【摘要】： 流感是由流感病毒引起的具有高度傳染性的一種急性呼吸道疾病,由于它的傳播速度比較快,且流感病毒容易發(fā)生變異,所以每年都會造成不同規(guī)模的流感流行。本研究首先借助近幾年對上海地區(qū)的流感監(jiān)測信息通過構(gòu)建抗原進化圖譜和基因進化圖譜呈現(xiàn)上海地區(qū)近年人群甲型H3N2和H1N1亞型流感病毒的抗原進化和基因進化特點,并深入探討了其差異的原因。同時我們篩選出了人群甲型H1N1亞型流感病毒HA1片段的潛在免疫顯性位點并建立了其抗原變異預測模型,主要內(nèi)容和研究結(jié)果如下： 1.為研究上海地區(qū)近年來人群甲型H3N2和H1N1亞型流感病毒血凝素抗原進化和基因進化的特點,并深入探討其抗原變異的規(guī)律,我們構(gòu)建了相應(yīng)的抗原進化圖譜和基因進化圖譜。結(jié)果顯示上海地區(qū)2007-2008年甲型H3N2亞型流感病毒的抗原進化和基因進化基本上按各年份有聚集趨勢,大體來說比較相似。2007和2008年的部分病毒株之間存在著抗原交叉,可能會有交叉保護。2008年上海流行株和現(xiàn)在的疫苗株A/Brisbane/10/2007雖然HA1基因上變異不大,但在可能的潛在免疫顯性位點194處發(fā)生了變異,所以現(xiàn)行疫苗株可能對上海地區(qū)人群不能提供足夠有效的免疫保護。研究結(jié)果同時顯示上海地區(qū)2005-2008年甲型流感病毒H1N1亞型的抗原進化圖譜和基因進化圖譜大體上都是按各年份聚集。我們發(fā)現(xiàn)140位點的變異可能是近年上海地區(qū)H1N1亞型流感病毒抗原進化的主要原因。雖有一定進化,但是部分病毒株彼此存在抗原交叉保護說明這幾年H1N1進化還是比較緩慢。而對于病毒株A/Shanghai/MH79/2008由于疫苗株A/Brisbane/59/2007不能對其形成有效保護,應(yīng)該對它以后在上海的流行要提高警惕。 為研究上海地區(qū)歷年人群甲型H3N2亞型流感病毒的抗原進化和基因進化特點,我們根據(jù)H3N2抗原變異預測模型用基因序列數(shù)據(jù)預測了其抗原進化圖譜,同時我們也構(gòu)建了其HA1基因進化圖譜。結(jié)果顯示抗原進化和基因進化特點比較相似,都按年份呈現(xiàn)一定的規(guī)律性。 為研究全球歷年人群甲型H1N1亞型流感病毒血凝素抗原進化和基因進化特點,我們構(gòu)建了其抗原進化圖譜和基因進化圖譜。結(jié)果都顯示各年份進化呈現(xiàn)一定的規(guī)律性,但各年份之間交叉比較明顯,同時看出其基因進化呈現(xiàn)連續(xù)進化的過程,而抗原進化是個跳躍進化的過程,在氨基酸位點N54K、T127N、H193R處的變異可能是抗原進化跳躍最主要的原因。 2.我們重點研究了篩選人群甲型H1N1亞型流感病毒HA1片段潛在免疫顯性位點的方法并建立其抗原變異預測模型。采用了SVM-RFE(linear)、SVM-RFE(RBF)以及隨機森林模型作為篩選位點的方法,同時用支持向量機、隨機森林模型和Ridge偏最小二乘回歸作為預測模型,共10種組合方法,通過重復抽樣技術(shù)在相同條件下比較了10種組合方法的預測效果,用驗證分類誤差、特異度和靈敏度作為預測效果好壞的評價指標。研究發(fā)現(xiàn)隨機森林模型作為預測模型比其它預測模型的效果都要好,所以選用隨機森林模型作為H1N1亞型流感病毒抗原變異的預測模型。隨后確定了納入模型的變量個數(shù)為30,此時SVM-RFE(linear)/RF、SVM-RFE(RBF)/RF和RF/RF的預測一致率分別為88.05%、88.05%和87.66%,預測效果都比較好。 本研究篩選出了人群甲型H1N1亞型流感病毒HA1片段的23個潛在免疫顯性位點,分別是36,43,54,69,71,73,80,96,121,125,127,128,140,165,169,189,192,193,204,251,270,271,282,其中有10個屬于抗原決定簇區(qū)域。我們識別出了上海H1N1亞型流感病毒抗原進化圖譜中兩簇差異最重要的位點140。同樣全球歷年人群甲型H1N1亞型流感病毒抗原進化圖譜中兩大類間發(fā)生變異的位點N54K、T127N、H193R也屬于我們篩選出的23個潛在免疫顯性位點,進一步映證了我們的研究結(jié)果可靠。 3.比較研究了新甲型H1N1流感病毒和中國分離的甲型H1亞型流感病毒HA基因序列,構(gòu)建了其基因進化圖譜。結(jié)果顯示中國流行的人甲型H1亞型流感病毒以及WHO近年推薦的疫苗株和目前流行的新甲型H1N1流感病毒在HA基因進化上相距很遠,從基因?qū)用嫔咸崾疚覈巳杭韧庖吆鸵呙缈赡懿荒芴峁┯行ПＷo。此外,新甲型H1N1流感和中國豬甲型H1亞型流感病毒在HA基因進化上仍然存在有一定的距離,表明這次始于北美的新甲型H1N1流感并不是來自于中國。 通過本研究,對上海地區(qū)近年來人群甲型流感病毒的抗原進化和基因進化過程有了直觀的理解,同時深入了解了其抗原變異和基因變異的規(guī)律,為防控流感提供了有價值的線索。另外,篩選出的人群甲型H1N1亞型流感病毒HA1片段的潛在免疫顯性位點為WHO選擇流感疫苗和H1N1亞型流感病毒的進化研究提供了科學依據(jù)。我們建立的H1N1抗原變異預測模型可在不進行HI試驗的情況下預測抗原變異情況,節(jié)省了大量的人力、物力和時間,提高了WHO選擇流感疫苗株的效率。
文內(nèi)圖片：
圖片說明：抗原進化圖譜
[Abstract]:Influenza is a highly contagious, acute respiratory disease caused by the influenza virus, because it is fast in speed and the influenza virus is susceptible to variation, so each year will result in a different scale of influenza. In this study, the evolution and genetic evolution of influenza A (H3N2) and H1N1 (H1N1) virus in Shanghai in recent years were presented by means of the evolution of the antigen and the evolution of the gene, and the reason of the difference was also discussed. At the same time, we screened the potential immunodominant site of the HA1 fragment of the human influenza A (H1N1) subtype influenza virus and established its antigen variation prediction model. The main contents and the results of the study are as follows: 1. To study the characteristics of the evolution and gene evolution of the hemagglutinin antigen of influenza A (H3N2) and H1N1 (H1N1) influenza virus in Shanghai in recent years, and to probe into the law of its antigenic variation, we constructed the corresponding antigen evolutionary map and gene. The results show that the antigenic evolution and gene evolution of the influenza A H3N2 subtype influenza virus in Shanghai in 2007-2008 are basically similar in all years, and there is an anti-virus strain between 2007 and 2008. The original crossover may have cross-protection. In 2008, the Shanghai epidemic and the present vaccine strain A/ Bridlane/10/2007, while the HA1 gene was not large, but at the possible potential immunodominant site 194, the current vaccine strain may not provide enough effective immunity to the population in the Shanghai area The results of the study also show that the antigenic evolution and genetic evolution of the H1N1 subtype of the influenza A virus in Shanghai in 2005-2008 are generally in the following year. We found that the mutation of the 140 site may be the main cause of the evolution of the H1N1 influenza virus antigen in Shanghai in recent years. For reasons, there is some evolution, but some of the virus strains have an antigen cross-protection to each other, indicating whether the H1N1 