Labeling the trees in GARD NEXUS files?

[shjenkins94]

I was wondering if there was a good way to label the NEXUS files GARD outputs for downstream analyses like RELAX. Messed around with the Hyphy Batch Language for a bit but couldn't quite figure out how to import a GARD NEXUS file and export it without losing anything.

[spond]

Dear @shjenkins94,

If you use HBL's built-in ReadDataFile to load a partitioned NEXUS file, HyPhy parses the alignment partitions and the trees into complex memory structures. Re-exporting this structure losslessly (retaining the exact partition coordinates, model configurations, and metadata) via alignment.ExportPartitionedNEXUS is complex and often strips out custom NEXUS annotations.

Because we only need to annotate the tree topologies in the file, the most robust and 100% lossless approach is to treat the NEXUS file as text and perform string replacement inside the TREES block.

Since the NEXUS file is just text, a simple Python script using regular expressions is extremely reliable, easy to customize, and easy to integrate into any command-line pipeline (attached)

Running the script:

python3 label_gard.py -i input.nex -o output.nex -r "species_A|species_B" -l test

This will find all occurrences of species_A and species_B in every tree in the NEXUS file and label them as {test} (e.g. converting (species_A:0.1,species_B:0.1) to (species_A{test}:0.1,species_B{test}:0.1)).

This raw text transformation inside the tree definitions block preserves your file formats, character blocks, and sequence data perfectly intact for downstream tools like RELAX.

Best,
Sergei

label_gard.py

[shjenkins94]

I think the one issue with that is that looking at the label-tree.bf script it has options to label internal nodes as well.

[shjenkins94]

I wrote a label-gard-nexus.bf file that I think might work. Not sure if I should try adding it here or to hyphy-analyses