Hello, I am working on two Agilent-039494 SurePrint G3 Human GE v2 8x60K Microarray datasets. One dataset only includes metastatic samples to the A site, and the other includes metastatic samples to the B site only. I merged this dataset to find differentially expressed genes. I plot PCA, to check batch effect, and I added PCA graph below. My two biological group coming from two different dataset. How can ı remove batch effect?

If all of the samples in group 155 are one type, and all the other samples in 438 are the other, then no. The biological and technical variability are completely confounded.