Hi,

In our lab are very interesting in to perform ancestral sequence reconstruction (ASR) of amino acid sequences by using an "asymmetric" matrix (20x20) for each site.
In the attachment you can find an example for a protein of 255 amino acids. 
We are wondering if we can work on a batchfile of Hyphy and try to implement this, do you think is this possible?. If yes, which batch file do you recommend to start working with? 

Many thanks!

All the best,
Miguel 

Hi,

I've been trying to make progress on this, see attachment. Now I have specified the asymmetric matrix into a batch file that is called from the main batch file (just by following the procedure used for empirical amino acid models).
Apparently, it works well with an asymmetric matrix. 

My questions are:
- What is the order of aa states in the 20x20 matrix and the freqs vector?, I mean, for example position 0 which aa state is?, and so on for the 20 (just need to know the aa letters behind the numbers..). I did not find this info in the manual of Hyphy.
- There is a paramter "t" in the matrix that I guess should be used to normalize the matrix but I do not know from where is it coming and which value should take.
- Finally, I could apply this script to perform ASR per site, generate an ASR file per site, and then concatenate all those files. But I am wondering if there is another way to do this easier and more elegant..
Many thanks! 

Best,
Miguel