Lung Cancer - New Information from Old Data: Preliminary Results

In Biology and Medicine, it is possible to generate large amounts of experimental data at a high speed rate. For example, microarrays can provide large amounts of data for genetic relative expression in illnesses of interest such as cancer in short time. These data, however, are stored and often times abandoned when new experimental technologies arrive. This work re-examines lung cancer microarray data with a multiple criteria optimization-based strategy developed in our research group. This strategy does not require any adjustment of parameters by the user and is capable to converge consistently to important genes -potential biomarkers- even in the presence of multiple and incommensurate units across microarrays. Groups with distinct smoking habits (never smoker, current smoker) and gender are contrasted to elicit a set of highly differentially expressed genes, several of which are already associated to lung cancer and other types of cancer. The list of genes is provided with a discussion of their role in cancer, as well as the possible research directions for each of them. It is also recognized at this point that experimental validation is necessary to confirm the role of genes for which not enough evidence is found in the literature. Fundamentally, these genes with little information represent the best opportunities for biological discovery from existing data.