StefanPatman
commented
1 year ago Thank you for your feedback. It is not clear to me if there are use cases where the full genes are required in the output. I decided to add a toggle for this behavior at the export step for Nexus/IQTree, which is turned off by default. For PartitionFinder output, the program already behaved as expected; the full markers were not present when reading frames were defined.
You may try the changes if you install from source and use the updated Concatenator backend:
git clone https://github.com/iTaxoTools/ConcatenatorGui.git
cd ConcatenatorGui
pip install git+https://github.com/iTaxoTools/concatenator.git . -f packages.html -v(requires Python, plus MSVC on Windows)
This change will also be included in the executables once we revisit Concatenator in the future.
@cebos found that the output from Concatenator includes both the full genes and the genes subsetted by codon position. Not sure if this is the desired behaviour? PartitionFinder only wants the codon positions, not the full genes, as far as I can tell. Supplying the file to ModelFinder in IQtree resulted in 54 partitons instead of the expected 41, due to this duplication. IQTree doesn't throw an error, but somehow it is then working with two rates for the same loci… which is also weird.
alignment.nex.txt