CEMiTool: error in overrepresentation analysis (Enrichment is null)/mod_ora function
0
1
Entering edit mode
June. ▴ 10
@june-23879
Last seen 17 months ago

Hello,

I'd like to use CEMiTool for ORA analysis for budding yeast. I downloaded GO gene sets with Enrichment Browser, read it with read_gmt function and obtained more than 19000 entrez IDs and GO terms. Then, I converted those IDs to gene names and I found many row repeats of gene sets and NAs all over this new GMT file, reaching to around 39,000 rows. I removed those extra rows and ended up with around 3-4 thousand gene sets. But, using mod_ora on this final GMT file, I get the following error and I don't know why?

No gene set have size > 10 ...
--> return NULL...
No gene set have size > 10 ...
--> return NULL...
No gene set have size > 10 ...
--> return NULL...
Warning messages:
1: In FUN(X[[i]], ...) : Enrichment for module M1 is NULL
2: In FUN(X[[i]], ...) : Enrichment for module M2 is NULL
3: In FUN(X[[i]], ...) : Enrichment for module M3 is NULL
4: In .local(cem, ...) :
Enrichment is NULL. Either your gmt file is inadequate or your modules really aren't enriched for any of the pathways in the gmt file.


By the way, initially I couldn't find modules and I changed set_beta to 2 to see if it differs and showed me 3 modules. I don't know what I did was correct and if the current error is related to that or not and how to solve it?

CEMiTool • 212 views
0
Entering edit mode

Hi, June. Are you certain that the gene names in your expression file and your gmt file match? That's what the warning message is saying when it means your gmt file might be inadequate. Otherwise, as the warning message also states, there might not be any gene sets in your gmt file which contain at least 10 genes of your modules. Also, setting your beta value to 2 is quite low, you might want to first look at the beta x R2 curve by using the following code after running the cemitool function:

cem <- plot_beta_r2(cem)
show_plot(cem, "beta_r2")


Ideally, you'd select a low-ish beta value which stabilizes the R2 values at around 0.8.