河麂（Hydropotes inermis）及鹿类动物物种分子识别研究

发布时间：2018-04-21 01:31

本文选题：河麂 + 鹿科　；参考：《沈阳师范大学》2017年硕士论文

【摘要】：鹿类动物在生物多样性和生态系统中具有非常重要地位。由于生态环境的破坏、栖息地丧失和非法捕猎等因素的共同作用,其野生种群数量在急剧减少,有的甚至已经灭绝,许多鹿类动物已被列为濒危物种。河麂(Hydropotes inermis)等鹿类动物的保护还没有引起足够的重视。DNA条形码序列能够实现从一个未知样本到已知物种的鉴别,但这个目标的实现需要可靠的数学运算和分析方法。由于需要分类的生物类群不同或采用的DNA条形码基因位点及序列长度的不同,在物种识别能力和效果上存在一定的差异。我们以河麂保护为目的,对鹿类动物进行分子识别分析方法比较研究,为河麂等鹿类动物的皮毛、茸、肉等制品的鉴定提供基础依据。本研究的样本以粪便、毛发等无损伤取样为主,提取基因组DNA。以线粒体细胞色素C氧化酶亚单元I基因(COI)片段作为DNA条形码和细胞色素b基因标记识别河麂等鹿类动物,并进行河麂在鹿类动物的系统发育地位的分析。对鹿类动物的8个物种30个样本的COI基因和Cytb基因部分片段进行了测序,得到700bp的COI序列和792bp长度的Cytb基因序列。从GenBank和BOLD数据库下载了相关13个物种的43条COI序列,并以此作为DNA条形码参考序列。依据相关理论和算法对测序获得的COI序列样本进行分类指派分析,采用11种指派分析方法包括条形码空隙探查法、邻接树法、条形码逻辑公式算法、最近邻算法、决策树算法、规则RIPPER算法、随机森林法、支持向量机算法、神经网络反向传播算法、模糊数据算法、贝叶斯算法。结果显示,这些方法在鹿类动物分类中的解析能力和准确性存在较大差异。随机森林法和支持向量机法对鹿类动物DNA条形码分类最为可靠。综合分析各种因素,作者认为对各种方法表现性能影响最大的因素是分子数据多态性。通过增加基因位点、增加测序长度和样本量可以改进多数方法的预测性能,另外,选择对分类样本最适用的分析方法可以明显提高分析质量。为进一步探讨基因位点数和序列长度对识别效果的影响,以及为评估河麂的系统地位,我们将COI基因和Cytb基因合成1492bp长度的串联数据,采用最大似然法和贝叶斯法构建系统发育树。与单独采用COI基因构建的系统树相比,序列样本聚类关系基本一致,但置信值更高。河麂在鹿类动物中属于相对古老的物种,与狍的亲缘关系最近。本次研究有目的地采集了河麂的2个地理亚种样本,但物种识别分析和系统发育分析并没有将二者明显区分开,即未能将朝鲜河麂与华南河麂有效划分,这可能反映了DNA条形码分析在亚种级别的局限性,或这2个地理亚种分歧时间较晚。
[Abstract]:Deer play an important role in biodiversity and ecosystem. Due to the destruction of ecological environment, habitat loss and illegal hunting and other factors, its wild population is rapidly reduced, some have even been extinct, many deer have been listed as endangered species. The protection of deer, such as Hydropotes inermisis, has not attracted enough attention. DNA barcode sequences can identify species from an unknown sample to a known species, but the realization of this goal requires reliable mathematical operations and analysis methods. There are some differences in the ability and effect of species recognition because of the difference of the biological groups that need to be classified or the different DNA barcode gene loci and sequence length. We compared and studied the molecular recognition and analysis methods of deer in order to provide the basis for the identification of fur, antler, meat and other products of deer such as Heji. In this study, samples were collected mainly from feces and hair, and genomic DNAs were extracted. The mitochondrial cytochrome C oxidase subunit I gene was used as DNA bar code and cytochrome b gene marker to identify deer and other deer, and the phylogenetic status of Heji in deer was analyzed. The COI gene and Cytb gene partial fragment of 30 samples from 8 species of deer were sequenced. The COI sequence of 700bp and the Cytb gene sequence of 792bp length were obtained. 43 COI sequences of 13 species were downloaded from GenBank and BOLD databases and used as DNA barcode reference sequences. According to the relevant theories and algorithms, the COI sequence samples obtained by sequencing are classified and assigned. Eleven methods of assignment analysis are used, including barcode gap detection method, adjacency tree method, bar code logic formula algorithm, nearest neighbor algorithm, decision tree algorithm, etc. Rule RIPPER algorithm, Random Forest algorithm, support Vector Machine algorithm, Neural Network back Propagation algorithm, Fuzzy data algorithm, Bayesian algorithm. The results showed that there were significant differences in analytical ability and accuracy of these methods in deer classification. Random forest method and support vector machine method are the most reliable methods for DNA barcode classification of deer. Comprehensive analysis shows that molecular data polymorphism is the most important factor affecting the performance of various methods. The prediction performance of most methods can be improved by adding gene loci, sequencing length and sample size. In addition, selecting the most suitable analysis method for classified samples can obviously improve the analysis quality. In order to further study the effect of site number and sequence length on recognition, and to evaluate the systematic status of Heji, series data of the length of 1492bp synthesized by COI gene and Cytb gene were used. The phylogenetic tree was constructed by maximum likelihood method and Bayesian method. Compared with the system tree constructed by COI gene alone, the clustering relationship of sequence samples is basically consistent, but the confidence value is higher. River muntjac is a relatively old species in deer, and has the closest relationship with roe deer. In this study, two geographic subspecies samples of Heji were collected, but the species identification analysis and phylogenetic analysis did not distinguish them clearly, that is to say, they could not be effectively divided between Korean and South China river muntjac. This may reflect the limitation of DNA barcode analysis at subspecies level or the late divergence of these two geographic subspecies.
【学位授予单位】：沈阳师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：Q959.842

【参考文献】