Question

Don't let TopGo to trimm long GO names

0

Entering edit mode

morteza.aslanzadeh ▴ 30

@mortezaaslanzadeh-23584

Last seen 4.6 years ago

I'm using TopGo to do GO analyses. When I get my enriched terms, the long term names get corrupted. I have some text then .... For example:

GO:0000122  negative regulation of transcription by ...
GO:0006886  intracellular protein transport
GO:0010976  positive regulation of neuron projection...

Any help would be appreciated?

My entire code:

library(topGO)
library(GO.db)
library(biomaRt)
library(Rgraphviz)
library(tidyverse)
library(dbplyr)

prebg_genes <- read.table("mart_mmu_protein_coding_genes.csv", header=F, sep=",")
bg_genes <- as.character(prebg_genes[,1])

pre_candidate_list_1 <- read.table("All_targets_ENSMUSG.txt", header=T, sep=",")
candidate_list_1= as.character(pre_candidate_list_1[,1])

db <- useMart('ENSEMBL_MART_ENSEMBL',dataset='mmusculus_gene_ensembl', host="www.ensembl.org")
go_ids <- getBM(attributes=c('go_id', 'ensembl_gene_id', 'namespace_1003'), filters='ensembl_gene_id', values=bg_genes, mart=db)
go_ids <- read.table("go_ids_mmu_protein_coding_genes.txt", header=T, sep="\t")
listAttributes(db)
View(go_ids)

gene_2_GO <- unstack(go_ids[,c(1,2)])

pre_keep1 <- candidate_list_1 %in% go_ids[,2]
keep1 <- which(pre_keep1==TRUE)
candidate_list_final1 <- candidate_list_1[keep1]

geneList1 <- factor(as.integer(bg_genes %in% candidate_list_final1))
names(geneList1) <- bg_genes

GOdata1 <- new('topGOdata', ontology='BP', allGenes = geneList1, annot = annFUN.gene2GO, gene2GO = gene_2_GO)

classic_fisher_result1 <- runTest(GOdata1, algorithm='classic', statistic='fisher')

weight_fisher_result1 <- runTest(GOdata1, algorithm='weight01', statistic='fisher') 

applied to the topGOdata object.
allGO1 <- usedGO(GOdata1)
all_res1 <- GenTable(GOdata1, weightFisher=weight_fisher_result1, orderBy='weightFisher', topNodes=length(allGO1))

p.adj1 <- round(p.adjust(all_res1$weightFisher,method="BH"),digits = 4)

pre_all_res_final1=cbind(all_res1,p.adj1)
all_res_final1 <- pre_all_res_final1[order(pre_all_res_final1$p.adj1),]
write.table(all_res_final1,"summary_topGo_analysis_all_DE_miRs_targets.csv",sep=",",quote=F)

go software error topgo rna-seq • 2.7k views

ADD COMMENT • link updated 23 months ago by James W. MacDonald 67k • written 4.6 years ago by morteza.aslanzadeh ▴ 30

1

Entering edit mode

Has anyone had a similar problem with showSigOfNodes? The GO term definitions are getting cut off. I know they are long but I would at least like the option to display the full definition. Here is an example of my problem: TopGO showSigOfNodes graph with GO term definitions cut off

ADD REPLY • link 2.6 years ago Marie ▴ 10

0

Entering edit mode

Hi, have you found a solution to this?

ADD REPLY • link 23 months ago ar14g12 • 0

0

Entering edit mode

Did you not notice that this has been answered as part of this thread that you posted in?

ADD REPLY • link 23 months ago James W. MacDonald 67k

0

Entering edit mode

Crossposted on biostars: https://www.biostars.org/p/440805/#440807. Could you show head(all_res1) and all_res1$Name or similar where the terms are stored?

ADD REPLY • link 4.6 years ago Lluís Revilla Sancho ▴ 760

0

Entering edit mode

head(all_res1) gives

GO.ID                 Term                            Annotated  Significant Expected weightFisher
GO:0045944 positive regulation of transcription by ...      1200          48    16.49      8.1e-11
GO:0010976 positive regulation of neuron projection...       360          20     4.95      3.2e-05
GO:1900748 positive regulation of vascular endothel...         6           3     0.08      5.0e-05
GO:2000543         positive regulation of gastrulation         8           3     0.11      0.00014
GO:0070936           protein K48-linked ubiquitination        58           6     0.80      0.00014
GO:0007638                     mechanosensory behavior        15           4     0.21      0.00019

and all_res1$Name is NULL

ADD REPLY • link 4.6 years ago morteza.aslanzadeh ▴ 30

0

Entering edit mode

For example the GO:0045944 positive regulation of transcription by ... should be written in the full name like GO:0045944 positive regulation of transcription by RNA polymerase II

ADD REPLY • link 4.6 years ago morteza.aslanzadeh ▴ 30

score 3 · Answer 1 · 2020-05-29

3

Entering edit mode

morteza.aslanzadeh ▴ 30

@mortezaaslanzadeh-23584

Last seen 4.6 years ago

I found the answer in Github. So basically we have to add numChar=1000 to all_res1 i.e:

all_res1 <- GenTable(GOdata1, weightFisher=weight_fisher_result1, orderBy='weightFisher', topNodes=length(allGO1), numChar=1000)

numChar: truncates GO term descriptions at 1000 chars (basically, disables truncation)