KnowEnG's Gene Prioritization Pipeline

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

This pipeline ranks the rows of a given spreadsheet, where spreadsheet's rows correspond to gene-labels and columns correspond to sample-labels. The ranking is based on correlating gene expression data (network smoothed) against pheno-type data.

There are four prioritization methods, using either pearson or t-test as the measure of correlation:

Options	Method	Parameters
Simple Correlation	simple correlation	correlation
Bootstrap Correlation	bootstrap sampling correlation	bootstrap_correlation
Correlation with network regularization	network-based correlation	net_correlation
Bootstrap Correlation with network regularization	bootstrapping w network correlation	bootstrap_net_correlation

Note: all of the correlation methods mentioned above use the Pearson or t-test correlation measure method.

---

How to run this pipeline with Our data

---

1. Clone the Gene_Prioritization_Pipeline Repo

 git clone https://github.com/KnowEnG/Gene_Prioritization_Pipeline.git

2. Install the following (Ubuntu or Linux)

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

3. Change directory to Gene_Prioritization_Pipeline

cd Gene_Prioritization_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_pearson	pearson correlation
make run_bootstrap_pearson	bootstrap sampling with pearson correlation
make run_net_pearson	pearson correlation with network regularization
make run_bootstrap_net_pearson	bootstrap pearson correlation with network regularization
make run_t_test	t-test correlation
make run_bootstrap_t_test	bootstrap sampling with t-test correlation
make run_net_t_test	t-test correlation with network regularization
make run_bootstrap_net_t_test	bootstrap t-test correlation with network regularization

---

How to run this pipeline with Your data

---

Follow steps 1-3 above then do the following:

* Create your run directory

mkdir run_directory

* Change directory to the run_directory

cd run_directory

* Create your results directory

mkdir results_directory

* Create run_paramters file (YAML Format)

Look for examples of run_parameters in ./Gene_Prioritization_Pipeline/data/run_files/zTEMPLATE_GP_BENCHMARKS.yml

* Modify run_paramters file (YAML Format)

set the spreadsheet, network and phenotype data file names to point to your data

* Run the Gene Prioritization Pipeline:

• Update PYTHONPATH enviroment variable

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

• Run (in test directory with env_setup as described above)

python3 ../src/gene_prioritization.py -run_directory ./run_dir -run_file zTEMPLATE_GP_BENCHMARKS.yml

---

Description of "run_parameters" file

---

Key	Value	Comments
method	correlation or net_correlation or bootstrap_correlation or bootstrap_net_correlation	Choose gene prioritization method
correlation_measure	pearson or t_test	Choose correlation measure method
gg_network_name_full_path	directory+gg_network_name	Path and file name of the 4 col network file
spreadsheet_name_full_path	directory+spreadsheet_name	Path and file name of user supplied gene sets
phenotype_name_full_path	directory+phenotype_response	Path and file name of user supplied phenotype response file
results_directory	directory	Directory to save the output files
number_of_bootstraps	5	Number of random samplings
cols_sampling_fraction	0.9	Select 90% of spreadsheet columns
rwr_max_iterations	100	Maximum number of iterations without convergence in random walk with restart
rwr_convergence_tolerence	1.0e-2	Frobenius norm tolerence of spreadsheet vector in random walk
rwr_restart_probability	0.5	alpha in V_(n+1) = alpha * N * Vn + (1-alpha) * Vo
top_beta_of_sort	100	Number of top genes selected
top_gamma_of_sort	50	Number of top genes reported
max_cpu	4	Maximum number of processors to use in the parallel correlation section

gg_network_name = STRING_experimental_gene_gene.edge
spreadsheet_name = CCLE_Expression_ensembl.df
phenotype_name = CCLE_drug_ec50_cleaned_NAremoved_pearson.txt

---

Description of Output files saved in results directory

---

• Any method saves separate files per phenotype with name {phenotype}_{method}_{correlation_measure}_{timestamp}_viz.tsv. Genes are sorted in descending order based on visualization_score.
  

Response	Gene_ENSEMBL_ID	quantitative_sorting_score	visualization_score	baseline_score
phenotype 1	gene 1	float	float	float
...	...	...	...	...
phenotype 1	gene n	float	float	float

• Any method saves sorted genes for each phenotype with name ranked_genes_per_phenotype_{method}_{correlation_measure}_{timestamp}_download.tsv.

Ranking	phenotype 1	phenotype 2	...	phenotype n
1	gene (most significant)	gene (most significant)	...	gene (most significant)
...	...	...	...	...
n	gene (least significant)	gene (least significant)	...	gene (least significant)

• Any method saves spreadsheet with top ranked genes per phenotype with name top_genes_per_phenotype_{method}_{correlation_measure}_{timestamp}_download.tsv.

Genes	phenotype 1	...	phenotype n
gene 1	1/0	...	1/0
...	...	...	...
gene n	1/0	...	1/0

References:

• The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity https://www.nature.com/articles/nature11003