AUTHOR=Chang Zhiqiang , Miao Xiuxiu , Zhao Wenyuan 

TITLE=Identification of Prognostic Dosage-Sensitive Genes in Colorectal Cancer Based on Multi-Omics

JOURNAL=Frontiers in Genetics

VOLUME=Volume 10 - 2019

YEAR=2020

URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2019.01310

DOI=10.3389/fgene.2019.01310

ISSN=1664-8021

ABSTRACT=Several studies have already identified the prognostic markers in colorectal cancer (CRC) based on somatic copy number alteration (SCNA). However, we still know a little about why they could be the prognostic marker. Gene dosage effect is one important mechanism of Copy Number, and dosage sensitive genes are more likely to be the driver genes. Thus, our work proposes a new pipeline to identify the prognostic genes which is dosage-sensitive as well. 
The data of RNAseq, somatic copy number of CRC from TCGA were assayed to scree out the SCNAs. After identifying the differential expressed genes with wilcoxon rank-sum test in alteration samples with |SCNA|>0.3, cox-regression was used to finding the candidate prognostic genes. Then a iterative algorithms was built to identify the stable prognostic genes with log-Rank test in alteration samples and non-alteration samples with different cutoff (from 0.1 to 0.5, step by 0.2) of SCNA. Finally, pearson correlation coefficient was calculated between gene expression and SCNA as the dosage effect score. The cell line data from CCLE was to test the consistence of the dosage effect. Differential co-expression network was built to discover their function in CRC
As a result, 6 amplified genes (NDUFB4, WDR5B, IQCB1, KPNA1, GTF2E1 and SEC22A) were found to be associated with pool prognosis, and they show a stable prognostic classification in more than 50% cutoff of SCNA. The average of dosage effect score was 0.5918±0.066, 0.5978±0.082 in TCGA and CCLE respectively, and they showed great stability in different data-set. In the network of differential co-expression, these 6 genes have the top degree. Function enrichment analysis revealed that gene NDUFB4 and GTF2E1 affect cancer-related functions such as transmembrane transport and transformation factors.
 In conclude, the pipeline for identifying the prognostic dosage-sensitive genes in CRC was proved to be stable and reliable.