真核基因剪接位点二级结构特征

利用RNA二级结构的预测程序,通过对148个人类基因由EST验证的发生剪接的784个供体位点和受体位点以及101个人类基因由EST验证的发生选择性剪接的418个供体位点和受体位点附近的二级结构的预测,寻找真核基因剪接位点及选择性剪接位点的二级结构特征。通过详细研究剪接位点及选择性剪接位点在RNA二级结构中的结构分布,获得了一些剪接位点及选择性剪接位点在RNA二级结构中结构分布的规律性。结果表明,在RNA剪接及选择性剪接过程中,顺式作用元件具有结构特定性。这种结构特定性可以从结构上对真核基因剪接及选择性剪接机制进行一些合理的解释,有利于对真核基因剪接位点的识别及其表达调控机理的进一步理解。     

老鼠基因组盒式外显子和内含子保留型可变剪接位点预测

老鼠和人类基因组的同源性超过90%,老鼠基因组的研究为人类基因组序列研究提供了参考数据.统计分析了老鼠盒式外显子和内含子保留型剪接位点附近的序列保守性特征,并据此分别利用基于多样性指标的支持向量机和二次判别法对老鼠基因组中这两种剪接类型的供体端和受体端可变剪接位点进行了预测.独立检验结果表明,盒式外显子和内含子保留型的供体端和受体端可变剪接位点的预测均能达到较高的识别精度.     

隐Markov模型在剪接位点识别中的应用

剪接位点的识别是基因识别中的一个重要环节。由于现有的基因识别算法主要关注编码区的整体特性 ,而并不着重考虑个别位点的信息 ,因此难以准确地识别出剪接位点。考虑到剪接位点附近的保守序列的相邻碱基之间应该存在某种相关性 ,利用一阶 Markov链建立了表述这种相关性的模型 ,在此基础之上 ,设计了专门用于剪接拉点识别的隐马氏模型 (HMM)方法。实验结果表明 ,用 HMM描述剪接位点附近序列符合实际情况 ,并且利用这一方法进行剪接位点的识别可以很好地提取位点附近保守序列在边缘分布与条件分布 (转移概率 )上的统计特征。使用该方法对真实剪接位点和虚假剪接位点进行识别 ,识别率均可达 90 %以上。     

基于位置关联权重矩阵及DNA结构信息预测人类剪接位点

人类剪接位点的识别是当前研究的一个重要课题.根据人类剪接位点附近区域的保守性,以位置关联权重矩阵及DNA结构信息作为特征输入参数,应用支持向量机(SVM)对人类基因组中的供体端和受体端剪接位点做了预测.对于供体端,5-fold交叉检验总体预测精度为92.55%,3-way data split检验总体预测精度为92.25%;受体端5-fold交叉检验总体预测精度为90.70%,3-way data split检验总体预测精度为89.87%.     

基于特征挖掘与融合的剪接位点识别

在基于保守序列这一信号特征识别剪接位点的基础上．挖掘了可用于剪接位点识别的其他多个特征（包括剪接位点上、下游序列的碱基组成。剪接位点信号和上、下游序列的碱基组成随位点邻近序列C＋G含量的变化等统计特征）,建立了描述这些特征的模型。设计了能有效融合这些特征对剪接位点进行识别的对数线性模型,开发了剪接位点识别程序SpliceKey．测试结果表明：SpliceKey识别剪接位点的精度不仅较WAM方法有显著的提高,而且也优于国际上最新发布的剪接位点识别软件DGSplice．SpliceKey已提供网络服务：http：／／infosci．hust．edu．cn／SpliceKey／．     

Takagi-Sugeno模型在剪接位点识别中的应用

为降低基因剪接位点识别算法复杂度和计算量,根据剪接位点上下游序列的保守特性及碱基组成随位点邻近序列GC含量变化等统计特征,建立Takagi-Sugeno模糊模型.通过模型输出值和阈值比较,判断真实的剪接位点.基于模糊似然函数的模糊聚类算法确定模型结构和前件参数,并结合最小二乘法完成该模型后件参数的识别.仿真结果表明,该算法简单,可使模糊模型的结构辨识和参数辨识同时完成,从而实现模糊模型的快速识别;能够很好地提取剪接位点附近保守序列的统计特征,为剪接位点的识别提供一种新的方法.     

