Plotting GO annotation results

Question

[seta]

Hi all friends,

I used TRAPID for GO annotation of a plant de novo assembled transcriptome, the output is a large text file (about 16 MB) like below:

 ab29703    GO:0019219
ab29703 GO:0009059
ab29707 GO:0044446
ab29707 GO:0006810
ab29707 GO:0044424
Contig6742  GO:0044260
Contig6742  GO:0003824
Contig6742  GO:0016772

I plan to use WEGO (enter link description here) for categorizing and plotting the GO annotation results, but the input of WEGO is something like:

ab29703 GO:0019219  GO:0009059
ab29707 GO:0044446  GO:0006810  GO:0044424
Contig6742  GO:0044260  GO:0003824  GO:0016772

Could you please help me out how I can change the original output format to accept by WEGO? Any comments and suggestion for using other tools for plant GO annotation and plotting would be highly appreciated.

Answer 1

[Benn]

Hi Seta,

I had a kinda similar question on the bioconductor site https://support.bioconductor.org/p/77134/

You can use dplyr in R to get what you want.

Good luck!

Ben

[seta]

Thanks for your reply. However, my question differed with yours; actually, I would like to have all GO term from one gene in every row of a text file. Please let me know if you have any suggestion.

Best

[Benn]

Yes, but the principle is the same right? Switch the genes with the go terms?

[seta]

Sorry, I am not much familiar with R. Could you please put the appropriate command for using "dplyr" to get GO term from one gene in every row of a text file?

Thanks for your help,

[Benn]

Assuming you have a tab-del txt file named "exampl.txt"

library(dplyr)

df<-read.table("exampl.txt")

reshaped <- group_by(df, V1) %>% summarise(GO = paste(V2, collapse = "\t"))

write.table(reshaped, "reshaped.txt", sep = "\t", quote = FALSE, row.names = FALSE, col.names = FALSE)

[seta]

Thank you very much b.nota. I try it.

[seta]

I tried your suggested command, it worked well.

again thanks

Answer 2

[ivivek_ngs]

Are you only restricted to using WEGO? you can simply do it yourself as well using R or as @b.nota said. In that case you will have to write a small script to count the number of GO terms each gene is having and then print that matrix and make any plots per row for those those genes , wither a barplot or even a heatmap.

[seta]

Thanks, vchris_ngs. No, I'm not restricted to use WEGO. Regarding your comment, I also have another txt file with below information in every row:

go  description     num_transcripts transcripts

GO:0000033  alpha-1,3-mannosyltransferase activity  3   Contig12621 Contig15208 ab76639

In fact, there is the number of transcript with a given GO term in this file. however, I don't know how to use it for plotting.

[ivivek_ngs]

Now you have it , for each gene the corresponding GO terms. as you put your example.

[seta]

No, I have a given GO term for multiple genes, while I want to have all corresponding GO terms for each gene in every row of a txt file.

[ivivek_ngs]

Thats what I wrote that b.nota has already shown it, sorry I did not mention his name in my comment.