GENETIC MINING OF LEUKEMIA USING MICROARRAY CLASSIFIED DATA

Abstract

Acute Lymphoblastic Leukemia (ALL) is the most common cancer in children and adults. At present, diagnosis, prognosis and treatment decisions are made
based upon blood and bone marrow laboratory testing. With advances in microarray technology it is becoming more feasible to perform genetic assessment of
individual patients as well. By utilizing the information from a few microarray tests that have produced information about both the SNP paroles of patients and additionally quality expression information. By using Singular Value Decomposition (SVD) on Illumina SNP, Affymetrix and cDNA gene- expression data has performed aggressive attribute selection using random forests to reduce the number of attributes to a manageable size. Then by exploring clustering and prediction of patient-specific properties such as disease sub-classification, and especially clinical outcome. By determining that integrating multiple types
of data can provide more meaningful information than individual datasets, if combined properly. This method is able to capture the correlation between the
attributes. The most striking result is an apparent connection between genetic  background and patient mortality under existing treatment regimes.