Support Vector Machine Approach for Cancer Detection using Amplified Fragment Length Polymorphism (AFLP) Screening Method

Waiming KONG1, Lawrence THAM2, Kee Yew Wong, Patrick Tan, Keng Wah CHOO, Nanyang Polytechnic;, Nanyang Polytechnic

Support Vector Machine is used to cluster data obtained from Amplified Fragment length Polymorphism screening of gastric cancer and normal tissue samples. AFLP was performed on genomic DNA samples of 58 gastric tumors which were obtained from cancer patients and 16 normal genomic DNA samples which were obtained from peripheral blood samples from healthy volunteers. Using the electrophoresis peak intensity measurements of the amplified fragments of the cancer and normal tissues, SVM was able to distinguish gastric cancer from normal tissue samples with a sensitivity of 0.98 and selectivity of 0.75. As AFLP is a low cost procedure which requires minimum prior sequence knowledge and biological material, SVM prediction of AFLP screening data is a potential tool for gastric cancer screening and diagnosis.