evolution or the ratio in the past few years The virus strain A/ Shanhai/ MH79/2008 cannot be effectively protected by the vaccine strain A/ Bridane/59/2007, and it should be mentioned later in Shanghai In order to study the antigenic evolution and genetic evolution of the influenza A H3N2 subtype influenza virus in the Han population in Shanghai, we used the gene sequence data of the H3N2 antigen variant prediction model to predict its antigen evolution map, and we also constructed the HA. The results showed that the evolution of the antigen and the character of the evolution of the gene were similar, both in the same year. In order to study the evolution of the hemagglutinin antigen and the genetic evolution of the influenza A (H1N1) influenza virus in the global years, we constructed the antigen. The results show that the evolution of the evolution shows a certain regularity, but the cross-correlation between the different years is obvious, and the evolution of the gene is a process of continuous evolution, and the evolution of the antigen is a process of jumping and evolution, in the amino acid position. The variation at points N54K, T127N, H193R may be anti- 2. We focus on the potential immunodominant sites of the HA1 fragment of the influenza A (H1N1) influenza virus in the screening population. Based on the method of SVM-RFE (linear), SVM-RFE (RBF) and random forest model as the screening site, SVM-RFE (linear), SVM-RFE (RBF) and random forest model are used as the screening site, and the support vector machine, the random forest model and the Ridge partial least-squares regression are used. As a prediction model, a total of 10 combination methods were used, and the prediction effect of 10 combinations was compared under the same conditions by the repeated sampling technique, and the error, specificity and spirit of the method were used to verify the classification error. In this paper, the random forest model is selected as the H1N1 (H1N1) model as the prediction model, and the effect of the model is better than that of other prediction models. The prediction model of the variant of influenza virus antigen was determined. The number of variables included in the model was 30. At this time, the consistent rate of SVM-RFE (linear)/ RF, SVM-RFE (RBF)/ RF and RF/ RF was 88.05%, 88.05%, respectively. In this study,23 potential immunodominant sites,36,43,54,69,71,73,80,96,121,125,127,128,140,165,169,189,192,193,204,251,270,271, 282,10 of which belong to the region of the antigenic determinant. We identified the anti-influenza virus of the Shanghai H1N1 subtype. The most important sites of the two clusters in the original evolutionary map are the two major sites in the evolution of the influenza A (H1N1) influenza virus antigen in the same global calendar year. The mutation sites, N54K, T127N, and H193R, also belong to the 23 potential immunogens that we screened. The results of this study are more reliable.3. A comparison of the new influenza A (H1N1) virus and the Chinese-isolated influenza A (H1) The results show that the influenza A H1 subtype influenza virus in China and the vaccine strain recommended by WHO in recent years and the new influenza A (H1N1) virus in recent years are far from the evolution of the HA gene, from the gene level It is suggested that the prior immunization and the vaccine of the Chinese population may not provide effective protection. In addition, the new influenza A (H1N1) and the influenza A H1 subtype of the Chinese swine still have a certain distance in the evolution of the HA gene, indicating that this The new H1N1 influenza, which started in North America, is not from China. By this study, the evolution of the antigen and the evolution of the gene of the influenza A virus in the population in Shanghai in recent years have a visual understanding, and at the same time, it is well understood The law of antigenic variation and gene mutation provides a valuable clue for the prevention and control of influenza. In addition, the potential immunodominant position of the HA1 fragment of the influenza A/ H1N1 subtype influenza virus of the selected population is the WHO choice. The evolutionary study of the influenza vaccine and the H1N1 subtype of influenza virus provides a scientific basis. The model of the H1N1 antigen variation established by us can predict the variation of the antigen without the HI test and save the large amount.
【學位授予單位】：復旦大學
【學位級別】：博士
【學位授予年份】：2010
【分類號】：R392.1

