Institute of Molecular Systems Biology, ETH, Wolfgang-Pauli-Strasse 16, 8093 Zurich, Switzerland.
Bioinformatics. 2011 Dec 15;27(24):3407-14. doi: 10.1093/bioinformatics/btr580. Epub 2011 Oct 20.
Deciphering the response of a complex biological system to an insulting event, at the gene expression level, requires adopting theoretical models that are more sophisticated than a one-to-one comparison (i.e. t-test). Here, we investigate the ability of a novel reverse engineering approach (System Response Inference) to unveil non-obvious transcriptional signatures of the system response induced by prion infection.
To this end, we analyze previously published gene expression data, from which we extrapolate a putative full-scale model of transcriptional gene-gene dependencies in the mouse central nervous system. Then, we use this nominal model to interpret the gene expression changes caused by prion replication, aiming at selecting the genes primarily influenced by this perturbation. Our method sheds light on the mode of action of prions by identifying key transcripts that are the most likely to be responsible for the overall transcriptional rearrangement from a nominal regulatory network. As a first result of our inference, we have been able to predict known targets of prions (i.e. PrP(C)) and to unveil the potential role of previously unsuspected genes.
Supplementary data are available at Bioinformatics online.
在基因表达水平上，破译复杂生物系统对侮辱性事件的反应需要采用比一对一比较（即 t 检验）更复杂的理论模型。在这里，我们研究了一种新的反向工程方法（系统响应推理）的能力，以揭示朊病毒感染引起的系统反应的非明显转录特征。
为此，我们分析了先前发表的基因表达数据，从中推断出小鼠中枢神经系统中转录基因-基因相关性的全尺度模型。然后，我们使用这个名义模型来解释朊病毒复制引起的基因表达变化，旨在选择受此干扰影响最大的基因。我们的方法通过识别最有可能导致整体转录重排的关键转录物，揭示了朊病毒的作用模式，这些转录物来自名义监管网络。作为我们推断的第一个结果，我们已经能够预测朊病毒的已知靶标（即 PrP(C)），并揭示以前未怀疑的基因的潜在作用。
补充数据可在生物信息学在线获得。
