BACKGROUND: Using DNA microarrays, we have developed two novel models for tumor classification and target gene prediction. First, gene expression profiles are summarized by optimally selected Self-Organizing Maps (SOMs), followed by tumor sample classification by Fuzzy C-means clustering. Then, the prediction of marker genes is accomplished by either manual feature selection (visualizing the weighted/mean SOM component plane) or automatic feature selection (by pair-wise Fisher's linear discriminant). RESULTS: The proposed models were tested on four published datasets: (1) Leukemia (2) Colon cancer (3) Brain tumors and (4) NCI cancer cell lines. The models gave class prediction with markedly reduced error rates compared to other class prediction approaches, and the importance of feature selection on microarray data analysis was also emphasized. CONCLUSIONS: Our models identify marker genes with predictive potential, often better than other available methods in the literature. The models are potentially useful for medical diagnostics and may reveal some insights into cancer classification. Additionally, we illustrated two limitations in tumor classification from microarray data related to the biology underlying the data, in terms of (1) the class size of data, and (2) the internal structure of classes. These limitations are not specific for the classification models used.

MOTIVATION: Radiation Hybrid Mapping (RHM) is a technique used to order a set of markers on a genome and estimating physical distances between them. RHM provides information on marker placement independent from other methods such as sequencing, and can therefore be used for example in genome sequencing to help ordering contigs. A radiation hybrid framework can be constructed by choosing a set of markers so that the chromosome coverage is good and so that the markers can be ordered with high confidence. Automatically constructing RHM frameworks is a computationally challenging problem. RESULTS: We have developed a new method for constructing radiation hybrid frameworks. Given a relatively large set of markers for a chromosome, the algorithm aims to select an ordered subset that makes up a framework, and that contains as many markers as possible. The algorithm has a time complexity that is better than any of the existing methods that we are aware of. Furthermore, we propose a method for comparing if two frameworks are consistent, giving a visual presentation as well as quantitative measures of how well the two frameworks agree. Applying our method on marker sets from 22 human chromosomes and comparing the resulting frameworks with previously published frameworks, we demonstrate that our automatic method efficiently constructs frameworks with good coverage of each chromosome and with high degree of agreement on the marker ordering.