CummeRbund: unable to use getGenes to retrieve short names in R

Question

[CandiceChuDVM]

Hi all,

I am now using CummeRbund in R (version 3.2.4) to visualize my Cuffdiff (version 2.2.1) outputs.

I was hoping to get short names such as MMP9... for my genes. However, I can only get NCBI Reference Sequence such as NM_001002930 NM_001002932 NM_001002938...etc.

Here is my R script:

cuff_data <- readCufflinks(dir=refCuffdiff,rebuild=T,gtfFile=gtfFilePath,genome=genomePath)
diffGeneIDs <- getSig(cuff_data,level="genes",alpha=0.01)    
diffGenes<-getGenes(cuff_data,diffGeneIDs)
featureNames(diffGenes)

Here is part of my output:

840  XLOC_024368                                   <NA>  
841  XLOC_024378                                   <NA>  
842  XLOC_024418                                   <NA>  
843  XLOC_024432                                   <NA>  
844  XLOC_024434                                   <NA>  
845  XLOC_024442                                   <NA>  
846  XLOC_024444                           NM_001003241  
847  XLOC_024451                           NM_001003219  
848  XLOC_024474                           NM_001197143  
849  XLOC_024482                                   <NA>  
850  XLOC_024503                              NR_128749

(1) I was wondering why most of are <na>. Is it caused by the lack of annotation in dog genome (CanFam3.1.gtf)? Or it has something to do with the addFeatures function?

(2) For the sample that has an ID, I was hoping to get short names instead of the NCBI Reference Sequence. How can I achieve my goal?

I have checked this post on SEQanswer but it couldn't solve my problem.

Thanks a lot!!!

[Sukhi Singh]

The problem is coming from your annotation file that you provided to Cuffdiff. Check whether it has the gene names there? For a quick test, do head gene_exp.diff inside the folder where you store the cuffdiff results, I reckon there are no gene names there only.

[CandiceChuDVM]

Yes. You are right. I have been working with dog genome, and I downloaded the gtf file from UCSC. The head gene_exp.diff result looks like this:

 test_id    gene_id gene    locus   sample_1    sample_2    status  value_1 value_2 log2(fold_change)   test_stat   p_value q_value significant
 XLOC_000001    XLOC_000001 -   chr1:722268-735387  rapid   slow    OK  8.01499 9.6448  0.267051    0.375532    0.52065 0.998058    no
 XLOC_000002    XLOC_000002 -   chr1:758468-805972  rapid   slow    OK  8.42177 10.402  0.304661    0.388298    0.49805 0.998058    no
 XLOC_000003    XLOC_000003 -   chr1:2710923-2742056    rapid   slow    OK  3.31617 2.22304 -0.576982   -0.664651   0.25485 0.998058    no
 XLOC_000004    XLOC_000004 -   chr1:2837400-2921732    rapid   slow    OK  7.15783 7.10424 -0.0108414  -0.0171568  0.9757  0.998058    no
 XLOC_000005    XLOC_000005 -   chr1:3053815-3066829    rapid   slow    OK  5.6806  4.76451 -0.253717   -0.370415   0.5147  0.998058    no
 XLOC_000006    XLOC_000006 -   chr1:3275469-3391717    rapid   slow    OK  7.30296 6.84825 -0.0927457  -0.104054   0.8612  0.998058    no
 XLOC_000007    XLOC_000007 -   chr1:4304020-4312108    rapid   slow    OK  5.7179  5.45483 -0.067951   -0.0962702  0.8682  0.998058    no
 XLOC_000008    XLOC_000008 NM_001193298    chr1:5070802-5112269    rapid   slow    OK  11.1779 16.6043 0.57091 0.452823    0.4478  0.998058    no
 XLOC_000009    XLOC_000009 -   chr1:8181111-8305073    rapid   slow    OK  7.05367 5.78846 -0.285193   -0.407552   0.5128  0.998058    no

Does that mean most of the gene have not been annotated?

Answer 1

[Sukhi Singh]

The GTF file you used with Cuffdiff, lacks this information. Use this annotation file from Ensembl, and redo the cuffdiff step, will solve your problem.

More information about Dog genome: http://www.ensembl.org/Canis_familiaris/Info/Index