PanCancer Protein Networks

*Correspondence: Won-Min Song (won-min.song@mssm.edu or wonmin1984@gmail.com)

Data download

The github page houses companion R codes to reproduce results and figures for pan-cancer protein network study. The input data can be dowloaded at https://www.synapse.org/#!Synapse:syn30595684, or using the following R codes via synapser package:

library(synapser)

library(synapserutils)

synLogin("synapse_username","password")

files <- synapserutils::syncFromSynapse("syn30595684")

The .tar.gz file must be extracted under "Data" folder.

Data folder organization:

• CrossCancer_Module_Overlap: Fisher's Exact Test (FET) results to to evaluate significance of overlap between modules from different cancer types
• CrossCancer_Module_Presevation: Module preservation statistics when modules from different cancer types are compared.
• DEG: PanCancer TCGA transcriptome-based differential expression analysis between tumor and matched normal (limma results)
• DEP: Differentially expressed protein results between tumor and matched normal samples (limma results)
• MEGENA: MEGENA outputs for modules and networks. Modules and network edge list are concatenated across different cancer types, and are marked for their respective cancer types. Subfolders include ModuleRanking (ranked list of modules by enrichment of DEP signatures per cancer), Individual_MEGENA_Proteome (MEGENA output per cancer type from proteome data), hubs (hub statistics per cancer type).
• KD_stratification: holds enrichment of CRISPRi signature from Archilles data set (FDR < 0.05), or LINCS gene perturbation signatures in network hubs from proteome and transcriptome data (KD_Stratification_Evaluaton.Transcriptome/Proteome.txt). Also holds broader list of network key driver list (summarized_KDA_results.txt).
• LINCS_validation: holds .RDS files for GSEA analysis results (from fgsea R package) to evaluate enrichments of LINCS gene perturbation signatures in respective target gene neighborhoods. Proteome/Transcriptome subfolders hold the results by using proteome-/transcriptome-network neighborhoods.
• MSigDB: MSigDB v6.2 signatures
• MutationalDrivers: List of mutational driver genes from TCGA PanCancer studies.
• QCed_Proteome: Batch, age and gender adjusted protein expression data that were used for network constructions.
• PRO_Clinical_Signatures.RDS: Pre-compiled protein signatures in .RDS format for quick loading in R workspace

Main codes:

• analyze_DEPs.R: codes to reproduce DEP results in Figure 2.
• MoM_analysis.v1.R: codes to analyze and reproduce Modules-of-Modules results in Figure 3.
• summarize_transcriptome_preservation_wt_sunburst.R: codes to analyze and reproduce transcriptome preservation results in Figure 4.
• analyze_network_connectivity.R: codes to analyze protein network hubs and nominated key network regulators via KDA in Figure 5.