package 'ChAMP' demo/test code fails at champ.CNA()
Entering edit mode
Last seen 3.0 years ago
United States

hi; Running the 'ChAMP' demo with its test data per vignette instructions ending with :

champ.process(directory = testDir)

all seems to work until this happens:


Preparing files for ComBat
Your data is being logit transformed before batch correction
Beginning batch correction
Found 3 batches
Found 1  categorical covariate(s)
Standardizing Data across genes
Fitting L/S model and finding priors
Finding parametric adjustments
Adjusting the Data
Your data is being inverse logit transformed
Run Probe Lasso DMR Hunter
[1] "contrast C T"
You have found 97281 significant MVPs with a BH adjusted P-value below 0.05
You have found 2933 DMRs.
You have found 2761 significant DMRs with a dmr.p < 0.05.
Run champ.CNA
You have chosen Control as the reference and this does not exist in your sample sheet (column Sample_Group). The analysis will run with ChAMP blood controls.
champ.CNA is using the samples you have defined as champCtl as the reference for calculating copy number aberrations.
As you are using the ChAMP controls Combat cannot adjust for batch effects. Batch effects may affect your dataset.
Saving Copy Number information for each Sample
Error in sort(abs(diff(genomdat)))[1:n.keep] :
  only 0's may be mixed with negative subscripts


Any thoughts?


Hamid Bolouri

> traceback()
4: trimmed.variance(genomdat[ina], trim)
3: smooth.CNA(CNA.object)
2: champ.CNA(intensity = myLoad$intensity, pd = myLoad$pd, batchCorrect = batchCorrect,
       resultsDir = resultsDir, sampleCNA = sampleCNA, plotSample = plotSample,
       groupFreqPlots = groupFreqPlots, freqThreshold = freqThreshold)
1: champ.process(directory = testDir)


> sessionInfo()

R version 3.1.2 (2014-10-31)
Platform: x86_64-unknown-linux-gnu (64-bit)

 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C
 [9] LC_ADDRESS=C               LC_TELEPHONE=C

attached base packages:
[1] stats4    parallel  stats     graphics  grDevices utils     datasets
[8] methods   base

other attached packages:
 [1] IlluminaHumanMethylation450kmanifest_0.4.0
 [2] ChAMP_1.4.0
 [3] Illumina450ProbeVariants.db_1.1.1
 [4] ChAMPdata_1.1.2
 [5] minfi_1.12.0
 [6] bumphunter_1.6.0
 [7] locfit_1.5-9.1
 [8] iterators_1.0.7
 [9] foreach_1.4.2
[10] Biostrings_2.34.1
[11] XVector_0.6.0
[12] GenomicRanges_1.18.4
[13] GenomeInfoDb_1.2.4
[14] IRanges_2.0.1
[15] S4Vectors_0.4.0
[16] lattice_0.20-29
[17] Biobase_2.26.0
[18] BiocGenerics_0.12.1
[19] BiocInstaller_1.16.1

loaded via a namespace (and not attached):
 [1] annotate_1.44.0       AnnotationDbi_1.28.1  base64_1.1
 [4] beanplot_1.2          cluster_1.15.3        codetools_0.2-9
 [7] DBI_0.3.1             digest_0.6.4          DNAcopy_1.40.0
[10] doRNG_1.6             genefilter_1.48.1     grid_3.1.2
[13] illuminaio_0.8.0      impute_1.40.0         limma_3.22.1
[16] marray_1.44.0         MASS_7.3-35           Matrix_1.1-4
[19] matrixStats_0.12.2    mclust_4.4            mgcv_1.8-3
[22] multtest_2.22.0       nlme_3.1-118          nor1mix_1.2-0
[25] pkgmaker_0.22         plyr_1.8.1            preprocessCore_1.28.0
[28] quadprog_1.5-5        R.methodsS3_1.6.1     RColorBrewer_1.0-5
[31] Rcpp_0.11.3           registry_0.2          reshape_0.8.5
[34] rngtools_1.2.4        RPMM_1.20             RSQLite_1.0.0
[37] siggenes_1.40.0       splines_3.1.2         stringr_0.6.2
[40] survival_2.37-7       sva_3.12.0            tools_3.1.2
[43] wateRmelon_1.6.0      XML_3.98-1.1          xtable_1.7-4
[46] zlibbioc_1.12.0


software error • 2.1k views
Entering edit mode
schasse • 0
Last seen 7.6 years ago
United States

I'v run into the same exact situation. I haven't been able to sort this out. I have actual data sets that run as well but they fail differently and at different steps in the pipeline. I'm beginning to suspect that the install is faulty.

Let me know if you've solved the problem. Thanks.