【參考文獻】

相關(guān)期刊論文 前10條

1 王勇,陳淑霞,薛穎,董杰,金奇,侯云德;中國歷年H_3N_2亞型人流行性感冒病毒血凝素基因的序列測定及分析[J];病毒學報;2002年02期

2 程從升;舒躍龍;張智清;;流感病毒的反向遺傳學研究進展[J];病毒學報;2007年01期

3 董麗波;張燁;溫樂英;趙翔;黃維娟;藍雨;郭俊峰;李梓;王敏;董婕;郭元吉;舒躍龍;;1995～2005年中國H3N2亞型人流感病毒血凝素基因變異與流行相關(guān)性研究[J];病毒學報;2007年05期

4 李哲謙;劉書朋;嚴壯志;黃海;;基于支持向量機的蛋白質(zhì)相互作用預測[J];電子測量技術(shù);2008年05期

5 朱爾一;林燕;莊贊勇;;偏最小二乘變量篩選法在毒品來源分析中的應(yīng)用[J];分析化學;2007年07期

6 李偉紅;龔衛(wèi)國;陳偉民;梁毅雄;尹克重;;基于SVM RFE的人臉特征選擇方法[J];光電工程;2006年05期

7 劉頌,王京燕;抗流感病毒藥物的研究進展[J];國外醫(yī)學.藥學分冊;2005年02期

8 修文瓊;翁育偉;沈曉娜;謝劍鋒;楊式芹;吳冰珊;王美愛;;1996～2004年中國福建省甲3亞型流感病毒(H3N2)的分子進化[J];中國科學(C輯:生命科學);2008年06期

9 陳啟買;陳森平;;支持向量機的一種特征選取算法[J];計算機工程與應(yīng)用;2009年23期

10 張麗新;王家欽;趙雁南;楊澤紅;;機器學習中的特征選擇[J];計算機科學;2004年11期

相關(guān)博士學位論文 前3條

1 張文彤;甲型流感病毒H3抗原進化及變異規(guī)律研究[D];復旦大學;2005年

2 唐發(fā)明;基于統(tǒng)計學習理論的支持向量機算法研究[D];華中科技大學;2005年

3 蔣露芳;上海地區(qū)人群甲型流感的傳播與病毒抗原變異的研究[D];復旦大學;2007年

相關(guān)碩士學位論文 前4條

1 武嬌;偏最小二乘回歸模型及其在教育統(tǒng)計中的應(yīng)用[D];陜西師范大學;2002年

2 彭國蘭;隨機森林在企業(yè)信用評估中的應(yīng)用[D];廈門大學;2007年

3 張小丹;基于支持向量機的基因表達數(shù)據(jù)特征選取方法研究[D];蘇州大學;2008年

4 楊吉星;上海地區(qū)甲型流感病毒HA基因和抗原變異研究[D];復旦大學;2009年

，

本文編號：2509883