Microarray Meta Analysis

Question

[kalyanimeha]

How to perform the meta-analysis of 15 GEO dataset in R. Is there any good and simple package which can be understood easily by me.

Answer 1

[h.mon]

Are you a first-grade student? Graduate student? Master, PhD, post-doctoral researcher? I will assume you are at PhD / post-doctoral level:

There are several methods and papers (for example, Meta-analysis methods for combining multiple expression profiles: comparisons, statistical characterization and an application guideline and Comparison study of microarray meta-analysis methods).

Are all the 15 GEO sets from the same microarray platform, or different platforms? Probably the most challenging part is to annotate all platforms to a common denominator.

There are indeed several R/BioConductor packages, for example, GeneMeta, crossmeta or RankProd. As you don't give details about the datasets you want to analyse, it is up to you to explore these packages and find out which one is more fit to your needs. There is also at least one server to perform meta-analyses of gene-expression (NetworkAnalyst for statistical, visual and network-based meta-analysis of gene expression data).

[kalyanimeha]

Helpful. I have different GPL file of GEO dataset. I am an M.Sc student. I am not getting how to start metanalysis with the help of R package. Data Set=

GSE35503-Agilent

GSE35503-Affymetrix Human Genome

GSE16186-Sentrix HumanRef-8 v2 Expression BeadChip

GSE88947-Agilent-021529

GSE86885-Illumina HumanHT-12

GSE32474-Affymetrix Human Genome

GSE90274-Illumina HiSeq 2000 (Homo sapiens)

GSE44186=Affymetrix Human Genome

GSE33755-Agilent-014850 Whole Human Genome Microarray 4x44K

GSE28974-Illumina human-6 v2.0 expression beadchip

GSE36970-Affymetrix Human Genome U133

GSE27313-Affymetrix Human Genome U133 Plus

GSE19664-Affymetrix Human Genome U133 Plus

I am new to meta-analysis, I will be grateful if you helped me. Thanx

[h.mon]

It seems crossmeta is a good choice then. Look at its vignette, and try to follow it - I suggest at first with its example, then start with two or three of the arrays of your interest. Only after succeeding with these steps, perform the whole analysis.

Take notes all the way, and when posting questions about the problems you encounter, post the code and the error message.