phylobayes mpi error

Question

[xinwangcau]

Has anyone used Phylobayes mpi version (pb_mpi1.6f)? I perform phylobayes to analysis my data, the command is following:

mpirun -np 30 pb_mpi -d seven_mt_protein_combine.phylip seven_mt_protein_pb

but It produced these errors:

read data from file : seven_mt_protein_combine.phylip
error: does not recognise character I
error: does not recognise character I
.....

anyone know how to handle with these problem? I tried PhyloBayes (4.1), there is no errors with same data,but it run really slowly. Any suggestion?

[Andrew Watson]

We are having the same error. Has this been resolved?

We have been running pb_mpi on two linux servers successfully for some time now.

The error occurred with a new alignment, but we went back to check some older ones we have successfully used in the past with pb_mpi and came up against the same error message (characters "P" in the first case and "i" in the second). Both files are read by pb without any error message.

At least one of these servers was updated recently. I will try to compile 1.6f on mac now and see if it works with the same input files...

Answer 1

[Andrew Watson]

I was under the impression that Phylobayes only accepts .phy files so did not think of this until I asked around.

If anyone is still having this problem then it can be easily worked around by inputting the sequence in .nexus format, so that the data type is explicitly stated.