Detecting robust gene signature through integrated analysis of multiple types of high-throughput data in liver cancer

To investigate the robust gene signature in liver cancer, we applied an integrated approach to perform a joint analysis of a highly diverse collection of liver cancer genome-wide datasets, including genomic alterations and transcription profiles.

1-class Significance Analysis of Microarrays coupled with ranking score method were used to identify the robust gene signature in liver tumor tissue.

In total, 1,625,051 gene expression measurements from 16 public microarrays, 2 pairs of serial analyses of gene expression experiments, and 252 loss of heterozygosity reports obtained from 568 publications were used in this integrated study. The resulting robust gene signatures included 90 genes, which may be of great importance to liver cancer research. A system assessment analysis revealed that our integrative method had an accuracy of 92% and a correlation coefficient value of 0.88.

The system assessment results indicated that our method had the ability of integrating the datasets from various types of sources, and eliciting more accurate results, as can be very useful in the study of liver cancer.