This dissertation, "Structure Comparison in Bioinformatics" by Zeshan, Peng, 彭澤山, was obtained from The University of Hong Kong (Pokfulam, Hong Kong) and is being sold pursuant to Creative Commons: Attribution 3.0 Hong Kong License. The content of this dissertation has not been altered in any way. We have altered the formatting in order to facilitate the ease of printing and reading of the dissertation. All rights not granted by the above license are retained by the author. Abstract: Abstract of thesis entitled "Structure Comparison in Bioinformatics" Submitted by Zeshan PENG for the degree of Doctor of Philosophy at The University of Hong Kong in February 16, 2006 Massive genomic and proteinomic structures have been revealed. All such structures can be categorized into primary structures (one-dimensional sequences), secondary structures (tree oriented structures) and tertiary structures (three-dimensional struc- tures). Thisthesis studiesstructurecomparisonproblemsinBioinformatics forfinding similarities or consensuses of input structures. For primary structure comparisons, we study the problem of online recognition of tandem-repeat-free strings, where a tandem repeat is a string formed by two identical halves and a string without any substring being a tandem repeat is tandem-repeat- free. Wepresentefficientalgorithmsfortheproblemanditsdynamicversions. Wealso study the constrained sequence comparison problem in which an optimal alignment or aconsensusofinputsequencesisreturnedsuchthatitincludesaconstrainedsequence. Space-efficient algorithms ontwo sequences andanapproximate algorithm on multiple sequences are presented. For secondary structure comparisons, we study the Maximum Constrained Agree- ment SubTree (MCAST) problem for phylogenetic analysis of species, where two evolu- tionarytrees, S andT, togetherwithanagreement subtreeofthem, P, areprovidedas inputs, and the objective is to finda Maximum Agreement SubTree (MAST)of S andT such that it includes P as a subtree. A linear time reduction is presented such that the MCAST problem can be solved efficiently by the algorithms for the MAST problem. We study the local forest edit distance problem on two ordered labeled forests, whose objective is to find two most similar subforests of input forests. Efficient dynamic programming algorithms are presented for it. Also we study the constrained forest edit distance problem in which two ordered labeled forests, E and F, and a common subtree of them, G, are provided as inputs, and the objective is to find a sequence ofedit operations with minimum total cost on nodes of E and F such that the resulting forests are identical and include G as a subtree. We present a dynamic programming algorithm for this problem. Fortertiarystructurecomparisons, westudytheproteintertiarystructurecompar- ison problem for finding similar fragments of m input protein tertiary structures. We transform protein triangle representations for finding good superpositions. We build a sequential m-partite graph from m transformed protein structures such that we can find their similar fragments by searching a maximum sequential m-clique matching or a maximum sequential m-star matching. We show that these two matching prob- lems are NP-complete. Furthermore, we present an efficient algorithm for comparing two protein tertiary structures and an efficient heuristic algorithm for comparing m protein tertiary structures. (419 words) DOI: 10.5353/th_b3627129 Subjects: Structural bioinformaticsComputational biology