Gappa "Invalid Color Normalization with min >= max" error

[Kenta Renard]

Dear All,

I am trying to do some dispersion analysis on some placement data using Gappa but I get this error message and I am not sure how to fix it.

My command line input was as below:

./gappa analyze dispersion --jplace-path placement_file.jplace --mass-norm relative --method all --color-list spectral --write-nexus-tree --edge-values both

Best,

Kenta

[Lucas Czech]

Hi Kenta,

thanks for using gappa!

So, the most likely case where this error can happen is when all values that are being visualized there (the edge dispersion values) are identical (probably, 0, for one or another reason). In order to figure out what exactly is going on there, could you please share one of the jplace files with me that is causing the error?

Thanks and so long
Lucas

Dr. Lucas Czech

Postdoctoral Fellow

 

University of Copenhagen

Faculty of Health and Medical Sciences

Globe Institute, Section for GeoGenetics

 

lucas...@sund.ku.dk

 

 

[Kenta Renard]

Dear Lucas,

Thank you, it works now.

I also had another minor issue with the heat-tree module. In both the NEXUS tree output and the SVG it applies a branch support of 1.0 to all bipartitions. Is there a way to remove/disable this?

Best wishes,

Kenta

[Lucas Czech]

Hi Kenta,

good to know!

Hm, gappa does not add any branch support to the output. It might however keep existing values in there, depending on the format. So my guess is that you already have branch support values in the input somehow? If that's the case, you can simply remove them beforehand. If not, could you please share an example, so that I can investigate?

Cheers
Lucas

[Lucas Czech]

Hi Kenta,

thanks for sharing the files. Indeed, as suspected, the jplace file contains the values already, and gappa is simply printing them onto the tree. The values are in the file as inner node labels, not as branch support values. Which program did you use to create that tree? I'd recommend to have a look at our publication on issues with inner node labels and branch support values in newick tree files: https://doi.org/10.1093/molbev/msx055

As for removing them: Gappa has a helper command (https://github.com/lczech/gappa/wiki/Subcommand:-clean-tree) to remove these from the original newick file. You'd have to apply this to the tree before running the placement. If your placement is fast, you can just re-do it. If you have a lot of jplace files already, and do not want to re-run, let me know. I can probably add an option to just not print inner node labels, to make this a bit easier.

Hope that helps, so long
Lucas

On 7/3/24 13:22, Kenta Renard wrote:

Dear Lucas,

I have attached the .jplace file here (and example of the heat tree output as an SVG). Worth noting is that I used this .jplace file with other modules and this error did not happen.

Best,

Kenta