Computer aided analysis of Restriction Landmark Genomic Scanning images from tumor and cell line models.

摘要

Restriction Landmark Genomic Scanning (RLGS) was used in this dissertation to detect genetic changes such as loss of heterozygosity, amplification, and deletion, in a variety of tumor and cell line models. New image processing and analysis software packages and methods were also used to extract useful data from the complex RLGS spot patterns. The first study employed a novel color hybridization Macro to detect genetic alterations in Methylnitrosourea (MNU) induced rat mammary tumors that were also treated with a punitive chemopreventive agent 4-Hydroxyphenylretinamide (4HPR). Results indicated that the MNU group had 35 altered spots compared to the untreated group and the NMU-4HPR contained fewer genetic alterations when compared to the control group. In the cervical cancer project, the group employed the same color hybridization macro in concert with an image warping system to detect alterations between different types premalignant, cancerous, and normal tissue. Analysis indicated 20 alterations between the premalignant and cancerous groups and the cervical normal tissue control. The third study was responsible for mapping the RLGS alterations between K-ras transformed NIH3-3T3 and NIH-3T3 control cell lines. This study utilized a new software package called GELLEB II+ to perform the image processing and analysis of the RLGS spot patterns. The software package and a relational database program, called COG, was extremely effective in the Lapacian segmentation, spot matching, and subsequent data mining of the RLGS gels. Results showed 12 alterations between the experimental and control groups. Two spots that were amplified in the K-ras transformed NIH3-3T3 group were statistically significant (p 0.01) With this in mind, these alterations may represent key events in the transformation process of K-ras. In the human bronchial epithelial High-LET Alpha Particle study, RLGS and the GELLAB II+ system in concert with Java based Webgel showed a total of 13 alterations between the non-High-LET Alpha Particle group in the irradiated samples.; RLGS was shown to be an effective tool in delineating genetic alterations in a variety of models when used with specialized image processing and analysis software packages.