基於極值定理適應模型之A型流感病毒序列的演化模擬

Abstract

大部分物種之間的表型演化像是形態學、生理學以及物種間行為的差異可以解釋為天擇的適應結果。當環境改變後，適應機制會使物種由原本的狀態進一步的改變為新的表現型更好的狀態。在這篇研究中，我們提供了一個分析A型流感病毒(IAV)演化的演化模擬。演化包含了上述的演化架構並可產生一個由初始A型流感病毒演化至給予的特定目標A型流感病毒的演化路徑。由模擬出的路徑，我們可以獲得並分析每個演化中的突變步驟，這很難由系統發生樹分析獲得。由於演化機制包含了複雜的演化機率、基因多效性以及群體遺傳學，因此我們基於序列模型來設計胺基酸序列中狀態空間的適應模擬，而非使用連續的狀態來避免適應在收斂到最佳狀態解時的過程太過複雜。為了驗證模擬器的可行性，我們測試兩個A型流感的內部蛋白PB1與PB2以及一個外部蛋白HA。演化模擬器提供了A型流感病毒蛋白質適應變異性的結果且符合先前的相關研究，並且提供預測有可能的演化趨勢，以便於近一步研究。

Most phenotypic evolution within species as well as morphological, physiological, and behavioral differences between species can be explained by adaptation due to natural selection. After a change in environment, the task of adaption for a species is to move the population from its current state toward a new phenotypical optimum one. In this study, we have developed an evolution simulator to analyzes the evolution of influenza A viruses (IAV). The simulator extends previous adaptive models and generates hypothetical evolutionary trajectories from an initial IAV to a set of pre-specified target IAVs. From the simulated trajectories, we are able to observe and analyze every single hypothetical mutational step in evolution, which cannot be achieved easily from a phylogenetic tree. Because the task of adaptation is complicated by random mutations, pleiotropy, and population genetics, we adapted the sequence-based models to design the simulator that considers adaptation in the discrete state space of amino acid sequences instead of a continuous space to avoid a long extended adaptive walk toward the optimum state. To demonstrate the feasibility of the simulator, we have tested it on IAV’s two internal proteins, PB1 and PB2, and one surface glycoprotein, HA. The evolution simulator provided results on the adaptive variability of IAV proteins, which correlated with previous studies’ findings, and indicated its potential of predicting prospective adaptation trends that warrant further investigation.