RNA除了可以在蛋白質合成的過程中扮演攜帶基因密碼的角色之外，它在細胞當中還能行使許多生物功能，像是基因調控、RNA修飾以及染色體複製。因為RNA 的功能主要會取決於其三級結構，所以能夠有效率且精準地比較出兩個RNA三級結構的工具對於了解RNA的功能是重要的。iPARTS2是我們實驗室在2016年所發展出來的工具，它會同時使用RNA一級與三級結構資訊來比對的兩個RNA的三級結構。雖然iPARTS2有考慮RNA的一級與三級結構資訊，但是它並沒有考慮到RNA的二級結構資訊。事實上，有一些研究已指出二級結構資訊在比對兩個RNA的三級結構是有幫助的。在本研究中，我們開發了一套名為iPARTS3的工具，它可以利用RNA一級、二級與三級結構資訊去比對兩個RNA的三級結構。iPARTS3 利用下面三個步驟計算出兩個RNA三級結構比對：第一步驟，我們會利用所謂的結構字元集(Structural Alphabet)方法將輸入的兩個 RNA 三級結構編碼成結構字元編碼序列(SA-encoded Sequences)。第二步驟，我們同時考慮RNA一級、二級與三級結構資訊去設計出一個動態規劃(Dynamic Pro-gramming)演算法，這個演算法可以計算出兩個結構字元編碼序列之間最佳的比對結果。第三步驟，我們把MaxSub演算法應用在兩個結構字元編碼序列的最佳比對結果上，使得在兩個RNA三級結構的重疊中比對在一起的C3’原子數目愈多愈好。最後，我們的實驗結果已證實了iPARTS3在RNA三級結構比對的品質上確實比其它的工具，像是iPARTS2、SARA與SETTER，有比較好的表現。然而，在RNA功能註解的品質上，iPARTS3雖不及iPARTS2，但仍可勝過SARA與SETTER。

In addition to transmission of genetic codes for protein synthesis, RNA is capable of per-forming a wide range of biological functions in cells, including gene regulation, RNA modifi-cation and chromosome replication. Because the functions of RNAs are largely determined by their three dimensional (3D) structures, tools capable of efficiently and accurately comparing two RNA 3D structures are important for understanding the functions of RNA. iPARTS2 was developed by our Lab in 2016 for aligning two RNA 3D structures based on primary and ter-tiary structure information. Although iPARTS2 considers the primary and tertiary structure information, it doesn’t consider secondary structure information. In fact, several studies have shown that secondary structure information is useful to compare two RNA 3D structures. In this study, we develop an RNA pairwise structure alignment tool called iPARTS3, which con-siders the primary, secondary and tertiary structure information of the input RNAs. iPARTS3 calculates an alignment of two RNA 3D structures by the following steps. First, we use the so-called structural alphabet (SA) approach to reduce the input RND 3D structures into two SA-encoded sequences. Second, we consider the primary, secondary and tertiary structure in-formation of RNAs to design a dynamic programming algorithm that can compute an optimal alignment of the two SA-encoded sequences. Third, we apply MaxSub algorithm to the opti-mal alignment of the two SA-encoded sequences such that in the superimposition of two RNA 3D structures, the number of aligned C3’ atom is as maximum as possible. Finally, our experimental results have shown that iPARTS3 indeed has better performance than the other tools, such as iPARTS2, SARA and SETTER, in terms of the quality of structural alignment. However, in terms of the quality of RNA functional annotation, iPARTS3 is inferior to iPARTS2, but it still outperforms SARA and SETTER.

中文摘要..........................................1
Abstract.........................................2
Acknowledgement..................................3
Contents.........................................4
List of figures..................................5
List of tables...................................6
Chapter 1 Introduction..........................7
Chapter 2 Algorithm of iPARTS3.................10
2.1 Encoding RNA 3D as 1D SA-encoded sequence...10
2.2 Dynamic programming algorithm...............11
2.3 MaxSub algorithm............................18
Chapter 3 Experimental results.................19
3.1 Testing datasets and method parameters......19
3.2 Experiments.................................20
Chapter 4 Conclusion..........................23
References......................................24

[1] Gesteland,R.F., Cech,T. and Atkins,J.F. (2006) The RNA world : the Nature of Modern RNA Suggests a Prebiotic RNA World, 3rd edn, Cold Spring Harbor Laboratory Press, NY.
[2] Kolodny,R. and Linial,N. (2004) Approximate protein structural alignment in polynomial time. Proceedings of the National Academy of Sciences of the United States of America, 101, 12201–12206.
[3] Capriotti,E. and Marti-Renom,M.A. (2008) RNA structure alignment by a unit-vector approach. Bioinformatics, 24, i112–i118.
[4] Capriotti,E. and Marti-Renom,M.A. (2009) SARA: a server for function annotation of RNA structures. Nucleic Acids Research, 37, W260–W265.
[5] Hoksza,D. and Svozil,D. (2012) Efficient RNA pairwise structure comparison by SET-TER method. Bioinformatics, 28, 1858–1864.
[6] Čech,P., Svozil,D. and Hoksza,D. (2012) SETTER: web server for RNA structure com-parison. Nucleic Acids Research, 40, W42–W48.
[7] Chung-Han Yang, Cheng-Ting Shih, Kun-Tze Chen, Po-Han Lee, Ping-Han Tsai, Jian-Cheng Lin, Ching-Yu Yen, Tiao-Tin Lin and Chin Lung Lu. (2016) iPARTS2: an improved tool for pairwise alignment of RNA tertiary structures, version 2. Nucleic Acids Research, 44, W328–W332.
[8] Jiang,T., Lin,G., Ma,B., and Zhang,K. (2002) A general edit distance between RNA structures. Journal of Computational Biology, 9, 371–388.
[9] Ortiz,A.R., Strauss,C.E. and Olmea,O. (2002) MAMMOTH (matching molecular models obtained from theory): an automated method for model comparison. Protein Science, 11, 2606–2621.
[10] Siew,N., Elofsson,A., Rychlewski,L. and Fischer,D. (2000) MaxSub: an automated measure for the assessment of protein structure prediction quality. Bioinformatics, 16, 776–785.
[11] Henikoff,S., Henikoff,J.G. (1992) Amino-acid substitution matrices from protein blocks. Proceedings of the National Academy of Sciences of the United States of America, 89, 10915–10919.
[12] Leontis,N.B. and Zirbel,C.L. (2012) Nonredundant 3D structure datasets for RNA knowledge extraction and benchmarking. In: Leontis,NB and Westhof,E (eds). RNA 3D Structure Analysis and Prediction. Springer, NY, pp. 281–298.
[13] Makio Tamura, Donna K. Hendrix, Peter S. Klosterman, Nancy R. B. Schimmelman, Steven E. Brenner and Stephen R. Holbrook (2004) SCOR: Structural Classification of RNA, version 2.0. Nucleic Acids Research, 32, D182–D184.