As @Rekins suggested InSilico DB has a "merging" R-Bioconductor package to combine public datasets from GEO. If you are not using R you can also combine data from the online platform (https://insilicodb.org)

Example:

# Retrieve 2 datasets

eset1 = getDataset(gse="GSE10072", gpl="GPL96", norm="ORIGINAL", genes=TRUE);

eset2 = getDataset(gse="GSE7670", gpl="GPL96", norm="ORIGINAL", genes=TRUE);

#combine them

esets = list(eset1, eset2);

eset = merge(esets, method="NONE");

#plot them

plotMDS(eset, targetAnnot="Disease", batchAnnot="Study");

InSilico DB packaged various batch removal effects methods so line 4 could be replaced with:

eset = merge(esets, method="XPN");

or

eset = merge(esets, method="COMBAT");

Hope this helps.

For more info Bioinformatics paper reference; InSilico DB and InSIlico Merging packages links, and blog link.

- Unlocking the potential of publicly available microarray data using inSilicoDb and inSilicoMerging R/Bioconductor packages -BMC Bioinfomatics [http://www.biomedcentral.com/1471-2105/13/335/abstract]

- inSilicoDb: an R/Bioconductor package for accessing human Affymetrix expert-curated datasets from GEO - Bioinformatics [http://bioinformatics.oxfordjournals.org/content/27/22/3204]

-Tutorial example : https://insilicodb.org/the-impact-of-batch-effects-when-merging-different-data-sets/

R-Bioconductor packages:

http://www.bioconductor.org/packages/2.12/bioc/html/inSilicoDb.html

and

http://www.bioconductor.org/packages/2.12/bioc/html/inSilicoMerging.html

insilicoDB, https://insilicodb.org/ may be useful as they allow a user to combine different published datasets for free.

As I don't know how compatible it is with R, you may need to export it.

go for Bioconductor as Claus said:

http://www.bioconductor.org/packages/2.12/bioc/html/virtualArray.html

As @Rekins suggested InSilico DB has a "merging" R-Bioconductor package to combine public datasets from GEO. If you are not using R you can also combine data from the online platform (https://insilicodb.org)

Example:

# Retrieve 2 datasets

eset1 = getDataset(gse="GSE10072", gpl="GPL96", norm="ORIGINAL", genes=TRUE);

eset2 = getDataset(gse="GSE7670", gpl="GPL96", norm="ORIGINAL", genes=TRUE);

#combine them

esets = list(eset1, eset2);

eset = merge(esets, method="NONE");

#plot them

plotMDS(eset, targetAnnot="Disease", batchAnnot="Study");

InSilico DB packaged various batch removal effects methods so line 4 could be replaced with:

eset = merge(esets, method="XPN");

or

eset = merge(esets, method="COMBAT");

Hope this helps.

For more info Bioinformatics paper reference; InSilico DB and InSIlico Merging packages links, and blog link.

- Unlocking the potential of publicly available microarray data using inSilicoDb and inSilicoMerging R/Bioconductor packages -BMC Bioinfomatics [http://www.biomedcentral.com/1471-2105/13/335/abstract]

- inSilicoDb: an R/Bioconductor package for accessing human Affymetrix expert-curated datasets from GEO - Bioinformatics [http://bioinformatics.oxfordjournals.org/content/27/22/3204]

-Tutorial example : https://insilicodb.org/the-impact-of-batch-effects-when-merging-different-data-sets/

R-Bioconductor packages:

http://www.bioconductor.org/packages/2.12/bioc/html/inSilicoDb.html

and

http://www.bioconductor.org/packages/2.12/bioc/html/inSilicoMerging.html

Hi Tusar,

Keep the same annotation in both(as many as you want) datasets [ load as dataframes] and simply cbind with match. e.g

Total

@Tushar, the issue with this method is the clinical annotations associated with your datasets will be missing. Please @Reema correct me if I'm wrong. The InSilico package method proposed above will return one eset with samples' phenotypes (pData(eset)).

