Question

Error in the fsva function after setting "newdat" argument to NULL

2

Entering edit mode

zicheng.hu ▴ 20

@zichenghu-12689

Last seen 6.8 years ago

I ran the example code from the sva page :

bioconductor.org/packages/release/bioc/vignettes/sva/inst/doc/sva.R

Everything works great. However, I got an error after I change the code in chunk 24 from

fsvaobj = fsva(trainData,trainMod,trainSv,testData)

to

fsvaobj = fsva(trainData,trainMod,trainSv,newdat=NULL)

Error message: Error in fsva(trainData, trainMod, trainSv, newdat = NULL) : object 'newV' not found

Why am I not allowed to set "newdat" argument to NULL here? I want to clean the data using sva for clustering and PCA analysis, so I don't have test data. Can I only adjust the training data using fsva?

sva • 1.0k views

ADD COMMENT • link updated 7.4 years ago by victor • 0 • written 7.6 years ago by zicheng.hu ▴ 20

score 0 · Answer 1 · 2017-08-22

0

Entering edit mode

victor • 0

@victor2wy-12605

Last seen 6 months ago

Canada

Did you ever figure this out?

I'm running into the same error.

ADD COMMENT • link 7.4 years ago victor • 0

0

Entering edit mode

I ended up using the removeBatchEffect function from limma.

SV = sva(expr,mod,mod0)
expr = removeBatchEffect(x=expr, batch=NULL, batch2=NULL, covariates=SV$sv)

ADD REPLY • link 7.3 years ago zicheng.hu ▴ 20