sampledoc
News and Announcements »

cluster_quality.py – compute the quality of a cluster

Description:

The input is a distance matrix (i.e. resulting file from beta_diversity.py).

Usage: cluster_quality.py [options]

Input Arguments:

Note

[REQUIRED]

-i, --input_path
Input distance matrix file
-m, --map
Mapping file
-c, --category
Column of mapping file delimiting clusters

[OPTIONAL]

-o, --output_path
Output path, prints to stdout if omitted
-s, --short
Print only the ratio of mean dissimilarities between/within clusters instead of more detailed output
--metric
Choice of quality metric to apply. Currently only one option exists, the ratio of mean(distances between samples from different clusters) to mean(distances between samples from the same cluster) Default: ratio

Output:

The output is either a single number (with -s), or a more detailed output of the similarity between and within clusters.

cluster quality based on the treatment category:

to compute the quality of clusters, and print to stdout, use the following idiom:

cluster_quality.py -i weighted_unifrac_otu_table.txt -m Fasting_Map.txt -c Treatment

cluster quality based on the DOB category:

to compute the quality of clusters, and print to stdout, use the following idiom:

cluster_quality.py -i weighted_unifrac_otu_table.txt -m Fasting_Map.txt -c DOB

Site index


sampledoc