Hi Artiz, Thanks for using EGSEA and apologies for the inconvenience error. Can you please post your design and contrast matrices here? And which version of EGSEA are you using? Cheers, Monther On Thu, Apr 13, 2017 at 3:25 AM, irizarnaunk [bioc] < noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User irizarnaunk <https: support.bioconductor.org="" u="" 12829=""/> wrote Question: > Ensemble EGSEA (EGSEA): base GSA method failures > <https: support.bioconductor.org="" p="" 94885=""/>: > > Hi all!! > > My name is Aritz Irizar and I am a postdoctoral fellow at the Icahn School > of Medicine at Mount Sinai. > > I'm trying to use EGSEA to generate pathway/gene-set enrichment results > from a RNA seq dataset that contains two experimental groups (disease vs > control) sampled at 4 timepoints. I perform 4 disease vs control contrasts, > one for each timepoint and I feed the corresponding voom object and the > contrast matrix to egsea in a for loop where, in each iteration, only one > of the GSA methods is used so I can see which of them work and which give > an error. > > For some reason, 9 of the 12 base GSA methods give an error (only CAMERA, > ROAST and FRY seem to work). > > See the code and the errors below please: > > #generate gene-level voom object > > dge_gene <- DGEList(counts = gcounts, genes = gannots) > dge_gene <- dge_gene[rowSums(cpm(dge_gene) > 0.2) > 5, keep.lib.sizes = F] > dge_gene <- calcNormFactors(dge_gene) > > #build model for differential expression analysis > > diffexpr_model <- model.matrix(~ ctrl_disease + day + RIN + total_RNA_ug + ctrl_disease:day, data = phenotypes) > colnames(diffexpr_model) <- make.names(colnames(diffexpr_model)) > > #perform voom transformation > pdf("diffexpr_voom_plot.pdf") > gene_voom <- voom(dge_gene, design = diffexpr_model, plot = TRUE, span = 0.5) > dev.off() > > #subset voom object to contain only genes with an Entrez Gene ID > entrez_voom <- gene_voom[which(!is.na(gene_voom$genes$EntrezID)),] > entrez_voom <- entrez_voom[-c(which(duplicated(entrez_voom$genes$EntrezID))),] > rownames(entrez_voom) <- entrez_voom$genes$EntrezID > rownames(entrez_voom$genes) <- entrez_voom$genes$EntrezID > > #build gene-set collections to compute enrichment for > > murine_gs_annots <- buildIdx(entrezIDs = entrez_voom$genes$EntrezID, species = "mouse", msigdb.gsets = c("c2", "c5", "c7"), gsdb.gsets = "none", kegg.exclude = c("Metabolism"), min.size = 10) > > #generate a symbolsMap > > symbols_map <- entrez_voom$genes[, c(2,8)] > colnames(symbols_map) <- c("FeatureID", "Symbols") > > #establish the list of independent gene-set enrichment methods for EGSEA to use > > egsea_methods <- egsea.base() > > #build contrasts for EGSEA > gsea_contrasts <- makeContrasts( > day5 = ctrl_diseaseD, > day12 = ctrl_diseaseD + ctrl_diseaseD.day12, > day17 = ctrl_diseaseD + ctrl_diseaseD.day17, > day36 = ctrl_diseaseD + ctrl_diseaseD.day36, > levels = diffexpr_model > ) > > #run EGSEA > #first pass to see which base GSA methods work in the dataset > > egsea_dir_1 <- "./single_gsa_results" > unlink(egsea_dir_1, recursive = TRUE, force = TRUE) > dir.create(egsea_dir_1) > > gsa_results_list <- list() > > for(i in 1:length(egsea_methods)){ > tryCatch({ > print(i) > gsa_results <- egsea(voom.results = entrez_voom, contrasts = gsea_contrasts, gs.annots = murine_gs_annots, symbolsMap = symbols_map, baseGSEAs = egsea_methods[i], report = FALSE, print.base = FALSE, verbose = TRUE, keep.limma = FALSE, keep.set.score = TRUE, egsea.dir = egsea_dir_1, num.threads = 7) > gsa_results_list[[i]] <- gsa_results > }, error = function(e){cat("ERROR:", conditionMessage(e), "\n")}) > } > > 1] 1 > > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running CAMERA for day5" > [1] " Running CAMERA for day12" > [1] " Running CAMERA for day17" > [1] " Running CAMERA for day36" > Running CAMERA on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and c5 gene sets" > [1] " Running CAMERA for day5" > [1] " Running CAMERA for day12" > [1] " Running CAMERA for day17" > [1] " Running CAMERA for day36" > Running CAMERA on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and c7 gene sets" > [1] " Running CAMERA for day5" > [1] " Running CAMERA for day12" > [1] " Running CAMERA for day17" > [1] " Running CAMERA for day36" > Running CAMERA on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and kegg gene sets" > [1] " Running CAMERA for day5" > [1] " Running CAMERA for day12" > [1] " Running CAMERA for day17" > [1] " Running CAMERA for day36" > Running CAMERA on all contrasts ... COMPLETED > [1] "EGSEA analysis has completed" > [1] 2 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running ROAST for day5" > [1] " Running ROAST for day12" > [1] " Running ROAST for day17" > [1] " Running ROAST for day36" > Running ROAST on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and c5 gene sets" > [1] " Running ROAST for day5" > [1] " Running ROAST for day12" > [1] " Running ROAST for day17" > [1] " Running ROAST for day36" > Running