jefworks/liger

Lightweight Iterative Gene set Enrichment in R

liger's Issues

Add title to plot

Add a 'main' parameter to GSEA plots to allow gene set names to be printed as a title.

compare to clusterProfiler

The Bioconductor package clusterProfiler does several tests, including one that is similar to GSEA.

 The weight on enrichment score calculation

Hi, I am trying to understand GSEA algorithm by your code .
I found that in the code functions.R line 41
es <- (abs(values) ^ power) * weight
is not like the paper's formula where a weight variable is added on .
The raw formula of P_hit(S, i) in the paper is like
es <- (abs(values ^ power) .
Am I miss something ?

Tutorial Webpages Missing

Hi Dr. Fan,

It appears that the webpages for the tutorials in the ReadME no longer work (or point towards current addresses). Hoping this can be restored?

Thanks!
Sam

Different p value estimates between bulk and single gene set testing

Thanks for creating this tool, I'm finding it very useful for my analysis. I am trying to first screen a number of gene sets using the bulk option and then visualize specific ones that are significant. However, I noticed that for the same gene set the p value being shown on the plot is different from the bulk testing. I understand that these are empirical p values so the exact values may not match. However, they are different by an order of magnitude (see below). Do you have insights into why this is the case?

gsea_out = bulk.gsea(
    values = fold_changes %>% 
        filter(gene %in% expressed_genes) %>%
        filter(!is.infinite(logFC)) %>% 
        {setNames(.$logFC, .$gene)},
    set.list = h_gene_sets,
    mc.cores = 10
)

for (gs in c('INTERFERON_GAMMA_RESPONSE', 'INTERFERON_ALPHA_RESPONSE')) {

    gsea(
        fold_changes %>% 
            filter(gene %in% expressed_genes) %>% 
            filter(!is.infinite(logFC)) %>% 
            {setNames(.$logFC, .$gene)},
        h_gene_sets[[gs]],
        main = str_remove(gs, 'HALLMARK_')
    )
  
}