I have also encountered an issue that occurs when PartitionFinder returns all data subsets in a single partition. The MrBayes block in the best_scheme.txt file includes the command 'prset applyto=(all) ratepr=variable'. This is inappropriate for an unpartitioned alignment, and I have observed it causing pathologies for the MrBayes analyses. Removing the ’prset applyto=(all) ratepr=variable' when the alignment is unpartitioned would easily resolve this issue.
From Andrew Magee via email:
I have also encountered an issue that occurs when PartitionFinder returns all data subsets in a single partition. The MrBayes block in the best_scheme.txt file includes the command 'prset applyto=(all) ratepr=variable'. This is inappropriate for an unpartitioned alignment, and I have observed it causing pathologies for the MrBayes analyses. Removing the ’prset applyto=(all) ratepr=variable' when the alignment is unpartitioned would easily resolve this issue.