@Tushar, Alain is right. You should carefully annotate/check annotation for both of your dataset before merging. Many to one, one to many and many to many these issues are quite common in annotation. Use InsilicoDB or virtual array whatever you like, but pay extra attention to your dataset annotation. Here's is an old article, which is basically for meta-analysis- but I am sure few of the steps would be common in your analysis as well.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2528050/

Ciao Tushar,

I would suggest you to give a try to ComBat: 'Combatting' Batch Effects When Combining Batches of Gene Expression Microarray Data. You can also specify covariates along with the different batches. I obtained pretty good results with it. It is part of the sva package for R.

http://www.bu.edu/jlab/wp-assets/ComBat/Abstract.html

Thanks to all of you...currently I am busy with trying your suggestion and will get back to here about my experience or may be another query

This a very user-friendly and reliable tool as well the mentioned above: http://arrayanalysis.org/. Maybe for the first time the best way if you create a workflow for the array analysis, and then find the appropriate tools for the process. This can be a combined tool, mostly one is not enough :)

When you have the same arrays batches it is not that difficult. You can try frma normalization that is nice to use in clinical setting:

source("http://bioconductor.org/biocLite.R")

biocLite("affy")

library("affy")

#Loading 3 batches of affy arrays

setwd("BATCH_1_LOCATION")

batch_1

@Bartosz-THANKS A LOT. I like this fRMA normalization methods. But do you have any suggestions for combining arrays from different microarray platforms?

@Bartosz, @Tushar Please note all array gene expression data for which CEL files are available have been fRMA normalized in InSilicoDB.

In my previous example, the "norm=ORIGINAL" parameter means you are retrieving the "original" normalization as submitted by the authors in GEO. This parameter can be changed to norm="FRMA".

The getDataset method is in our InSIlicoDB package (http://www.bioconductor.org/packages/2.12/bioc/html/inSilicoDb.html)

# Retrieve 2 datasets

>eset1 = getDataset(gse="GSE10072", gpl="GPL96", norm="FRMA", genes=TRUE);

>eset2 = getDataset(gse="GSE7670", gpl="GPL96", norm="FRMA", genes=TRUE);

###The getDataset method is in our InSIlicoDB package (http://www.bioconductor.org/packages/2.12/bioc/html/inSilicoDb.html)###

To combine them, you can use the merge method from our InSilicoDBMerging package (http://www.bioconductor.org/packages/2.12/bioc/html/inSilicoMerging.html)

#combine them

>esets = list(eset1, eset2);

>eset = merge(esets, method="NONE");

The merge method allows you to apply various batch removal effect algorithms:

>eset = merge(esets, method="NONE");

>eset = merge(esets, method="COMBAT");

>eset = merge(esets, method="XPN");

>eset = merge(esets, method="DWD");

>eset = merge(esets, method="GENENORM");

Just do not do this kind of analysis. Here is why:

Even within the same array platform, the lab effect will be bigger than the biological effect. With different array platforms you will see nothing else than technical differences. Check what R. Irizarry has to say about these attempts:

http://nar.oxfordjournals.org/content/42/D1/D938.full?keytype=ref&ijkey=OFo6cWXXhRu3HXn

Sorry,

Herbert

I agree, I would be very cautious with using different array platforms,

Look at the TCGA dataset - they were using all the same platforms within different labs and batch effects somethimes are striking! You can make a batch removal - even from TCGA you have a tool provided to do that.

I would not merge different array platforms in one analysis - for example for some marker identification... You can however try to use one dataset as a validation set for another one. With this approach you escape a need to merge different platform datasets.

Hope you will find your way,

Best!

Bartek

I suggest the ComBat from sva package

 @Tushar Tomar I read the FZD-X conference abstract, did you combine GEO datasets from different platform in validation cohort? What package did you use?

combat package