search

MarioniLab / scran

Clone of the Bioconductor repository for the scran package.
https://bioconductor.org/packages/devel/bioc/html/scran.html
40 stars 22 forks source link

Ask about trendVar step #21

Closed gxg2002cn closed 5 years ago

gxg2002cn commented 5 years ago

Hi again, I met an error when running trendVar in identifying highly variable genes as below. "Error in model.frame.default(formula = to.fit ~ kept.means, weights = c(1_EpiAPST2 = 59L, : variable lengths differ (found for 'kept.means')" If set use.spikes=FALSE then no error. My questions are:

  1. How to fix the error?

  2. If it is a big differece use or don´t use spick-in here? Thank you very much!

LTLA commented 5 years ago

If this is the same data set as #18, then you can't use spike-ins because they're all NaN. You have no choice but to fit the curve to endogenous genes - see the comments at http://bioconductor.org/packages/devel/workflows/vignettes/simpleSingleCell/inst/doc/xtra-3-var.html#32_when_spike-ins_are_unavailable.

gxg2002cn commented 5 years ago

Yes, it´s the same data as #18. I will remove the ERCC. Thanks a lot for your help! Best regards, Xiaogang

On Sun, 14 Oct 2018 at 13:36, Aaron Lun notifications@github.com wrote:

If this is the same data set as #18 https://github.com/MarioniLab/scran/issues/18, then you can't use spike-ins because they're all NaN. You have no choice but to fit the curve to endogenous genes - see the comments at http://bioconductor.org/packages/devel/workflows/vignettes/simpleSingleCell/inst/doc/xtra-3-var.html#32_when_spike-ins_are_unavailable .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/MarioniLab/scran/issues/21#issuecomment-429619168, or mute the thread https://github.com/notifications/unsubscribe-auth/Aju_Emb7AKVw_BUBkkeym6LkPFl4C7Dwks5ukyGugaJpZM4XasfH .