Question

error in filtering genes with rowVars(gexpr(bg_chrX)) >1

0

Entering edit mode

fubeide • 0

@fubeide-13688

Last seen 7.3 years ago

Hi,

I'm doing RNA-Seq data analysis following nature protocol paper (Transcript-level expression analysis of RNA-seq experiments with HISAT, StringTie and Ballgown). But after going to step 10 (bg_liver_filt=subset(bg_liver_filt,"rowVars(gexpr(bg_liver_filt))>1",genomesubset=TRUE)), I found some genes in my results with expression of 0 in all samples. just like this:

Ar.10265 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
Ar.10715 0.180575 0.000000 0.160167 0.000000 0.000000 0.000000
Ar.10726 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
Ar.10727 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

(got by liver_gene_expr=gexpr(bg_liver_filt))

So, I think their must be something wrong with subset .........

Thanks for any help.

Beide

genefilter ballgown • 2.0k views

ADD COMMENT • link updated 6.9 years ago by importantrubbish • 0 • written 7.3 years ago by fubeide • 0

score 0 · Answer 1 · 2017-08-08

0

Entering edit mode

James W. MacDonald 67k

@james-w-macdonald-5106

Last seen 1 day ago

United States

subset is a function from base R that has been around since the dark ages, and used by millions daily. The likelihood that there is a bug in that function is infinintesimal compared to the likelihood that you made a mistake in your code, so I would go back and re-check your code first.

ADD COMMENT • link 7.3 years ago James W. MacDonald 67k

0

Entering edit mode

Thank you for your reply. But I think I didn't give enough info about the function subset. The function subset I used is not from base R, but from package ballgown..................

I go back and checked my code and found that after using the filtering criteria (rowVars(gexpr(bg_liver_filt))>1), I got the logic variable of each gene in my ballgown object ,just like:

Ar.10265 Ar.10266

FALSE FALSE

However, these genes were still kept in the filtered ballgown object. I don't know why it is so wired..........

ADD REPLY • link 7.3 years ago fubeide • 0

score 0 · Answer 2 · 2018-01-17

0

Entering edit mode

importantrubbish • 0

@importantrubbish-14803

Last seen 6.9 years ago

Hi, texpris better then gexpr I think. The ballgown object contains the information of introns, exons and transcripts , but no genes there. When filtering data with genes, the row names of the results may not match the right ones.

ADD COMMENT • link 6.9 years ago importantrubbish • 0