Skip to content

The repository of the article "A Network-Centric Framework for the Evaluation of Mutual Exclusivity Tests on Cancer Drivers", published in Frontiers in Genetics (IF 4:77)

Notifications You must be signed in to change notification settings

cansuyalcinn/netcentric

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

A Network-centric Framework for the Evaluation of Mutual Exclusivity Tests on Cancer Drivers

This is the original repository for the Network-centric Framework for the Evaluation of Mutual Exclusivity Tests project codes. The project involves the evaluation of mutual exclusivity methods: Discover, Discover_strat, Fisher's Exact Test, MEGSA, MEMO, and WExT. The results from these methods include pairwise mutual exclusivity p-values. Based on them, we apply our network-centric epistatic evaluation.

Getting Started

Prerequisites

Python: 3.8-3.9

Getting Started

Using Github clone

git clone  https://github.com/abu-compbio/netcentric
cd NetCentric
pip install -r requirements.txt

You can also run this project locally by following these steps:

  1. Download the repo
  2. Unzip NetCentric-main
  3. Open cmd/terminal and cd into the project
  4. Execute python -m pip install -r requirements.txt

Some of the datasets are given in the .zip file format. In order to unzip, you should run script_unzip_data.py located under the NetCentric.

python script_unzip_data.py

Input

1. Network edge and index files.

1. PPI Network:

The file is located at data/intact_nodupl_index_file.txt where the first column contains gene identifiers and the second column corresponds to gene symbols.

0	""CHEBI
1	100147744
2	1B
3	1EFV
...

The file is located at data/intact_nodupl_edge_file.txt where the first two columns contain the gene identifiers and the third column denotes the confidence level of the interaction.

7589	13441	0.99
9123	10446	0.98
4248	1740	0.98
3776	5279	0.98
...

2. HINT Network:

The file is located at data/hint_index_file.txt where the first column contains gene identifiers and the second column corresponds to gene symbols.

1 A1BG	
2 A1CF	
3 A2BP1	
...

The file is located at data/hint_edge_file.txt where the first two columns contain the gene identifiers.

2988 7255
7255 9111
7255 9109
...

3.PPI Network with 0.25 and 0.45 thresholds:

These index files are located at data/intact_index_file_0.25.txt and data/intact_index_file_0.45.txt where the first column contains gene identifiers and the second column corresponds to gene symbols.

1	""CHEBI
2	100147744
3	1B
4	1C
...

These edge file are located at data/intact_edge_file_0.25.txt and data/intact_edge_file_0.45.txt where the first two columns contain the gene identifiers.

MDM2	TP53
MYC	MAX
BRAF	MAP2K1
...

4.STRING Network:

The file is located at data/string_network.txt where it shows the edges between two genes (gene1 and gene2).

	gene1	gene2
0	A1CF	APOBEC3H
1	A1CF	APOBEC3G
2	A1CF	APOBEC3A
3	A1CF	APOBEC3C
...

2. Input mutation data

The mutation data includes pairwise mutual exclusivity p-values given for each method (discover, discover_strat, fishers, megsa, memo and wext). The files with the name of mutations_all_genes include all genes and intact_filtered include only ones in intact network.

The file is located at data/{method}mutation_filtered_ep_data/{cancerType}{method}result_mutations_all_genes{threshold}.txt

	gene1	gene2	pvalue
0	A2M	A2ML1	0.6584654889330113
1	A2M	ABCA1	0.5332913581418495
2	A2M	ABCA10	0.8971732886956303
...

The file is located at data/{method}mutation_filtered_ep_data/{cancerType}{method}_pairs_intact_filtered_subset{threshold}.txt

	gene1	gene2	pvalue	oddsratio
0	TCF7L2	CTNNB1	0.9015805073650888	1.6786858974358974
1	SMAD4	SMAD3	0.839665475908354	1.4567901234567902
2	EP300	TP53	0.0742406168447767	0.5221052631578947
...

The file is located at data/binary_matrices_all_genes_ep_mutation_filtered/ directory. Each row is a TCGA patient id and each column is a gene. The matrix contains 1 if the gene is mutated in the corresponding patient. Here, we only provide the mutation matrix for COADREAD.

	patients	A1BG	A1CF	A2M ...
TCGA-3L-AA1B-01A	0	0	0
TCGA-4N-A93T-01A	0	0	0
TCGA-4T-AA8H-01A	0	0	0
...

MLA: The file contains the corresponding MLA.

The file is located at data/MLA_ep_mutation_filtered_all_genes

A1BG	4.261253658699028
A1CF	5.095042391780406
A2M	5.539871662596874
...

3. Reference cancer genes

The file is located at data/known_cancer_genes directory.

