Generating heatmap with custom clustering

[ArnoHagenbeek]

Hi,

I have been trying to create a heatmap using phyloFlash_heatmap.R, but with the samples clustered by phylogeny. I have generated a distance matrix based on 18S, and am attempting to use it as the base of the sample clustering using the following command:
rscript phyloFlash_heatmap.R --custom-distance-matrix-sample=custom_distance_matrix_for_heatmap.tsv --library-name-from-file --cluster-samples=custom --min-ntu-count=300 *.csv
For many groups, the clustering becomes identical. However, for several samples, the phyloflash clustering does not match the trees given to me by programs such as iqtree or T-rex when they are provided with the same distance matrix.

Am I not providing the distance matrix correctly, or does phyloflash use a different logic to build the tree compared to these other programs?

Thanks!
Arno

Iqtree:

T-rex server:

Phyloflash_heatmap.R: