confusing results of enrichKEGG (clusterProfiler)

Question

[Ann]

I ran a KEGG enrichment analysis for co-expression clusters on a non-model organism using enrichKEGG. The list of top 5 enriched pathways looks good and seems biologically plausible in the context of our experiment, but the numbers in GeneRatio and BgRatio are confusing.

universe_set -- a vector of all unique KOs annotated with KofamKOALA for my de novo transcriptome. gene_set -- a vector of all unique KOs for my gene set (transcripts from one of the co-expression clusters)

Test code:

# TEST CODE
gene_set <- c("K23540", "K23487", "K22188")
length(gene_set) # 3

universe_set <- c("K23540", "K23487", "K22188", "K06843", "K04678", "K17601", "K22037", "K23897", "K04675", "K04437", "K09191", "K00476", "K23437", "K12487", "K09292", "K21469", "K06639", "K20523", "K07366", "K14688", "K04811", "K24738", "K12026", "K25615", "K10049", "K24496", "K06238", "K22647", "K19716", "K17613", "K13100", "K12261", "K22156")
length(universe_set) # 33

kegg_kofam <- enrichKEGG(gene = gene_set, 
                universe=universe_set,
                organism="ko",
                keyType='kegg',
                pAdjustMethod = "BH",
                minGSSize = 1,
                maxGSSize = 1000,
                pvalueCutoff = 1,
                qvalueCutoff = 1, 
                use_internal_data = F)

kegg_kofam_DF <- kegg_kofam[,1:9]

My gene_set vector has length 3 and my universe_set vector is 33. gene_set vector is part of universe_set.

And I expected to get GeneRatio k/3 and BgRatio M/33, but I get GeneRatio k/1 and BgRatio M/17.

Where could there be an error or snag?

[ATpoint]

The effective universe is the intersect between universe and the genes annotated in the pathway database. That is apparently just 17 out of 33 genes.