Average Rank-Based Score to Measure Deregulation of Molecular Pathway Gene Sets

BackgroundDeregulation of biological pathways has been shown to be involved in the turmorigenesis of a variety of cancers. The co-regulation of pathways in tumor and normal tissues has not been studied in a systematic manner.
ResultsIn this study we propose a novel statistic named AR-score (average rank based score) to measure pathway activities based on microarray gene expression profiles. We calculate and compare the AR-scores of pathways in microarray datasets containing expression profiles for a wide range of cancer types as well as the corresponding normal tissues. We find that many pathways undergo significant activity changes in tumors with respect to normal tissues. AR-scores for a small subset of pathways are capable of distinguishing tumor from normal tissues or classifying tumor subtypes. In normal tissues many pathways are highly correlated in their activities, whereas their correlations reduce significantly in tumors and cancer cell lines. The co-expression of genes in the same pathways was also significantly perturbed in tumors.
ConclusionsThe co-regulation of genes in the same pathways and co-regulation of different pathways are significantly perturbed in tumors versus normal tissues. Our method provides a useful tool for better understanding the mechanistic changes in tumors, which can also be used for exploring other biological problems.

背景：生物通路失调已被证实参与多种癌症的肿瘤发生过程。目前，针对肿瘤与正常组织中通路的共调控，尚未开展系统性研究。
结果：本研究提出一种名为AR-score（average rank based score，基于平均秩的评分）的新型统计量，用于基于微阵列基因表达谱（microarray gene expression profiles）评估通路活性。我们针对涵盖多种癌症类型及对应正常组织表达谱的微阵列数据集，计算并比较了各通路的AR-score。研究发现，相较于正常组织，众多通路在肿瘤组织中出现了显著的活性变化。仅需少数通路的AR-score，即可有效区分肿瘤与正常组织，或对肿瘤亚型进行分类。在正常组织中，多数通路的活性呈现高度相关性；而在肿瘤及癌细胞系中，这种相关性显著降低。同一通路内基因的共表达情况在肿瘤组织中也发生了显著扰动。
结论：相较于正常组织，肿瘤组织中同一通路内基因的共调控以及不同通路间的共调控均发生了显著扰动。本研究提出的方法为深入理解肿瘤的机制变化提供了实用工具，同时也可用于探索其他生物学问题。