一個使用全基因資訊建立演化樹的創新方法

Abstract

建立演化樹在計算上是個困難的問題。隨著加入分析的物種增多、可能的演化樹數量也急遽增多，使得其計算也非常耗時。要同時降低計算的耗時並且準確的建立演化關係一直是我們所必須克服的挑戰。 隨著現今定序技術不斷的進步、可以取得的全基因資料量日益增長，也因此開啟了我們使用全基因的資訊來建立演化樹的可能。目前已經有多種使用全基因建演化樹的方法被開發出來。然而，面對基因組內不同基因之間相異的演化速率，仍然還沒有一個足夠妥當的方法來解決；同時，龐大的序列也造成計算上的負荷。 我們引入一個概念：在三個物種A、B和C的同源基因中，A與B之間的演化距離和A與C之間的演化距離會保持一個穩定的比例關係。應用這個概念，我們提出一個嶄新的方法來建立的演化樹。我們更進一步使用遺傳演算法或是高斯牛頓法來加速搜尋最佳解。 從測試的結果顯示、我們的方法所建立出來的演化樹能夠穩固地維持演化關係。即在模擬基因有所缺失的情況下，演化關係的訊息仍然可以很好地被保留下來。所以，在全基因組的資料量不斷成長的現在，我們的方法有望提供建立全基因樹一個強力解答。

Phylogeny reconstruction is a difficult computational problem. As the number of included taxa grows, the number of possible solutions would increase dramatically and the computation would thus become quite time-consuming. Reducing the time cost and inferring phylogeny precisely are the challenges to overcome with. In addition, with the progress of sequencing technology, many whole genome sequencing data are available now and therefore constructing phylogeny using these whole genome data becomes more and more possible. There have been many methods developed using whole genome data. However, how to manage the various evolutionary rates among genes in a genome appropriately still remains unresolved. Since the ratio of the genetic distances between species A and B relative to that between species A and C should be conserved, we adopted this concept and proposed a novel method to construct our phylogenetic tree. Further, to look for the optimal tree, we applied genetic algorithms and Gaussian-Newton method to accelerate our searching process. The result showed that, our approach is robust on keeping the phylogenetic relationship. In condition that some genes were lost, the tendency of phylogenetic relationship would still be maintained well. Our method is hopefully to be an efficient way to infer the phylogeny for the increasing whole genome data.