BSAIC: Identification of Significant Copy Number Aberrations in Cancer

Because it helps to determine the location of driver genes and cancer suppressor genes, the identification of the significant aberration regions from the somatic copy numbers in cancer is vital to make clear the mechanisms of tumor initiation and progression. To address the problem of identifying Significant Copy Number Aberrations in cancer, in this paper, we have completed the following four tasks. (1) We have re-implemented GISTIC, JISTIC and SAIC in a Java project and published it as an open source project. (2) We have proposed an improved algorithm BSAIC based on both the significant detection algorithm SAIC and the famous parallel integrated learning method Bagging. (3) We have developed a mechanism for generating synthetic datasets aimed at such algorithms. (4) We have compared the performances of GISTIC, JISTIC, SAIC and BSAIC using the synthetic datasets, and proved the advantages of the BSAIC.