The absence of reliable, patient-specific models in cancer biology has been a chronic impediment to understanding heterogeneity, progression, metastasis, and chemotherapy resistance. Recent advances in technologies for cloning adult stem cells resident in normal, regenerative epithelia have been adapted to cloning stem cells from Barrett's esophagus. Investigators in the proposed Center have now extended this technology, in preliminary studies, to cloning the stem cells of dysplastic Barrett's and adenocarcinoma itself. The impact of this technology for resolving key questions in Barrett's and cancer biology in general, including the precise steps in tumorigenesis from precursor lesions, the functional and genomic intra-tumor heterogeneity and its impact on chemotherapy resistance, and lastly the nature and targetability of stem cells of precursor lesions and frank adenocarcinoma, is potentially transformative. As potentially transformative is a biorepository that exploits the regenerative and clonal nature of these stem cells to extend our knowledge of discrete cases in parallel fashion. In Aim 1, we will generate high-density arrays of stem cell clones corresponding to a topological sampling across four mucosal resections of patients with early adenocarcinoma. Aim 2 will generate large libraries of cancer stem cells from advanced cases of esophageal adenocarcinoma. Aim 3 will generate an interactive database that links datasets from both the arrays of clones derived from mucosal resections as well as those associated with analyses of the cancer stem cell libraries. We anticipate that such a virtual database of datasets would greatly augment downstream studies with the same libraries and discrete clones as each would come with such linked datasets that would be further extended by ongoing and future studies by investigators within the Center and across the community.