探討人類微型核醣核酸轉錄因子於蛋白質交互作用網路的協同調控特性與模型

Abstract

系統生物學已成為生物醫學研究中一個新興的領域，其著重於以系統的觀點來了解生物系統的運作。由於細胞內的各種生物反應過程，皆由許多基因與蛋白質共同參與，形成複雜的生物分子網路，因此，建構出完整基因調控網路的系統模型，對於了解細胞的生理，以及疾病的分子機制更顯重要。 近年來，微型核糖核酸（microRNA，簡稱miRNA）被發現普遍存在於動植物以及某些病毒之中，長度為~22個鹼基，並不會轉錄成蛋白質，但是會和細胞內與其部分序列互補的mRNA結合，破壞mRNA的穩定性或抑制其轉錄為蛋白質，進而調控基因在蛋白質層面上的表現。過去的研究發現，microRNA參與生物的發育、分化、生長及代謝等過程，調控基因的表現；此外，已發現microRNA與基因轉綠因子（transcription factor）存在一些特殊的相關性。雖然microRNA對單一基因的調控機制已經有許多研究，其分子反應機制也日漸明朗，但是因為單一的microRNA所調控的基因眾多，所以無法有效的辨別microRNA在生物反應機制中的功能或是角色，以及如何與轉錄因子合作調控仍然不清楚，需要進一步的研究與探討。 本論文結合了大規模的生物實驗數據以及電腦演算法預測的結果分別建立了microRNA調控蛋白質交互作用網路的模型，以及轉錄因子調控蛋白質交互作用網路的模型，利用電腦程式篩選並分析找出可能存在的調控模型，並分析這些模型對蛋白質交互作用網路中的影響。 本篇論文分析結果顯示，在蛋白質交互作用關係中會伴隨著被microRNA或是轉錄因子共調控（co-regulate）以及crosstalk（間接交談），也就是說microRNA或是轉錄因子會相互藉著兩個或兩個以上去影響特定的生物反應機制。特別是在crosstalk這種調控模型，是本論文發現的一種較為特殊的調控模型，結果顯示microRNA或是轉錄因子會藉由調控第三者的基因間接調控蛋白質交互作用，有50%以上（P值<0.001）的基因都會被此種模型調控，並且間接影響蛋白質交互作用。

Systems biology is a rising field of biomedical research. It uses systematic analysis to characterize or discover the mechanism of biological systems. Biological processes involve numerous genes and proteins, which interact with each other to form complicated molecular networks. Thus, finding and conducting a complete module of gene regulatory network is important for understanding mechanisms of cellular physiologies or diseases. In recent years, microRNAs (miRNAs, small RNA molecules with ~22 bps) have been widely discovered in plants, animals and viruses. MicroRNAs will not be translated to proteins but can recognize target genes based on complementary sequence similarity and then suppress the translations of target genes. Previous studies found that microRNAs are involved in growth development, differentiation and metabolism through down-regulating target genes. Furthermore, there exist some characteristic correlations between microRNAs and transcription factors (TFs). Although many studies have uncovered the mechanism of microRNA biogenesis and functions, their corporation with transcription factors remains unclear. One of the reasons is that microRNAs usually have numerous target genes, which made functional prediction or classification difficult and complicated. Here we build a model of microRNAs and transcription factors regulating human protein-protein interaction networks from experimental results and computational predictions. Using computational analysis finds out possible existing regulatory motifs and relations between protein-protein interaction networks. Our results show that protein-protein interaction networks might be regulated by microRNAs and/or transcription factors through their co-regulation and crosstalk regulatory motifs. In other words, more than two microRNAs or transcription factors might be involved in specific biological processes together. Most interestingly, we have identified crosstalk motifs for the first time and found that microRNAs or transcription factors might indirectly regulate protein-protein interactions through regulating a third gene. More than half of all the genes participate in crosstalk motifs (P-value < 0.001).