Analysis of variance, normal quantile-quantile correlation and effective expression support of pooled expression ratio of reference genes for defining expression stability

详细信息    查看全文

• 作者：Himanshu Priyadarshi^a ; ^{priyadarshimanshu@gmail.com} ; Rekha Das^b ; Shivendra Kumar^c ; Pankaj Kishore^d ; Sujit Kumar^e
• 关键词：Bioinformatics ; Mathematical biosciences
• 刊名：Heliyon
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：3
• 期：1
• 全文大小：607 K
• 卷排序：3

文摘

Identification of a reference gene unaffected by the experimental conditions is obligatory for accurate measurement of gene expression through relative quantification. Most existing methods directly analyze variability in crossing point (Cp) values of reference genes and fail to account for template-independent factors that affect Cp values in their estimates. We describe the use of three simple statistical methods namely analysis of variance (ANOVA), normal quantile-quantile correlation (NQQC) and effective expression support (EES), on pooled expression ratios of reference genes in a panel to overcome this issue. The pooling of expression ratios across the genes in the panel nullify the sample specific effects uniformly affecting all genes that are falsely reflected as instability. Our methods also offer the flexibility to include sample specific PCR efficiencies in estimations, when available, for improved accuracy. Additionally, we describe a correction factor from the ANOVA method to correct the relative fold change of a target gene if no truly stable reference gene could be found in the analyzed panel. The analysis is described on a synthetic data set to simplify the explanation of the statistical treatment of data.