Question

lumiR [Illumina] error from non-normalized data

0

Entering edit mode

Stane ▴ 40

@stane-10974

Last seen 6.3 years ago

Hello, I am trying to analyse two illumina GEO GSE38410 & GSE59522 non-normalized raw data file with bioconductor package 'lumiR'.

http://www.ncbi.nlm.nih.gov/geo/download/?acc=GSE38410&format=file&file=GSE38410%5Fnon%2Dnormalized%2Etxt%2Egz

http://www.ncbi.nlm.nih.gov/geo/download/?acc=GSE59522&format=file&file=GSE59522%5Fnon%5Fnormalized%2Etxt%2Egz

I am getting error 'subscript out of bounds' so I am guessing that those non-normalized files format may be off and would like another point of view as I am quite new to illumina.

Thanks,

illumina • 1.6k views

ADD COMMENT • link updated 7.7 years ago by Gordon Smyth 50k • written 7.7 years ago by Stane ▴ 40

1

Entering edit mode

I prefer using limma to normalise illumina array data, it can be done simply with the command neqc(yourdata) after reading the exported .txt files with read.ilmn() function. Sorry, that may not help, but it is probably something to consider.

ADD REPLY • link 7.7 years ago chris86 ▴ 420

0

Entering edit mode

Thank you for this other solution, I am going to check that. update: your method work with my other illumina files, but not for those two, I figure that they are missing all the illu columns and one even seems to be already log2 transform expression value :)

ADD REPLY • link 7.7 years ago Stane ▴ 40

score 1 · Answer 1 · 2016-08-23

You cannot use any specialist methods for Illumina arrays on these files because they are just matrices of numbers without all the associated information that is found in a raw Illumina output file. You could read the normalized data that is provided on GEO website. Or else read the non-normalised matrix into R using read.delim(), then apply normalizeQuantiles() and possibly log2() as appropriate to the matrix of numbers. normalizeQuantiles() is a function in the limma package.