May I have some assistance please? I have performed topGO analysis on a GeoMx DSP dataset. I set the parameters for the significance value and estimate in the code for differentially expressed genes in the topGO analysis, but when I look at the list go 'significant' genes in the TopGO output, some of the genes included aren't actually differentially expressed and are outside of the parameters I set out.
Please note that genesInTerm provides exactly what one might expect a function with that name might provide - the genes that are annotated to a term. You instead want the genes in that node that are also significant. I don't see a function intended to do that directly (under the hood, topGO just uses intersect between the significant genes and the genes in the term). Here is a self-contained example.
> library(topGO)
> example("genesInTerm")
> term.genes <- genesInTerm(GOdata, "GO:0032446")
> term.genes
$`GO:0032446`
[1] "1651_at" "1720_at"
[3] "1721_g_at" "1925_at"
[5] "1996_s_at" "31877_at"
[7] "33136_at" "34762_at"
[9] "35907_at" "37228_at"
[11] "37282_at" "38414_at"
[13] "38480_s_at" "38687_at"
[15] "39855_at" "40404_s_at"
[17] "40591_at" "40727_at"
[19] "41614_at" "41623_s_at"
[21] "41624_r_at" "838_s_at"
> sapply(term.genes, function(x) intersect(x, sigGenes(GOdata)))
GO:0032446
[1,] "38687_at"
[2,] "40404_s_at"
> termStat(GOdata, "GO:0032446")
Annotated Significant
GO:0032446 22 2
Expected
GO:0032446 3.48
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
You need to provide some example code and output that shows what you are seeing.
df <- data.frame( Gene = results2$Gene, PV = results2$
Pr(>|t|), Estimate = results2$Estimate ) rownames(df) <- df$Genevalid_genes <- rownames(df) %in% keys(org.Hs.eg.db, keytype = "SYMBOL") df_filtered <- df[valid_genes, ] cat("Dimensions of filtered data frame:", dim(df_filtered), "\n")
p_values <- df_filtered$PV scores <- df_filtered$Estimate significant_condition <- p_values < 0.05 & abs(scores) > 0.5
filtered_genes <- rownames(df_filtered)[significant_condition] filtered_scores <- scores[significant_condition] filtered_pvalues <- p_values[significant_condition]
filtered_results <- data.frame( Gene = filtered_genes, Score = filtered_scores, PValue = filtered_pvalues ) head(filtered_results)
all_genes <- rownames(df_filtered) geneList <- rep(0, length(all_genes)) names(geneList) <- all_genes geneList[filtered_genes] <- 1
geneSelectionFun <- function(geneList) { return(geneList == 1) }
GOdata <- new("topGOdata", description = "GO Enrichment for Significant Genes", ontology = "BP", allGenes = geneList, geneSelectionFun = geneSelectionFun, annot = annFUN.org, mapping = "org.Hs.eg.db", ID = "SYMBOL")
result <- runTest(GOdata, algorithm = "classic", statistic = "fisher")
enriched_terms <- GenTable(GOdata, classicFisher = result, topNodes = 10)
enriched_terms$classicFisher <- as.character(enriched_terms$classicFisher) enriched_terms$classicFisher[enriched_terms$classicFisher == "< 1e-30"] <- "1e-30" enriched_terms$log_p_value <- -log10(as.numeric(enriched_terms$classicFisher))
go_terms <- AnnotationDbi::select( GO.db, keys = enriched_terms$GO.ID, columns = c("TERM"), keytype = "GOID" )
enriched_terms_full <- merge(enriched_terms, go_terms, by.x = "GO.ID", by.y = "GOID") head(enriched_terms_full)
library(ggplot2)
enriched_terms_full$classicFisher <- as.numeric(enriched_terms_full$classicFisher) enriched_terms_full$log_p_value <- -log10(enriched_terms_full$classicFisher)
genes_in_term <- genesInTerm(GOdata, whichGO = "GO:0002181")[[1]]
interesction<-intersect(filtered_genes,genes_in_term)