Title :
Genome-wide germline single nucleotide polymorphisms for cancer classification in the Framingham Heart Study
Author :
Peterson, Leif ; Paranilam, Jaya ; Xu, Jiaqiong ; Monzon, Federico
Author_Institution :
Center for Biostat., Methodist Hosp. Res. Inst., Houston, TX, USA
Abstract :
We identified 90 germline single nucleotide polymorphisms (SNPs) that were informative for discriminative analysis of 9 major cancers among genotyped Framingham Heart Study participants. Support vector machines resulted in the greatest classification performance, which was in the range of 70-100%. The germline SNPs identified are based on DNA from peripheral blood lymphocytes obtained during non-invasive blood draws, and unlike SNPs in tumor DNA, may not be functionally related to tumor characteristics. Further validation studies are required in order to understand the role of the seeding, genetic selection, and lifetime cumulative effects of these germline SNPs in cancer development.
Keywords :
DNA; biology computing; blood; cancer; genomics; medical computing; support vector machines; tumours; DNA; Framingham heart study; cancer classification; discriminative analysis; genome-wide germline single nucleotide polymorphisms; noninvasive blood draws; peripheral blood lymphocytes; support vector machines; tumor characteristics; Bioinformatics; Biological cells; Cancer; DNA; Genomics; Kernel; Support vector machines;
Conference_Titel :
Neural Networks (IJCNN), The 2010 International Joint Conference on
Conference_Location :
Barcelona
Print_ISBN :
978-1-4244-6916-1
DOI :
10.1109/IJCNN.2010.5596665
