Add feature dispersion script

[rbeiko]

The purpose of Feature_Dispersion.py is to compute the sparsity of features on the phylogenetic tree. So, for example, a feature that is restricted to one part of a tree but is present in most or all of the genomes in that part will have a small dispersion value, whereas a feature that is present only in two very distantly related genomes will have a high dispersion. This captures the idea that transferred genes should have a patchy distribution on a phylogenetic tree.

The Python script requires:

• The reference tree (core_gene_alignment.tre) that is produced in the phylogenomics workflow
• The feature profile that is output by the annotation workflow

The script calculates the dispersion statistic and other bits of information, and prints these to a .tsv. Additionally, it produces a heatmap that considers the relationship between the dispersion statistic and count of genomes.

Optionally:

• The original samplesheet

If provided, the user can specify the samplesheet and indicate that they wish to calculate dispersion statistics based on columns in the samplesheet that they specify. These statistics are included in the .tsv but no corresponding heatmap is produced.

FeatureDispersion.tar.gz

The script should be run immediately after the completion of the phylogenomics subworkflow, if there is an accessible feature profile in the annotation directory (and, if specified, the samplesheet).