There are two ways to combine datasets:
If you are familiar with R you can use our R InSilico Merging package, this package has the advantage of wrapping batch effect removal methods like XPN and COMBAT.
From the web interface you can combine samples you pick from different datasets
For both methods raw data .CEL files must be available. InSIlico DB starts from the CEL files which are preprocessed using fRMA and SCAN algorithms, the results are then combined and optionally, if using the R package, passed to batch effect removal methods.
From the interface (limited to one platform at the time e.g HGU-133-plus_2):
- Click on a dataset of interest, scroll down to the clinical annotations, and use the tree view to select the samples you want by right-clicking the term of interest and click on the "add to basket" option. If the tree does not show these options .CEL files are not available.
Go to the next study and repeat the operation.
Go to the sample basket and edit the clinical annotations to standardize them.
Save your new dataset and you are ready to go!
Note in this example I'm choosing lymph node status=Neg and editing ln -> lymph node status