Tumorigenesis is thought to be a multistep process in which genetic alterations accumulate, ultimately producing the neoplastic phenotype. A model was proposed to explain the genetic basis of colorectal neoplasia that included several salient features. First, colorectal tumors appear to occur as a result of the mutational activation of oncogenes coupled with the inactivation of tumor-suppressor genes. Second, mutations in at least four or five genes are required to produce a malignant tumor. Third, although the genetic alterations often occur in a preferred sequence, the total accumulation of changes, rather than their chronologic order of appearance, is responsible for determining the tumor's biologic properties. Several different genetic alterations were identified that occur during colorectal tumorigenesis. Activational mutation of the ras oncogene was found in approximately 50% of colonic carcinomas and in a similar percentage of intermediate-stage and late-stage adenomas. Allelic deletions were discovered of specific portions of chromosomes 5, 17, and 18, which presumably harbor tumor-suppressor genes. The target of allelic loss events on chromosome 17 has been shown to be the p53 gene, which is mutated, not only in colonic cancer, but also in a large percentage of other human solid tumors. The gene dcc recently was identified; this candidate tumor-suppressor gene on chromosome 18 appears to be altered in colorectal carcinomas. The protein encoded by the dcc gene has significant sequence similarity to neural cell adhesion molecules and other related cell-surface glycoproteins. By mediating cell-cell and cell-substrate interactions, this class of molecules may have important functions in mediating cell growth and differentiation. Alterations of the dcc gene may interfere with maintenance of these controls and thus may play a role in the pathogenesis of colorectal neoplasia. Another candidate tumor-suppressor gene also was identified on chromosome 5, mcc (for mutated in colorectal cancers). The mcc genetic alterations include one tumor with somatic rearrangement of one mcc allele and several tumors with somatically acquired point mutations in the coding region. Studies currently are ongoing to (1) identify additional tumor-suppressor gene candidates, (2) increase our understanding of normal tumor-suppressor gene function, and (3) demonstrate the functional tumor-suppressor ability of these genes both in vivo and in vitro.