Logo image
Improved Algorithms for the Gene Team Tree Problem
Dissertation

Improved Algorithms for the Gene Team Tree Problem

Lin, Chien-Hsin
Doctor of Philosophy (PHD), 國立清華大學, 資訊工程學系
2012

Abstract

演算法 資料結構 計算生物 基因組 基因組樹 基因序列比對 algorithms data structures bioinformatics gene teams gene team trees comparative genomics
Comparing multiple genome sequences is an important method to discover new biological insights. If a group of genes remain physically close to each other in multiple genomes, often called a conserved gene cluster, then the genes may be either historically or functionally related. A gene team is a set of genes that appear in two or more species, possibly in a different order yet with the distance between adjacent genes in the team for each chromosome always no more than a certain threshold δ. A gene team tree is a succinct way to represent all gene teams for every possible value of δ. In this dissertation, new efficient algorithms are presented for the problem of constructing a gene team tree of two chromosomes. For this problem, Zhang and Leong had an O(n lg^2 n)-time algorithm, where n is the number of genes. In this dissertation, two improved algorithms are presented, which require, respectively, O(n lg n lglg n) and O(n lg n α(n)) time, where α(n) is the inverse of Ackermann's function. Similar to Zhang and Leong's gene-team-tree algorithm, the presented algorithms can be extended to k chromosomes with the time complexities increased only by a factor of k, where k > 2 is an integer. In practice, the distance between two genes is integer. In addition to the two improved algorithms, assuming that the positions of genes are integers, this dissertation presents an O(n lg n + z)-time algorithm, where z is the maximum distance between any two adjacent genes. For real-world applications, z is usually smaller than n and thus the third algorithm is more efficient than the other two algorithms. Similarly, it can be extended to k chromosomes. The extended algorithm requires O(kn lg n + z) time.

Metrics

1 Record Views

Details

Logo image