ROAST on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and c7 gene sets" > [1] " Running ROAST for day5" > [1] " Running ROAST for day12" > [1] " Running ROAST for day17" > [1] " Running ROAST for day36" > Running ROAST on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and kegg gene sets" > [1] " Running ROAST for day5" > [1] " Running ROAST for day12" > [1] " Running ROAST for day17" > [1] " Running ROAST for day36" > Running ROAST on all contrasts ... COMPLETED > [1] "EGSEA analysis has completed" > [1] 3 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running SAFE for day5" > [1] "ERROR: SAFE encountered an \n\t\t\t\t\t\t\terror Error in runsafe(voom.results = voom.results, contrast = contrast, gs.annot = gs.annot, : Invalid contrasts selected.\n" > ERROR: ERROR: One of the GSE methods failed on this > dataset (safe). > Remove it and try again. > See error messages for > more information. > [1] 4 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running GAGE for day5" > [1] "ERROR: GAGE encountered an \n\t\t\t\t\t\t\terror Error in rungage(voom.results = voom.results, contrast = contrast, gs.annot = gs.annot, : Invalid contrasts selected.\n" > ERROR: ERROR: One of the GSE methods failed on this > dataset (gage). > Remove it and try again. > See error messages for > more information. > [1] 5 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running PADOG for day5" > [1] "ERROR: PADOG encountered an \n\t\t\t\t\t\t\terror Error in runpadog(voom.results = voom.results, contrast = contrast, gs.annot = gs.annot, : Invalid contrasts selected.\n" > ERROR: ERROR: One of the GSE methods failed on this > dataset (padog). > Remove it and try again. > See error messages for > more information. > [1] 6 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > Running PLAGE on all contrasts ... COMPLETED > ERROR: incorrect number of dimensions > [1] 7 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > Running ZSCORE on all contrasts ... COMPLETED > ERROR: incorrect number of dimensions > [1] 8 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > Running GSVA on all contrasts ... COMPLETED > ERROR: incorrect number of dimensions > [1] 9 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > Running SSGSEA on all contrasts ... COMPLETED > ERROR: incorrect number of dimensions > [1] 10 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running GLOBALTEST for day5" > [1] "ERROR: GLOBALTEST encountered an \n\t\t\t\t\t\t\terror Error in runglobaltest(voom.results = voom.results, contrast = contrast, : Invalid contrasts selected.\n" > ERROR: ERROR: One of the GSE methods failed on this > dataset (globaltest). > Remove it and try again. > See error messages for > more information. > [1] 11 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running ORA for day5" > [1] " Running ORA for day12" > [1] " Running ORA for day17" > [1] " Running ORA for day36" > Running ORA on all contrasts ... COMPLETED > ERROR: attempt to select less than one element in get1index > [1] 12 > [1] "The ensemble mode was disabled." > [1] "EGSEA analysis has started" > [1] "Log fold changes are estimated using limma package ... " > [1] "EGSEA is running on the provided data and c2 gene sets" > [1] " Running FRY for day5" > [1] " Running FRY for day12" > [1] " Running FRY for day17" > [1] " Running FRY for day36" > Running FRY on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and c5 gene sets" > [1] " Running FRY for day5" > [1] " Running FRY for day12" > [1] " Running FRY for day17" > [1] " Running FRY for day36" > Running FRY on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and c7 gene sets" > [1] " Running FRY for day5" > [1] " Running FRY for day12" > [1] " Running FRY for day17" > [1] " Running FRY for day36" > Running FRY on all contrasts ... COMPLETED > [1] "EGSEA is running on the provided data and kegg gene sets" > [1] " Running FRY for day5" > [1] " Running FRY for day12" > [1] " Running FRY for day17" > [1] " Running FRY for day36" > Running FRY on all contrasts ... COMPLETED > [1] "EGSEA analysis has completed" > Warning messages: > 1: In mclapply(args.all, rungsva.contrast, mc.cores = num.workers) : > all scheduled cores encountered errors in user code > 2: In mclapply(args.all, rungsva.contrast, mc.cores = num.workers) : > all scheduled cores encountered errors in user code > 3: In mclapply(args.all, rungsva.contrast, mc.cores = num.workers) : > all scheduled cores encountered errors in user code > 4: In mclapply(args.all, rungsva.contrast, mc.cores = num.workers) : > all scheduled cores encountered errors in user code > > From the errors, I cannot really figure out what went wrong in each case. > Besides, I could really use some advice on what criteria to use when > selecting the base GSA methods to include in the egsea run. > > Any suggestions? > > Thanks a lot, > > Aritz > > > > ------------------------------ > > Post tags: egsea > > You may reply via email or visit Ensemble EGSEA (EGSEA): base GSA method failures >