limma, best practices review paper?

Question

[RNAseqer]

Is Limma still one of the better packages for DE analysis of microarray data or has it been surpassed by other programs?

Related to that question, could anyone recommend a recent review article detailing best-practices and/or comparing program accuracy/sensitivity for microarray DE analysis?

Thanks!

Answer 1

[Kevin Blighe]

I cannot define the word "better", but Limma is to microarray analyses what PLINK is to GWAS association testing. It became the leading microarray differential expression analysis tool long ago and has much general functionality for microarray data processing, too. The package is downloaded >10,000 times per month from the Bioconductor repository.

One issue with microarray data processing is the diversity of microarray architectures, which translates into the fact that an analysis pipeline can be very different from one platform to another. Oligo and affy are used to process Affymetrix arrays, while limma can be used for Agilent arrays, as a general guide. However, limma can be used to conduct the diff. expression analysis for all microarrays. Other packages, like aroma.affymetrix, are used for processing Affymetrix copy number arrays.

As a result of the above, there is no main 'all-encompassing' pipeline. Code is scattered all across the WWW to process microarrays.

One comprehensive tutorial includes An end to end workflow for differential gene expression using Affymetrix microarrays

A good review on microarrays, generally, is by John Quackenbush: Microarray data normalization and transformation.

Kevin

[RNAseqer]

As always, I really appreciate the advice. Thanks for bringing some clarity to this for me.