I want to use the groupGO function from cluster profiler, but do not exactly understand the level argument. The help page (hence ?groupGO) says: "Specific GO level." When I play around with it (giving it different numbers), I still don't seem to understand what and how this argument works.

> ggo_all <- groupGO(gene = All_genes, keytype = "ENSEMBL", OrgDb = org.Hs.eg.db, ont = "BP", level = 20, readable = F)
> dim(as.data.frame(ggo_all))
[1] 2 5
> ggo_all <- groupGO(gene = All_genes, keytype = "ENSEMBL", OrgDb = org.Hs.eg.db, ont = "BP", level = 2, readable = F)
> dim(as.data.frame(ggo_all))
[1] 33  5
> ggo_all <- groupGO(gene = All_genes, keytype = "ENSEMBL", OrgDb = org.Hs.eg.db, ont = "BP", level = 3, readable = F)
> dim(as.data.frame(ggo_all))
[1] 585   5
> ggo_all <- groupGO(gene = All_genes, keytype = "ENSEMBL", OrgDb = org.Hs.eg.db, ont = "BP", level = 10, readable = F)
> dim(as.data.frame(ggo_all))
[1] 13782     5
> ggo_all <- groupGO(gene = All_genes, keytype = "ENSEMBL", OrgDb = org.Hs.eg.db, ont = "BP", level = 15, readable = F)
> dim(as.data.frame(ggo_all))
[1] 1046    5

What I would like to do is, to get all GO terms for my genes. But I don't understand what the level argument should be set on to achieve this.

> sessionInfo()
R version 3.4.3 (2017-11-30)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.1252
[2] LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
 [1] grid      parallel  stats4    stats     graphics  grDevices utils    
 [8] datasets  methods   base     

other attached packages:
 [1] Rgraphviz_2.22.0      topGO_2.30.1          SparseM_1.77         
 [4] GO.db_3.5.0           graph_1.56.0          BiocInstaller_1.28.0
 [7] org.Hs.eg.db_3.5.0    AnnotationDbi_1.40.0  IRanges_2.12.0       
[10] S4Vectors_0.16.0      Biobase_2.38.0        BiocGenerics_0.24.0  
[13] clusterProfiler_3.6.0 DOSE_3.4.0           

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.15        pillar_1.2.1        compiler_3.4.3     
 [4] plyr_1.8.4          tools_3.4.3         digest_0.6.15      
 [7] bit_1.1-12          lattice_0.20-35     RSQLite_2.0        
[10] memoise_1.1.0       tibble_1.4.2        gtable_0.2.0       
[13] pkgconfig_2.0.1     rlang_0.2.0         fastmatch_1.1-0    
[16] igraph_1.1.2        DBI_0.8             rvcheck_0.0.9      
[19] fgsea_1.4.1         gridExtra_2.3       stringr_1.3.0      
[22] bit64_0.9-7         glue_1.2.0          qvalue_2.10.0      
[25] data.table_1.10.4-3 BiocParallel_1.12.0 GOSemSim_2.4.1     
[28] purrr_0.2.4         tidyr_0.8.0         DO.db_2.9          
[31] ggplot2_2.2.1       reshape2_1.4.3      blob_1.1.0         
[34] magrittr_1.5        matrixStats_0.53.1  splines_3.4.3      
[37] scales_0.5.0        colorspace_1.3-2    labeling_0.3       
[40] stringi_1.1.6       lazyeval_0.2.1      munsell_0.4.3

The Gene Ontology is a directed acyclic graph, so there are levels to the graph. At level 1, for e.g., the BP ontology, there is only one term, biological process. At the next level, there are something like 33 terms, that are more specific than biological process, but less specific than the next level. The lower the level, all things equal, the more specific the term becomes, and the more terms there are, and the fewer genes there are for a term. Using data from example(groupGO),

> as.data.frame(groupGO(gcSample[[1]], "org.Hs.eg.db", ont = "BP", level = 1))[,1:2]
   ID        Description
GO:0008150 biological_process

And if we choose level 2

> z <- as.data.frame(groupGO(gcSample[[1]], "org.Hs.eg.db", ont = "BP", level = 2))
> dim(z)
[1] 33  5

> z[1:10,2]
 [1] reproduction                                                  
 [2] metabolic process                                             
 [3] cell killing                                                  
 [4] immune system process                                         
 [5] sulfur utilization                                            
 [6] phosphorus utilization                                        
 [7] growth                                                        
 [8] behavior                                                      
 [9] cell proliferation                                            
[10] carbohydrate utilization                                      

And if we choose level 3

> zz <- as.data.frame(groupGO(gcSample[[1]], "org.Hs.eg.db", ont = "BP", level = 3))
> dim(zz)
[1] 585   5
> zz[1:10,2]
 [1] sexual reproduction                     
 [2] asexual reproduction                    
 [3] reproductive process                    
 [4] multicellular organism reproduction     
 [5] reproduction of a single-celled organism
 [6] reproduction of symbiont in host        
 [7] oxidation-reduction process             
 [8] nitrogen compound metabolic process     
 [9] catabolic process                       
[10] biosynthetic process          

And note that the first 6 terms here are children of 'reproduction', which itself is a biological process.