Hii I am handling microarray data and I have two datasets with the same platform (GPL750) one of which contains tumor samples and another one has normal samples. I want to merge these two datasets for differential expression analysis. Is it required batch correction?
If the data is from two different experiments, you should correct for batch effect indeed. It sounds however, that you have an impossible design: tumors in one batch and normal in the other batch. It is not possible to correct for batch in this case, since there is no overlap between the batches. You need tumors and/or normals in both batches.