Omics Integrator is a package designed to integrate proteomic data, gene expression data and/or epigenetic data using a protein-protein interaction network. It is comprised of two modules, Garnet and Forest.
Contact: Seyma Unsal Beyge [unsalseyma@gmail.com]
All data included in P100 dataset is provided in "./raw_data/" folder.
Some of the scripts require python module of cmapPy
1. prepare_raw_data_files.py
2. prepare_seed_protein_list.py
3. prepare_interactome.py
4. find_all_possible_drug_networks.py
- For parameter tuning purposes, forest-tuner repository can be used.
- For separation score calculation, method described in Menche et.al.,2015 is used.
Running prepare_raw_data_files.py
Find significantly perturbed genes and/or phosphoproteins and create the files:
- p values of each gene/phosphoprotein in each significantly perturbed condition
- significant gene read outs
- log2FoldChanges of those significantly perturbed gene/protein list
If only landmark genes are chosen, all files will be prepared for 978 landmark genes.
Options:
| Option | Option | Destination | Description |
|---|---|---|---|
| -h | --help | help | show this help message and exit |
| -c | --cell | cellline | (Required)Cell line of interest. Choices: "A375","A549","MCF7","PC3","YAPC","NPC". |
| -p | --pval | pvalThreshold | p-value threshold that is desired to get the significant conditions (Default is 0.05) |
| -L | --L1000 | l1000File | Path to the raw L1000 data file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_Level3_INF_mlr12k_n1667x12328.gctx" |
| -P | --P100 | P100File | Path to the raw P100 data file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_Level3_INF_mlr12k_n1667x12328.gctx" |
| -ci | --cellinfo | cellFile | Path to the cell info file. Default = "./raw_data/GSE101406_Broad_LINCS_cell_info.txt" |
| -pt | --pert | pertFile | Path to the perturbant info file. Default = "./raw_data/GSE101406_Broad_LINCS_pert_info.txt" |
| -lg | --Lgene | LgeneFile | Path to the L1000 gene info file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_gene_info.txt" |
| -la | --Lanalyte | LanalyteFile | Path to the L1000 analyte info file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_inst_info.txt" |
| -pi | --PinstFile | PinstFile | Path to the P100 instrument info file. Default = "./raw_data/GSE101406_Broad_LINCS_P100_inst_info.txt" |
| -pa | --Panalyte | PanalyteFile | Path to the P100 analyte info file. Default = "./raw_data/GSE101406_Broad_LINCS_P100_analyte_info.txt" |
| -lm | --lmGenes | lmGenesFile | Path to the L1000 Landmark genes file. Default = "./raw_data/landmark_genes_978.txt" |
| . | --iflandmark | iflandmark | True or False to select which genes will be used in transcriptomic data (L1000). If True, only landmark genes will be used in the analysis. Landmark genes are listed in the file whose path can be given with --lmGenesoption. Default = False |
| . | --outpath | outputpath | Path to the directory which will hold the output files. Default = "./raw_data/" |
Running prepare_seed_protein_list.py
Given the drugname, cell line name of interest and pvalues for L1000 and P100 data, this script finds out seed protein list including transcriptomic, phosphoproteomic and drug targetome data. Then it calculates corresponding prize values and outputs the prize file as "{cell_line}_{drugname}_prizefile.txt".
Options:
| Option | Option | Destination | Description |
|---|---|---|---|
| -h | --help | help | show this help message and exit |
| -c | --cell | cellline | (Required) Cell line of interest. Choices: "A375","A549","MCF7","PC3","YAPC","NPC". |
| -d | --drug | drugname | (Required) Drug of interest |
| -P | --pval_p100 | pval_p100 | (Required) Pvalue necessary to collect significantly phophorylated proteins from P100 data. |
| -L | --pval_l1000 | pval_l1000 | (Required) Pvalue necessary to collect TFs related with significantly transcribed genes from L1000 data. |
| -lf | --LFfile | LFtestfile | (Required) Path to the L1000 F test pvalues file. |
| -pf | --PFfile | PFtesetfile | (Required) Path to the P100 F test pvalues file. |
| . | --l1000_fc | l1000_fc | (Required) Path to the L1000 Fold Change file. |
| . | --p100_fc | p100_fc | (Required) Path to the P100 Fold Change file. |
| -r | --regnet | regnet_file | Path to the Regulatory Network file. Default = "./raw_data/human_regulatory_network/human.source" |
| -t | --tf_file | tf_file | Path to the DBTF file. Default = "./raw_data/DBTF_List.txt" |
| -dt | --targetome_file | targetome_file | Path to the drug targets file. Default = "./raw_data/drug_targets.txt" |
| -lg | --Lgene | LgeneFile | Path to the L1000 gene info file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_gene_info.txt" |
| -ps | --Panalyte | PanalyteFile | Path to the P100 analyte info file. Default = "./raw_data/GSE101406_Broad_LINCS_P100_analyte_info.txt" |
| . | --outpath | outpath | Path to the directory which will hold the output files. Default = "./raw_data/prize_files/" |
Running prepare_interactome.py
Given the drugname, cell line name of interest, path to the reference interactome paths to the raw L1000 data, gene-inst-pert info files, path to the file including list of landmark genes processes the reference interactome to exclude hub nodes, low-expressed genes, then perfroms link prediction and adds edges passing localization filters. Finally outputs the interactome file as "{cell_line}_{drugname}_iref_processed_interactome.txt".
Options:
| Option | Option | Destination | Description |
|---|---|---|---|
| -c | --cell | cellline | (Required) Cell line of interest. Choices: "A375","A549","MCF7","PC3","YAPC","NPC". |
| -d | --drug | drugname | (Required) Drug of interest. |
| -i | --int | interactome | Path to the reference interactome. Default='./raw_data/iref_mitab_miscore_2013_08_12_interactome_no_selfloop.txt |
| -lg | --Lgene | LgeneFile | Path to the L1000 gene info file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_gene_info.txt" |
| -la | --Lanalyte | LanalyteFile | Path to the L1000 analyte info file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_inst_info.txt" |
| -L | --L1000 | l1000File | Path to the raw L1000 data file. Default = "./raw_data/GSE101406_Broad_LINCS_L1000_Level3_INF_mlr12k_n1667x12328.gctx" |
| . | --localization | localization | Path to the localization info file. Default = "./raw_data/subcellular_location.txt" |
| . | --expthreshold | expthreshold | The level of expression value that will used for excluding the low expressed genes. Default = 2.0. |
| . | --outpath | outputpath | Path to the directory which will hold the output files. Default = "./raw_data/interactomes/{drugname}/" |
| . | --lpmethod | lpmethod | The link prediction method to use for interactome preparation. Options are "Adamic/Adar", "Jaccards", "Preferential Attachment", "Resource Allocation" and None. If None, any link prediction is not used and the preparation procedure is terminated. Default='Adamic/Adar' |
Running find_all_possible_drug_networks.py
This script designed as a case study which runs Forest module of Omics Integrator.
It includes several variables such as paths to the prize and edge files, parameters needed by Forest, outpaths etc. These variables within the script should be updated accordingly.