We propose to become members of the consortium to develop mouse models for cancer. We propose to contribute a number of excellent models for human gastrointestinal cancer that we have developed. In addition, we propose to develop and examine a number of new models for GI and other cancers. The Albert Einstein College of Medicine is designated as a Comprehensive Cancer Center by the National Cancer Institute and provides a number of core facilities for generating transgenic and gene knock-out mice. The barrier facility where the mice will be housed allows maintenance and breeding of mice in a pathogen-free facility, making it possible to easily ship mice to other facilities. Many of the other core facilities of the Cancer Center, such as the Oligonucleotide and DNA Sequencing Facility, Image Analysis Facility, Protein Sequencing and Synthesis Facility, Cell Sorting Facility and Monoclonal Antibody Facility, are important assets for the proposed work. We have also developed and/or implemented several new technologies that would be important for the proposed activity. These include high throughput bacterial clone screening, preparation of bacterial clone maps of regions of the genome and use of the Affymetrix oligonucleotide arrays for detection of changes in gene expression. We have also built a robot to prepare high density microarrays of cDNA inserts from validated mouse and human cDNA clones, methods to hybridize them with complex probes and laser scanning devises to read, record and analyze the data. The infrastructure also includes a very robust histopathology with particular expertise on gastrointestinal cancer. Several genes whose germ line mutations in humans lead to a cancer predisposition phenotype and other genes which are consistently found to be mutated in colorectal tumors have been identified. During the past several years, we generated mice with mutations in 14 different genes implicated in human colorectal cancer. These include Apc, Mcc, Ki-ras, N-ras, Smad2, Smad4, Msh2, Msh3, Msh4, Msh5, Msh6, Mlh1. Many of these mice with single gene mutations or in combinations show a high degree of tumor suspectibility. We now propose to validate some of these models, test additional combinations and create new lines of genetically modified mice to serve as GI tumor models. Since many of these mice develop tumors of other tissue types as well, it is possible that combining our gene modifications with those developed by other members of the consortium may provide for novel models for other human tumors.