"])={};
(_hyphyAssociativeArray["0"])["Name"] = "myTree";
(_hyphyAssociativeArray["0"])["Root"] = 7;
 

It looks pretty ugly, but what the array is quite simple:
key "0" (i.e. flatTree[0]) stores the information about the tree itself and where the root is located in the array (in this case it's at index 7). Then all nodes are laid out in post-order traversal (left-right-parent).
For each node you get a series of attributes (its name, what node is its parent, what depth in the tree it is, whether or not it has children etc). If all you want to look up is the name of the left-most child of the root node, something like this should do the trick:

Code:

rootIndex =(flatTree [0])["Root"];
leftChildIndex = ((flatTree[rootIndex])["Children"])[0];
fprintf (stdout, "\nThe name of root's left child is ", (flatTree[leftChildIndex])["Name"], "\n");
 

Cheers.
Sergei
[code]

Dear Sergei,

Finally with all those improvements all seem to be very well.
So, I tried to analyze my simulated data and I found a bad surprise. I know the real MRCA and I see that when I write the command
ACCEPT_ROOTED_TREES=1;
before ASR, I found a big error (respect to the real MRCA) about 20% (that is for any MRCA node, Node1 (ingroup) or Node0 (outgroup)), however if I do not write that command the error is only of 5% (like other software of ASR).
Then, thinking about that, I found that the tree (with the active ACCEPT_ROOTED_TREES=1; command) of the LikelihoodFunction in the ASR has the length of branches very different, where the outgroup had a very very long branch.

For example, for the same alignment:
- Without the command:
Results from the likelihood optimization:
Log Likelihood = -6051.63159193931;
Tree givenTree=((((((seq00001:0.00140814,seq00008:0.00562907)Node6:0.0510105,seq00010
:0.0550151)Node5:0.00451872,((seq00002:0.0115897,s
eq00005:0.0157931)Node11:0.0124699,seq00004:0.0280265)Node10:0.0258714)Node4:0.0
468666,seq00007:0.136883)Node3:0.0368719,(seq00003:0.04
23708,seq00006:0.047428)Node16:0.0849871)Node2:0.222705,seq00009:0.350053,outgro
up:0.123559);
       [RECONSTRUCTING ANCESTORS]
       [WRITING ANCESTORS TO Results/sequences00001_ancestors.fas]


- With the command:
Results from the likelihood optimization:
Log Likelihood = -6642.04685957071;
Tree givenTree=(((((((seq00001:0.0013966,seq00008:0.00564641)Node6:0.0513962,seq00010
:0.0546931)Node5:0.00396752,((seq00002:0.0115815,s
eq00005:0.015796)Node11:0.0124217,seq00004:0.0280487)Node10:0.0263444)Node4:0.04
58819,seq00007:0.137755)Node3:0.0318315,(seq00003:0.041
6297,seq00006:0.0477905)Node16:0.0913618)Node2:0.0270367,seq00009:0.535481)Node1
:0.09,outgroup:7499.98);
       [RECONSTRUCTING ANCESTORS]
       [WRITING ANCESTORS TO Results/sequences00001_ancestors.fas]


I cannot understand why those trees are so different in the length of branches, perhaps the ASR does not work fine with rooted trees.  :-/
I attach you the complete Batch File on this.

Cheers,
Miguel

Dear Sergei,

That was great again!   
Thank you very much!!

Cheers,
Miguel