Search
Question: Gage and setp argument
0
gravatar for peter.chernek
18 months ago by
peter.chernek0 wrote:

Hi,

 

 

         I am trying to use gage package to do gene set enrichment analysis and I am able to get a result from the main gage function but now want to make a scatterplot and heatmap of the results. The issue I am having is how to correctly define the setp argument using the results of the main function which includes both greater than and less than. Can someone give me a clue on how to create proper input from the results. I can include a snapshot or code from my script. 

 

 

 

 

Thanks,

 

Peter Chernek, MPH

ADD COMMENTlink modified 18 months ago • written 18 months ago by peter.chernek0

I do not know how to take the results just part of the gage function so that I can use them in the setp argument of the next esset.grp function

 

results<-gage(exprs = y, ......

 

 

Example:

 

<- esset.grp(gse16873.kegg.p$greater,

ADD REPLYlink written 18 months ago by peter.chernek0
I guess you are talking about sigGeneSet() function. You can look into the documentation: ? sigGeneSet Here is the example: … gse16873.kegg.p <- gage(gse16873, gsets = kegg.gs, ref = hn, samp = dcis) gse16873.kegg.sig<-sigGeneSet(gse16873.kegg.p, outname="gse16873.kegg")
ADD REPLYlink written 18 months ago by Luo Weijun1.4k

Thank you Weijun! I will look into sigGeneSet() and send you more of my code and any error messages I am getting trying to create the scatter plot/heatmap of my results.

ADD REPLYlink written 18 months ago by peter.chernek0

Hi Weijun,

 

            My results are a list of results$greater, results$less, and results$stat.  This is just a snapshot of a  part of my results$greater so you can see yourself: As you can see each patient has a p-value and the V1 here is my own gene-set I included in my code which are specific targets we are interested in. So V1 is a gene-set of miRNA 1304 targets. The codes for the result are below along with the error I am getting using the sigGeneSet function

 

   p.geomean stat.mean p.val q.val set.size tcga-2a-a8vl-01 tcga-2a-a8vo-01 tcga-2a-a8vt-01 tcga-2a-a8vv-01
V1 0.001575478  3.019888     0     0       52     0.002227978     0.002482561    0.0005434875    0.0007809031
   tcga-2a-a8vx-01 tcga-2a-a8w1-01 tcga-2a-a8w3-01 tcga-2a-aayf-01 tcga-2a-aayo-01 tcga-2a-aayu-01
V1     0.003331894     0.003869365     0.001480415     0.002733215     0.002390523    0.0009286663
   tcga-4l-aa1f-01 tcga-ch-5737-01 tcga-ch-5738-01 tcga-ch-5739-01 tcga-ch-5740-01 tcga-ch-5741-01
V1     0.001013193     0.001880055    0.0008417054     0.000344228     0.001480244     0.002461462
   tcga-ch-5743-01 tcga-ch-5744-01 tcga-ch-5745-01 tcga-ch-5746-01 tcga-ch-5748-01 tcga-ch-5750-01
V1    0.0006472154     0.002079714     0.001131573    0.0009498034     0.001089526     0.001169852

 

 results<-gage(exprs = y, gsets = as.list(miRNA1304.targets, kegg.gs), samp = z, full.table = TRUE, saaPrep = gagePrep, saaTest = gs.tTest, same.dir = TRUE)

 

sigGeneSet(results)
Error in setp$less[, "p.val"] : incorrect number of dimensions

ADD REPLYlink written 18 months ago by peter.chernek0

> dim(results$less)
[1]   1 555
> dim(results$greater)
[1]   1 555

dim (results) 

NULL

ADD REPLYlink written 18 months ago by peter.chernek0

Hello Weijun,

 

         I am trying to now use the geneData() function to visualize my target genes. Just a background on my data:

miRNAtargets is a gene set of 61 mRNA target genes with class equal data.frame

t(z) is an expression matrix with 547 patients as columns and 1 row of values for each patient.

y is large expression matrix of mRNA with 20,500 rows of genes and 550 patients as columns.

Now I am doing:

 essGene(gs=(miRNA1304.targets), exprs= (y), ref = NULL, samp = t(z), use.fold = TRUE, rank.abs = FALSE, use.chi = TRUE, chi.p = 0.05)
Error in exprs[a, ] : invalid subscript type 'list'

   

I hope you can help with this.

 

Thanks,

 

Peter

ADD REPLYlink written 18 months ago by peter.chernek0
Peter, You can check the function documentation ?essGene Which will tell you what’s the exact format of the inputs. For Example, gs should be a character vector, your miRNA1304.targets should be in that format. Ref and samp are numeric vectors (integers), while t(z) appears to be a matrix. In addition, I woiuld suggest you get familiar with R too. That will help. -------------------------------------------- On Wed, 5/18/16, peter.chernek [bioc] <noreply@bioconductor.org> wrote: Subject: [bioc] C: Gage and setp argument To: luo_weijun@yahoo.com Date: Wednesday, May 18, 2016, 8:58 PM Activity on a post you are following on support.bioconductor.org User peter.chernek wrote Comment: Gage and setp argument: Hello Weijun,            I am trying to now use the geneData() function to visualize my target genes. Just a background on my data: miRNAtargets is a gene set of 61 mRNA target genes with class equal data.frame t(z) is an expression matrix with 547 patients as columns and 1 row of values for each patient. y is large expression matrix of mRNA with 20,500 rows of genes and 550 patients as columns. Now I am doing:  essGene(gs=(miRNA1304.targets), exprs= (y), ref = NULL, samp = t(z), use.fold = TRUE, rank.abs = FALSE, use.chi = TRUE, chi.p = 0.05) Error in exprs[a, ] : invalid subscript type 'list'     I hope you can help with this.   Thanks,   Peter Post tags: gsea, gage You may reply via email or visit C: Gage and setp argument
ADD REPLYlink written 18 months ago by Luo Weijun1.4k

Thanks Weijun, I am slowly getting better at R although I only have about a little over a year experience with it. I will check the documentation, make sure everything is correct and then get back if still getting an error.

ADD REPLYlink written 18 months ago by peter.chernek0
0
gravatar for peter.chernek
18 months ago by
peter.chernek0 wrote:


 

ADD COMMENTlink modified 18 months ago • written 18 months ago by peter.chernek0
0
gravatar for Luo Weijun
18 months ago by
Luo Weijun1.4k
United States
Luo Weijun1.4k wrote:
You only include one gene set in your analysis, hence there is only 1 row in the result matrices. sigGeneSet assumes that you have multiple gene sets, and tries to select the significant ones (rows) from the result matrices. That’s why it threw error. In your case, you don’t really need to use sigGeneSet. Your gene set was very significant based on the p.val and q.val columns (both 0’s) from results$greater (test for up-regulation). HTHs.
ADD COMMENTlink written 18 months ago by Luo Weijun1.4k

Do you think using essGene() and then geneData() would be a good approach for visualization since I have one gene set?

ADD REPLYlink modified 18 months ago • written 18 months ago by peter.chernek0
0
gravatar for peter.chernek
18 months ago by
peter.chernek0 wrote:

Ok thank you Weijun! So what would be the best way to make a scatterplot or map to visualize the results$greater to show how the set is up-regulated among the entire mRNA expression matrix? How would you interpret the greater than result? Would you say the set is significantly upregulated in comparison to the exprs=y?

ADD COMMENTlink written 18 months ago by peter.chernek0
0
gravatar for peter.chernek
18 months ago by
peter.chernek0 wrote:

please ignore this post

ADD COMMENTlink modified 18 months ago • written 18 months ago by peter.chernek0
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 205 users visited in the last hour