rowVars function, used in ballgown analysis not available for R 3.5.2, alternatives?
2
1
Entering edit mode
2.6 years ago
RNAseqer ▴ 160

Hello,

I've hit a bit of a snag here: I had to update to the latest version of R to get ballgown working, but one of the functions I see included in filtering out low expression genes, rowVars, is not available for this version of R.

Using this tutorial: https://rpubs.com/kapeelc12/Ballgown

The rowVars function is employed in filtering:

bg_filt = subset(bg,"rowVars(texpr(bg)) >1",genomesubset=TRUE)


I have R studio, would it be possible to download an older version of R and switch to it temporarily to use rowVars, then switch back to continue the ballgown output analysis?

If so does anyone know what the last version of R rowVars worked on?

Or should I try to find an older version of ballgown, and downgrade my R version? I know there are many functions on bioconductor that do not necessarily work on the latest version of R, so is there a benifit to running older versions of R generally when working with these packages?

R rna-seq ballgown rowVars • 2.6k views
5
Entering edit mode
2.6 years ago

Is it the rowVars function from metaMA ?

install.packages('metaMA')
library(metaMA)
?rowVars

0
Entering edit mode

Huh, must be! I tried to get it as part of matrixstats but that didnt work, this download did. Thanks a bunch! This was a real help.

0
Entering edit mode

This comment has been moved to an answer so it can get accepted and mark this question as solved.

4
Entering edit mode
2.6 years ago

I would not recommend using ballgown in the first place - it performs VERY poorly compared to all the other DE tools out there (and have not been updated in years). Take a look at this recent benchmark (which agrees with most other benchmarks but is one of the few that actually test ballgown). Instead go for DESeq2 or edgeR - if you have StringTie data you can directly get them into R via tximport and use use DESeq2 as described here.

1
Entering edit mode

This is going to save me some time and heartache. Thanks for the heads-up!