This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Gene filtering returns NULL dimensions

Hi, I'm using R and bioconductor in order to conduct some gene analysis but I'm having trouble at filtering genes. I'm trying to create a dataset called small.eset by filtering my dataset but small.eset returns NULL dimensions. I was wondering if I'm doing some obvious mistake?

The dataset I'm using is #redacted and I'm sharing my condensed code (without plots and quality checks).

# Library and Data Imports ------------------------------------------------
library(affy)
library(affyPLM)
library(simpleaffy)
library(limma)
library(RColorBrewer)
library(genefilter) 
setwd("/Users/Documents/R Scripts/biostat_geo/data")
Data <- ReadAffy() #read all CEL files in working directory

# log2 transformation -----------------------------------------------------

ex <- exprs(Data)
qx <- as.numeric(quantile(ex, c(0., 0.25, 0.5, 0.75, 0.99, 1.0), na.rm=T))
LogC <- (qx[5] > 100) ||
  (qx[6]-qx[1] > 50 && qx[2] > 0)
if (LogC) { ex[which(ex <= 0)] <- NaN
exprs(Data) <- log2(ex) }

#background correction

eset <- expresso(Data, bgcorrect.method = "mas",
                 normalize.method = "loess", 
                 pmcorrect.method = "pmonly",
                 summary.method = "mas")

# filtering

esetunlog <- 2^exprs(eset)
ffun <- filterfun(pOverA(0.20,10))
t.fil <- genefilter(esetunlog,ffun)
small.eset <- (log2(esetunlog[t.fil,]))
r filtering affymetrix genefilter cel
• 1,449 views • link
written by purestdata • purestdata

Try debugging each filter step by running head for each result variable.

head(esetunlog)
head(ffun)
head(t.fil)
• 0 views • link
written by Ram Ram 
ffun <- filterfun(pOverA(p=0.20,A=100))

If I use 5 instead of 100 for A argument in pOverA;

ffun <- filterfun(pOverA(p=0.20,A=5))

it starts to populate numbers in the small.eset.

I'm reading the help page for pOverA and it states A is the number to be exceeded for gene selection. What is the proper way of defining this number?

In addition to that, it seems like I am able to produce a small dataset if I use nsFilter(). Are these two functions interchangeable?

Thanks for your time in advance!

• 0 views • link
written by purestdata • purestdata

I'm not familiar with the package you're using. I was merely guiding you on the debugging process. Someone familiar with the package might be able to help you with the specifics.

• 0 views • link
written by Ram Ram

0 answers

No answers yet.

Log in to answer this question.

More posts like this