clusterProfiler: BgRatio N value does not equal set gene universe size
1
0
Entering edit mode
@lindseypieper-22053
Last seen 23 months ago

I have a clarifying question for clusterProfiler's enrichKegg function. I am setting my own universe based on Kegg gene IDs I generated using Koala Blast. When I run enrichKegg, the Bg Ratio denominator, N, is less than the total of unique Kegg gene IDs in my set background universe (1511 vs 2863). From my understanding this should be equal. Has anyone else seen this issue or know what is going on? I've included by code and the output below with bolded BgRatios. Thanks!

koalaList %>% group_by(K_geneID) %>% summarize(Nobserved=n()) %>% arrange(desc(Nobserved)) #2863 features
test <- mergeList.wtvfnr.dmso.midexp %>%
  filter(padj<0.1 & abs(log2FoldChange)>(1)) %>%
  filter(!is.na(K_geneID)) %>%
  group_by(K_geneID) %>% summarize(Nobserved=n()) %>% arrange(desc(Nobserved)) #623 sig de features

enriched<-
mergeList.wtvfnr.dmso.midexp %>%
  filter(padj<0.1 & abs(log2FoldChange)>(1)) %>%
  filter(!is.na(K_geneID)) %>%
  pull(K_geneID) %>%
  as.character() %>%
  enrichKEGG(gene=., organism="ko", universe=koalaList$K_geneID)

enriched@result %>% as.tibble() %>% dplyr::select(-geneID) %>% arrange(desc(Count))

ID Description GeneRatio BgRatio pvalue p.adjust qvalue Count <chr> <chr> <chr> <chr> <dbl> <dbl> <dbl> <int> 1 ko01120 Microbial metabolism in diverse environments 89/369 235/1511 0.000000366 0.00000376 3.08e⁻ ⁶ 89 2 ko02010 ABC transporters 64/369 176/1511 0.000108 0.000781 6.39e⁻ ⁴ 64 3 ko01200 Carbon metabolism 43/369 94/1511 0.00000252 0.0000227 1.86e⁻ ⁵ 43

clusterProfiler • 941 views
ADD COMMENT
1
Entering edit mode
Guangchuang Yu ★ 1.2k
@guangchuang-yu-5419
Last seen 7 months ago
China/Guangzhou/Southern Medical Univer…

the secret is here, https://github.com/YuLab-SMU/DOSE/blob/master/R/enricher_internal.R#L59.

If the provided universe contains genes that have no annotation, they will be filtered out. Otherwise, it will downgrade the p value.

ADD COMMENT
0
Entering edit mode

many software do the wrong things to give you smaller p values, but clusterProfiler do not.

ADD REPLY
0
Entering edit mode

Thank you for the clarification.

ADD REPLY
0
Entering edit mode

Thank you for the clarification.

ADD REPLY

Login before adding your answer.

Traffic: 204 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

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

Powered by the version 2.3.6