Hsiao Tzu-Hung, Chen Hung-I Harry, Lu Jo-Yang, Lin Pei-Ying, Keller Charles, Comerford Sarah, Tomlinson Gail E, Chen Yidong
Greehey Children's Cancer Research Institute, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA.
Transl Cancer Res. 2013 Feb 1;2(1):6-17. doi: 10.3978/j.issn.2218-676X.2012.12.04.
Extracting maximal information from gene signature sets (GSSs) via microarray-based transcriptional profiling involves assigning function to up and down regulated genes. Here we present a novel sample scoring method called Signature-score (S-score) which can be used to quantify the expression pattern of tumor samples from previously identified gene signature sets. A simulation result demonstrated an improved accuracy and robustness by S-score method comparing with other scoring methods. By applying the S-score method to cholangiocarcinoma (CAC), an aggressive hepatic cancer that arises from bile ducts cells, we identified enriched oncogenic pathways in two large CAC data sets. Thirteen pathways were enriched in CAC compared with normal liver and bile duct. Moreover, using S-score, we were able to dissect correlations between CAC-associated oncogenic pathways and Gene Ontology function. Two major oncogenic clusters and associated functions were identified. Cluster 1, which included beta-catenin and Ras, showed a positive correlation with the cell cycle, while cluster 2, which included TGF-beta, cytokeratin 19 and EpCAM was inversely correlated with immune function. We also used S-score to identify pathways that are differentially expressed in CAC and hepatocellular carcinoma (HCC), the more common subtype of liver cancer. Our results demonstrate the utility and effectiveness of -score in assigning functional roles to tumor-associated gene signature sets and in identifying potential therapeutic targets for specific liver cancer subtypes.
通过基于微阵列的转录谱从基因特征集（GSSs）中提取最大信息涉及为上调和下调基因赋予功能。在此，我们提出一种名为特征分数（S分数）的新型样本评分方法，该方法可用于量化来自先前确定的基因特征集的肿瘤样本的表达模式。模拟结果表明，与其他评分方法相比，S分数方法具有更高的准确性和稳健性。通过将S分数方法应用于胆管癌（CAC），一种由胆管细胞产生的侵袭性肝癌，我们在两个大型CAC数据集中确定了富集的致癌途径。与正常肝脏和胆管相比，CAC中有13条途径富集。此外，使用S分数，我们能够剖析CAC相关致癌途径与基因本体功能之间的相关性。确定了两个主要的致癌簇及其相关功能。包含β-连环蛋白和Ras的簇1与细胞周期呈正相关，而包含转化生长因子-β、细胞角蛋白19和上皮细胞黏附分子的簇2与免疫功能呈负相关。我们还使用S分数来识别在CAC和肝癌更常见的亚型肝细胞癌（HCC）中差异表达的途径。我们的结果证明了S分数在为肿瘤相关基因特征集赋予功能角色以及识别特定肝癌亚型潜在治疗靶点方面的实用性和有效性。