  1. CGC genes: We download all the genes from Cancer Gene Census from COSMIC database.

  2. CGC_SNV genes: We try using a subset of CGC genes to include only those which have SNV type of mutations in cancer (378 out of 723 genes). To this end, we filter out the genes where the mutation type column consists of only A (amplification), D (large deletion) or T (translocation).

  3. IntoGen genes: We download Unfiltered driver results 05.tsv file (2020-02- 02 release) from https://www.intogen.org and include the genes where FILTER column is PASS, which results in 503 genes.

4. Tissue Specific Network (TSN)

Rather than using a common nonspecific network for all the cancer types, in this component of our evaluation framework we employ TSN based on the tissue in which the tumor develops. To construct the TSN for a particular tissue, we start with the original PPI network and remove the edges between the pairs of genes that are not co-expressed in the corresponding tissue. For this purpose, we download RNA-seq datasets from GTEX portal. In the main article it was discussed under the section "Network-centric ME Evaluations in Relation to TSN".

The file is located at data/gtex_tsn_fractions_intact_filtered_applied_threshold

MDM2	TP53	1.0
PAK1	RAC1	1.0
FADD	CASP8	0.9987163029525032
...

5. Results

The Mutual Exclusivity results will be available in the folder ME_results. The TSN results will be available in the folder tsn_results. The MLA results will be available in the folder MLA_results. These folders will appear under the main directory, when the results are ready.

The commands to run the mutual exclusivity algorithms are shown in running_mex_methods section.

Runs

The codes regarding various analyses given in the main article.

ME Evaluations Based on Defined Metrics

The main source code for the evaluation ME Tests. In the main article it was discussed under the section "ME Evaluations Based on Defined Metrics". This analysis code were also used in the section "Robustness Analysis of Evaluations Based on Defined Metrics". robustness_iterations value is given as parameter i: number of iteration in the code.

As output, you get tables with all analysis results in NetCentric/ME_results. To generate the algorithm for the given input, the following script should run (c: cancer type, t: threshold, i: number of iteration, m: methods, p: p_value threshold, -ni: network index file, -e:network edge file, -r: Reference cancer genes)

cd src
evaluations_on_metrics.py -c COADREAD -t 20 -i 100 -m discover discover_strat fishers megsa memo wext -p 0.05 -ni intact_nodupl_index_file.txt -e intact_nodupl_edge_file.txt -r Census_allFri_Apr_26_12_49_57_2019.tsv

In order to run the code with STRING network, keep the edge and index files empty and use the command given below. (-str: String network)

cd src
evaluations_on_metrics.py -c COADREAD -t 20 -i 100 -m discover discover_strat fishers megsa memo wext -p 0.05 -str string_network.txt -r Census_allFri_Apr_26_12_49_57_2019.tsv

ME Evaluations Based on Corrections via MLA

Scatterplots of percentage significance of mutual exclusivity runs vs mutation load association (MLA). In the main article it was discussed under the section "ME Evaluations Based on Corrections via MLA". As output, you get results in NetCentric/MLA_results/percent_sig_figures

cd src
evaluations_via_mla.py -c COADREAD -t 20 -m discover discover_strat fishers megsa memo wext

Scatterplots of percentage significance of mutual exclusivity runs vs mutation load association (MLA) when only CGC genes that have > 1 neighbors are included. As output, you get results in NetCentric/MLA_results/perc_sig_figures_for_multiple_neighbors

cd src
evaluations_via_mla_neighbors.py -c COADREAD -t 20 -m discover discover_strat fishers megsa memo wext

ME Evaluations Based on Corrections via TSN

In the main article it was discussed under the section "ME Evaluations Based on Corrections via TSN". (c: cancer type, t: threshold, m: methods, ti: tissue, th: tsn threshold). As output, you get tables with all analysis results in NetCentric/tsn_results.

cd src
evaluations_via_tsn.py -c COADREAD -t 20 -m discover discover_strat fishers megsa memo wext -ti Colon -th 0.0

ROC analysis based on tissue-specificity. As output, you get results in NetCentric/tsn_results/figure_tsn_AUROC (c: cancer type, t: threshold, m: methods, th: tsn threshold, p: percentage )

cd src
me_on_tsn_ntsn_roc_curve.py -c COADREAD -t 20 -m discover discover_strat fishers megsa memo wext -th 0.0 -p 0.25

About

The repository of the article "A Network-Centric Framework for the Evaluation of Mutual Exclusivity Tests on Cancer Drivers", published in Frontiers in Genetics (IF 4:77)

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 86.1%
  • Jupyter Notebook 12.8%
  • R 1.1%