Question

Calculating standard deviation for a few genes using RNAseq data

0

Entering edit mode

marhabaie • 0

@marhabaie-16361

Last seen 4.2 years ago

Is there a way to calculate standard deviation, standard error, or 95% confidence intervals for a few genes (~10) from RNAseq data? I am comparing RNAseq fold change values with qPCR ddCts and need to show some measure of uncertainty for RNAseq FC values. Need to add error bars to this graph I was hoping to be able to get some measure of uncertainty from edgeR (that I used to get DE genes) but as far as I found edgeR uses dispersion which is not what I need.

Thank you,

edger RNAseq • 1.9k views

ADD COMMENT • link updated 4.2 years ago by Steve Lianoglou ★ 13k • written 4.2 years ago by marhabaie • 0

2

Entering edit mode

I am not an expert on this, but AFAIK it is is not possible to obtain errors or confidence intervals. See this thread. https://support.bioconductor.org/p/61640/

ADD REPLY • link 4.2 years ago Guido Hooiveld ★ 4.1k

0

Entering edit mode

This does not seem to be related to any Bioconductor package (?), even though you have tagged EdgeR. It seems that you may want geom_errorbar() from ggplot2, but, again, this is not related specifically to any Bioconductor package.

ADD REPLY • link 4.2 years ago Kevin Blighe ★ 4.0k

0

Entering edit mode

Sorry for not being clear. I just edited the question. I was hoping to be able to get it out of edgeR that I used to get DE genes. My question is not how to plot, but is what data to use to calculate errors.

ADD REPLY • link 4.2 years ago marhabaie • 0

0

Entering edit mode

After looking more carefully, I realized that ultimately I want to calculate the correlation between qPCR and RNAseq data. For this, I do NOT need error bars as they are not relevant here in calculating the correlation. Am I correct?

ADD REPLY • link 4.2 years ago marhabaie • 0

score 4 · Accepted Answer · 2020-09-15

4

Entering edit mode

Steve Lianoglou ★ 13k

@steve-lianoglou-2771

Last seen 21 months ago

United States

One way to get confidence intervals for logFC estimates from RNA-seq data is to use limma::voom.

Run the standard voom -> lmFit -> eBayes -> topTable analysis pipeline, and set the confint parameter in the topTable() function to TRUE. You'll get CI.L and CI.R columns that provide the lower and upper 95% confidence intervals you're after.

ADD COMMENT • link 4.2 years ago Steve Lianoglou ★ 13k