Metaphlan output is different scale (i.e. fraction rather than reads)
cluster_features sets a bottom threshold before the log transform, and it is too high for metaphlan
need to add the bottom threshold as parameter
and create ShotgunExperiment class, which will set it automatically
Metaphlan output is different scale (i.e. fraction rather than reads) cluster_features sets a bottom threshold before the log transform, and it is too high for metaphlan need to add the bottom threshold as parameter and create ShotgunExperiment class, which will set it automatically