Additional important differences between the two studies relate to the number, origin and sampling of tissue, as well as medical information available. Sample-related differences should be of uttermost importance, although this is quite difficult to prove as no direct comparison was possible. The Stuttgart cohort consisted exclusively of liver tissue removed surgically from Caucasian donors in one hospital, using only one procedure for sample collection, freezing, storage, RNA isolation, quality assessment, DNA isolation and microarray analysis. Although all but one sample were resected because of liver cancer, this fact by itself should not affect genotype–phenotype relationships because only non-tumorous material was analyzed. In contrast the 427 samples of the Seattle cohort consisted mostly of postmortem material obtained from prospective organ donors who were presumably cancer-free, but the tissue quality may vary more widely because of warm ischemia before preservation and long storage times before cryopreservation. Furthermore, the Seattle cohort was collected in three independent centers, giving rise to differences among samples regarding tissue acquisition and storage protocols, criteria for RNA quality, etc. Another significant influential factor in the Seattle study
