UserGuide: GO Enrichment Analysis

GO Enrichment Analysis

We uses the GOs annotations file to perform enrichment analysis on differentially expressed gene. For this, you define several parameters:

1) for enrichment

• parameters$GO_threshold  the significant threshold used to filter p-values
• parameters$GO_min_num_genes  the minimum number of genes for each GO terms in the genome
• parameters$GO  gene set chosen for analysis "up", "down", "both" (up+down)
• parameters$GO_algo  algorithms for runTest function ("classic", "elim", "weight", "weight01", "lea", "parentchild")
• parameters$GO_stats  statistical tests for runTest function ("fisher", "ks", "t", "globaltest", "sum", "ks.ties")

2) for visualization

• parameters$Ratio_threshold  the min enrichment ratio to display GO in graph
• parameters$GO_max_top_terms  the maximum number of GO terms plot for each GO category
• parameters$GO_min_sig_genes  the minimum number of significant gene(s) behind the enriched GO-term to display GO in graph

If provided by the user, genes are linked to Gene Ontology (GO) annotations in the GO annotation file (see Input files description section) making GO enrichment analysis possible. Each gene can be annotated with several terms which define biological pathways in which the genes are involved. GO terms are classified into 3 categories (MF describing the molecular activity of a gene, BP describing a broader biological process in which the gene is involved in coordination with other genes, and CC describing the cellular location in which the gene performs its function).

GO enrichment is automatically performed using the topGO package on the DE genes (up, down, or both, depending on the parameters$GO) of each contrast of the experiment. The developed GOenrichment function generates numerous tables and plots.

Be careful: As enrichment tests are based on proportions, the lower the number of genes, the less reliable the test is. The interpretation of the enrichment is then hazardous. Do not draw hasty conclusions for gene lists of less than 100 genes.

The commands for running GO enrichment analysis are:

# Parameters for GO enrichment
parameters$GO_threshold = 0.05
parameters$GO_min_num_genes = 10
parameters$GO = "both" 
parameters$GO_algo = "weight01" 
parameters$GO_stats = "fisher" 

# Parameters for GO enrichment graphs
parameters$Ratio_threshold = 1
parameters$GO_max_top_terms = 10
parameters$GO_min_sig_genes = 5  

# run analysis
GOenrichment(resDEG, data, parameters)

A "DEG_test/GOenrichment/TOTAL_DEgenes/" directory will be created with all GO images and tables of statistics for results of both (up and down) DE genes (parameters$GO = "both" ).

A "DEG_test/GOenrichment/UP_DEgenes/" directory will be created with all GO images and tables of statistics for results of DE UP regulated genes (parameters$GO = "up" ).

A "DEG_test/GOenrichment/DOWN_DEgenes/" directory will be created with all GO images and tables of statistics for results of DE DOWN regulated genes (parameters$GO = "down" ).

Graphical outputs

Global graphs

Global graphs are created, for each contrast, highlighting either the p-value or the enrichment ratio in the 3 GO categories.

Example of global graphs:

Detailed graphs for each GO category

To satisfy the maximum number of users, individualized graphs by GO category are also created, regrouping the p-value and enrichment ratio information.

Example of detailed graphs:

Tabular outputs

Global ouput

Example of one statistical table:

GO.ID	Term	Annotated	Significant	Expected	statisticTest	Ratio	GO_cat
GO:0003735	structural constituent of ribosome	135	51	31.31	0.000086	1.628873	MF
GO:0004812	aminoacyl-tRNA ligase activity	51	24	11.83	0.000150	2.028741	MF
GO:0019843	rRNA binding	23	13	5.33	0.000570	2.439024	MF
...

Explications of some columns:

• Annotated: number of genes in your genome (the gene universe) annotated with the GO-term.
• Significant: number of genes in the list annotated with the GO-term.
• Expected: number of genes expected in the list if the proportion of the genes in the list was quite equal to its proportion in the gene universe (meaning no enrichment)
• statisticTest: p-value of the statistic test chosen
  
  

Tabular outputs for each enriched GO

Sub-directories are created in "DEG_test/GOenrihcment/TOTAL-UP-DOWN_DEgenes/" directories for each contrast in which the user can find many tables resuming informations on the genes behind each enriched GO-term (name of the genes, their description if an annotation file is provided, their DE status in all contrasts, and their normalized expression in CPM in all experimental conditions).

Example:

Gene	Gene_description	AC1vsAC2	AC1vsAC3	AC2vsAC3	BC1vsBC2	BC1vsBC3	BC2vsBC3	AC1vsBC1	AC2vsBC2	AC3vsBC3	AC1	AC2	AC3	BC1	BC2	BC3	GO_ID	GO_term	GO_cat
Gene_002357	phenylalanine-trna ligase beta subunit	1	0	0	0	0	0	0	-1	-1	266.349500	210.51620	221.547300	254.519300	266.1183000	300.3326000	GO:0000162	tryptophan biosynthetic process	BP
Gene_002384	tyrosine-trna ligase	1	0	0	0	0	0	0	-1	0	64.172870	51.79736	51.085950	60.415040	71.3212400	61.5877900	GO:0000162	tryptophan biosynthetic process	BP
Gene_003773	tyrosyl-trna synthetase	1	0	0	0	0	0	0	-1	-1	375.692400	300.71430	306.424000	333.957000	380.2730000	402.9832000	GO:0000162	tryptophan biosynthetic process	BP
...

Finally, for each analysis, a "NameOfTheContrast_SignificantGO" directory is created in which you can find, for each enriched GO-term, the genes that enabled enrichment.