I am working on minimum 10 different datasets of microarray, which combines of Affymetrix, Agilent, illumina etc. All data is pre-processed and log2 transformed. I would like to perform the analysis combining all data together. I applied limma batch-effect removal method, however, this method didn't really work. As far as I understand I can't use combat on negative values. So I was wondering if any one could suggest a approach to remove batch correction.
Thank you in advance