High pvalues when using clusterProfiler for seurat
Entering edit mode
10 weeks ago
Ahmed • 0

Hi, I am trying to run clusterProfiler::GSEA version 4.8.3 for each cluster of my SeuratObj

When ranking the DEG based on logFC I get decent ES/NES scores however my q/p values are usually > 0.05 sometimes even >0.5. However, when I re-run the code but use the area under the curve (auc) to rank my DEG I get way better NES and q/p values but worse ES scores.

Would appreciate your thoughts on this and which route would you suggest to follow.

    library(clusterProfiler )
    library(msigdbr )
    m_t2g <-m_df<- msigdbr(species = "Mus musculus", category = "C7") %>%  
    dplyr::select(gs_name, gene_symbol) %>% filter(grepl("DP|Tcell|TCELL|T_cell|THYMOCYTES|TH1|TH2|TH17|TCR|NKT", gs_name, ignore.case = TRUE))
    markers <- wilcoxauc(SeuratObj, 'seurat_clusters')
    cluster_genes <- markers %>%  filter(group == as.character(3))  %>% filter(padj <1) %>% arrange(desc(logFC)) %>% select(feature, logFC)
    ranks <- deframe(cluster_genes)
    rescp <- clusterProfiler::GSEA(ranks, TERM2GENE = m_t2g, eps = 0.0, minGSSize = 20, pvalueCutoff = 1)

LogFC rankingenter image description here

auc rankingenter image description here

clusterProfiler • 228 views

Login before adding your answer.

Traffic: 2435 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6