emad2/general_clustering_pipeline

Clusters columns of any data frame

KnowEnG's General Clustering Pipeline

This is the Knowledge Engine for Genomics (KnowEnG), an NIH BD2K Center of Excellence, General Clustering Pipeline.

This pipeline clusters a spreadsheet's columns, with various methods:

Options	Method	Parameters
K-means	K Means	kmeans
hierarchical clustering	hierarchical clustering	hclust
Linked hierarchical clustering	hierarchical clustering constraint	link_hclust
Bootstrapped hierarchical clustering	consensus hierarchical clustering	cc_ hclust
Bootstrapped K-means	consensus K Means	cc_kmeans
Bootstrapped Linked hierarchical clustering	consensus linked hierarchical clustering	cc_link_hclust

---

How to run this pipeline with Our data

---

1. Clone the General_Clustering_Pipeline Repo

 git clone https://github.com/KnowEnG-Research/General_Clustering_Pipeline.git

2. Install the following, for Linux

 apt-get install -y python3-pip libfreetype6-dev libxft-dev libblas-dev liblapack-dev libatlas-base-dev gfortran
 pip3 install pyyaml knpackage scipy==0.19.1 numpy==1.11.1 pandas==0.18.1 matplotlib==1.4.2 scikit-learn==0.17.1 

3. Change directory to General_Clustering_Pipeline

cd General_Clustering_Pipeline

4. Change directory to test

cd test

5. Create a local directory "run_dir" and place all the run files in it

make env_setup

6. Use one of the following "make" commands to select and run a clustering option:

Command	Option
make run_kmeans_binary	Clustering with k-means
make run_kmeans_continuous
make run_hclust_binary	Hierarchical Clustering
make run_hclust_continuous
make run_link_hclust_binary	Hierarchical linkage Clustering
make run_link_hclust_continuous
make run_cc_kmeans_binary	Consensus Clustering with k-means
make run_cc_kmeans_continuous
make run_cc_hclust_binary	Consensus Hierarchical Clustering
make run_cc_hclust_continuous
make run_cc_link_hclust_binary	Consensus Hierarchical linkage Clustering

---

How to run this pipeline with Your data

---

Follow steps 1-5 above then do the following:

* Create your run directory

mkdir run_dir

* Change directory to the run directory

cd run_dir

* Create your results directory

mkdir results

* Create run_paramters file (YAML Format)

Look for examples of run_parameters in the General_Clustering_Pipeline/data/run_files zTEMPLATE_cc_hclust.yml

* Modify run_paramters file (YAML Format)

Change processing_method to one of: serial, parallel depending on your machine.

processing_method: serial

set the data file targets to the files you want to run, and the parameters as appropriate for your data.

* Run the General Clustering Pipeline:

• Update PYTHONPATH enviroment variable

export PYTHONPATH='../src':$PYTHONPATH    

• Run

python3 ../src/general_clustering.py -run_directory ./run_dir -run_file zTEMPLATE_cc_net_nmf.yml

---

Description of "run_parameters" file

---

Key	Value	Comments
method	kmeans,hclust,link_hclust,cc_kmeans, cc_hclust, cc_link_hclust	Choose clustering method
affinity_metric	euclidean, manhattan, jaccard	Choose clustering affinity
linkage_criterion	ward, complete, average	Choose clustering affinity
spreadsheet_name_full_path	directory+spreadsheet_name	Path and file name of user supplied gene sets
results_directory	directory	Directory to save the output files
tmp_directory	./run_dir/tmp	Directory to save the temporary files
number_of_clusters	3	Estimated number of clusters
number_of_bootstraps	4	Number of bootstraps for cc_kmeans, cc_hclust and cc_link_hclust
rows_sampling_fraction	0.8	Select 80% of spreadsheet rows
cols_sampling_fraction	0.8	Select 80% of spreadsheet columns
top_number_of_rows	10	Top number of features to analyze
processing_method	serial or parallel or distribute	Choose processing method
parallelism	number of cores	Set number of cores for speed or memory
threshold	10	Threshold to define categorical data and continuous data in evaluation toolbox
nearest_neighbors	10	Number of Nearest Neighbors in cc_link_hclust method

spreadsheet_name = EXPR_GSE_METABRIC_lymphN_binary.tsv.gz

---

Description of Output files saved in results directory

---

• Output files of all methods save row by col heatmap variances per row with name row_variance_{method}_{timestamp}_viz.tsv.
  

	variance
row 1	float
...	...
row m	float

• Output files of all the methods save row by col heatmap with name row_by_col_heatmp_{method}_{timestamp}_viz.tsv.
  

	col 1	...	col n
row 1	float	...	float
...	...	...	...
row m	float	...	float

• Output files of all methods save col to cluster map with name col_labeled_by_cluster_{method}_{timestamp}_viz.tsv.
  

	cluster
col 1	int
...	...
col n	int

• Output files of all methods save row scores by cluster with name row_averages_by_cluster_{method}_{timestamp}_viz.tsv.
  

	cluster 1	...	cluster k
row 1	float	...	float
...	...	...	...
row m	float	...	float

• Output files of all methods save spreadsheet with top ranked rows per column with name top_row_by_cluster_{method}_{timestamp}_download.tsv.
  

	cluster 1	...	cluster k
row 1	1/0	...	1/0
...	...	...	...
row m	1/0	...	1/0

• All methods save silhouette number of clusters and corresponding silhouette score with name silhouette_average_{method}_{timestamp}_viz.tsv.
  

File Example: 
silhouette number of clusters = 3, corresponding silhouette score = 1