-Scott Chasse-

Entering edit mode

I tried to alert the ChAMP maintainer to this thread but my email bounced; if anyone knows how to contact them, let me know.

Entering edit mode

I did get ahold of the maintainer and advise them about this thread; I hope they will reply soon.

Entering edit mode
Thank you so much! -Scott- ________________________________ From: Dan Tenenbaum [bioc] [] Sent: Friday, March 06, 2015 4:32 PM To: Chasse, Scott A Subject: [bioc] C: package 'ChAMP' demo/test code fails at champ.CNA() Activity on a post you are following on<https:""> User Dan Tenenbaum<https:"" u="" 4256=""/> wrote Comment: package 'ChAMP' demo/test code fails at champ.CNA()<https:"" p="" 64133="" #65485="">: I did get ahold of the maintainer and advise them about this thread; I hope they will reply soon. ________________________________ You may reply via email or visit C: package 'ChAMP' demo/test code fails at champ.CNA()
Entering edit mode
yan.marky • 0
Last seen 7.7 years ago
United States


Has this issue been resolved for anyone? I still couldn't get the champ.CNA() function to work properly. I suppose this function should generate some sort of matrix containing copy numbers but the demo code for this function doesn't seem to produce any results. 


Please let me know if it is working for you guys, thanks!



Entering edit mode
Last seen 4.5 years ago

Changing all "C" values in the "Sample_Group" column to "Control"  in the csv file worked for me.

Entering edit mode

Where did you find out what columns are needed and what values they should have? That's very shady! I am trying to run it with tue "built-in" controls and guess that there may be the problem? 

Entering edit mode

But it looks like in champ.MVP (which is called by champ.lasso), it is expected (but not required) that the Sample_Group column contains the codes "C" and "T" for 'controls' and 'cases'.  

> champ.MVP
function (beta.norm = myNorm$beta, pd = myLoad$pd, adjPVal = 0.05, 
    adjust.method = "BH", = c("C", "T"), resultsDir = paste(getwd(), 
        "resultsChamp", sep = "/"), bedFile = TRUE) 
    groupLabel =
    checkLabel = unique(pd$Sample_Group)
    if (any(!(groupLabel %in% checkLabel))) {
        message("The group labels that have been defined ", groupLabel, 
            " do not exist in your sample sheet. Please edit the Sample_Group column or the parameter. ChAMP will use information in your Sample_group columnn.")
        groupLabel = checkLabel
    if (length(groupLabel) > 2) {
        message("Your dataset has more than two groups. ChAMP will compare the first two groups.")
    controls = pd[which(pd$Sample_Group == groupLabel[1]), ]
    test = pd[which(pd$Sample_Group == groupLabel[2]), ]

If you don't use "C" and "T" values in the Sample_Group column, then which group is mapped to the 'controls' group and which is mapped to the 'test' group seems to depend on the order of values in the Sample_Group column. For example:

>  tt <- c("Case","Case","Control")
> unique(tt)
[1] "Case"    "Control"
>  tt <- c("Control","Case","Case","Control")
> unique(tt)
[1] "Control" "Case" ​
Entering edit mode
apmishra • 0
Last seen 6.2 years ago

I'm running into the same problem and I'm using the built-in blood controls when calling champ.CNA so I cannot change the Sample Group values. Does anyone know how I can solve this error?

Entering edit mode

This thread is old - please post a new question with the code you've tried and your sessionInfo(). Tag the question with 'ChAMP. 


Entering edit mode
Yuan Tian ▴ 90
Last seen 3.2 years ago



ChAMP has been updated to version 2.6 and nearly all code has been recoded. I just know this thread was published two years ago...


Firstly I suggest you use champ functions step by step instead of champ.process() so that we could know which part might be wrong. 

Secondly I have tested champ.CNA before and see no problems. Please try nesest version package and see if you still have error. If you do please post the latest error message.

Thirdly yes in ChAMP, some comparision groups are compared by order of sample_group labels. But since we are focuing are finding difference between groups, the statistic and significant p value should be still correct.

Fourthly I remember in champ.CNA() there are two ways to calculate CNA. One is compared all sample to certain "control" pheno(there is a control dataset in package for 450k as well). And the other method is compare each sample to their mean values.Finally champ.CNA() will return plots. I don't remember sample_group's name would influence the result. Could you provide more error information on that?

Finally, sorry for not replying this thread...It has been published two years ago when I was still an undergraduate student.

I am in Spring Festival vocation now so I missed the reminder email. Please send your error message and question to Though I just become maintainer for couple month, I will see if I can help. But please do try use newest version package cause no one knows how version 1.4 is coded now.



Yuan Tian


Login before adding your answer.

Traffic: 697 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6