基于HMM的基因剪接供体位点识别改进研究

本文对人类基因供体剪接位点的数据进行词频统计,分析了位点序列的特征及其特征碱基之间的关系。通过改进传统隐马尔可夫模型（HMM）,使其能将各特征碱基以相应权值关联识别位点。结果显示改进模型的识别能力要强于传统模型。同时对特征位点较相似的序列有了更好的区分度。     

基于快速傅里叶变换的剪接特征提取

挖掘剪接特征是剪接位点识别算法的基础,在频域空间挖掘对位点识别有帮助的特征至关重要.利用基于快速傅里叶变换的剪接特征提取方法对其进行特征提取,该方法能够将时域信息转化到频域中,以此来构建所需的频域特征,为了比较还构建了位置特征与统计特征. 实验结果表明将频域特征加入剪接位点识别中能够有效地提高识别精度,这也表明将信号处理方法应用于生物信息学领域是可行有效的.      

RS-GA-KNN算法识别灵长类动物DNA序列剪接位点

以灵长类动物DNA序列的剪接位点识别资料为研究对象,将选定样本序列中各碱基编码作为原始变量数据,用粗糙集方法和遗传算法对原始变量数据进行变量筛选,即以粗糙集方法选取的变量为基础,用遗传算法进行变量的二次搜索,从样本序列各碱基中挑选出保守性强的碱基对应的变量构成变量集,采用最近邻聚类识别灵长类动物DNA序列剪接位点类型,总识别准确率达90．66％,明显高于直接使用原始变量数据或将粗糙集理论方法和遗传算法单独用于变量选取的识别结果．     

人工神经网络和支持向量机在剪接位点识别上的应用

将人工神经网络和支持向量机应用于剪接位点的识别中,并用标准测试数据集进行了5倍率交叉验证,测试结果显示人工神经网络和支持向量机对剪接位点的识别效果优于目前广泛使用的权阵列模型.     

基于序列信息理论预测线虫基因选择性剪切位点

基因的选择性剪切使得在DNA上一段相同的序列翻译成多个不同的蛋白质序列.选择性剪切的出现把剪切位点分为选择性供体位点、组成性供体位点、选择性受体位点和组成性受体位点.基于EBI中的线虫基因选择性剪切位点数据库,选取不同位点的单碱基频率和序列片段的三联体频数作为参数,利用位置权重矩阵和离散增量结合支持向量机,对选择性剪切位点进行了理论预测.对选择性供体位点和选择性受体位点的预测成功率分别为63.78%和72.63%,特异性分别为68.02%和83.96%.     

基于卷积神经网络的基因剪接位点预测

研究剪接位点可以更深入地探索剪接机制和基因预测方法,准确预测剪接位点至关重要。基于深度学习技术提出一种新的预测方法,无需人工提取样本特征,以基因序列的K-MER编码向量作为输入,采用训练后的卷积神经网络(CNN)模型进行预测。基于人类基因HS3D供体数据集,与传统机器学习方法进行预测比较,结果表明预测模型的主要性能指标,包含马修斯相关系数(MCC)、灵敏度(SN)均超过传统的机器学习方法。     

A new method for splice site prediction based on the sequence patterns of splicing signals and regulatory elements

It is of significance for splice site prediction to develop novel algorithms that combine the sequence patterns of regulatory elements such as enhancers and silencers with the patterns of splicing signals. In this paper, a statistical model of splicing signals was built based on the entropy density profile （EDP） method, weight array method （WAM） and K test; moreover, the model of splicing regulatory elements was developed by an unsupervised self-learning method to detect motifs associated with regulatory elements. With two models incorporated, a multi-level support vector machine （SVM） system was devised to perform ab initio prediction for splice sites originating from DNA sequence in eukaryotic genome. Results of large scale tests on human genomic splice sites show that the new method achieves a comparative high performance in splice site prediction. The method is demonstrated to be with at least the same level of performance and usually better performance than the existing SpliceScan method based on modeling regulatory elements, and shown to have higher accuracies than the traditional methods with modeling splicing signals such as the GeneSplicer. In particular, the method has evident advantage over splice site prediction for the genes with lower GC content.     

A role for branchpoints in splicing in vivo

The nucleotides immediately surrounding intron/exon junctions of genes transcribed by RNA polymerase B can be derived from 'consensus' sequences for donor and acceptor splice sites by only a few base changes. Studies in vivo have underlined the importance of these junction nucleotides for splicing. In higher eukaryotes, no evidence has been found for specific internal intron sequences involved in splicing. However, the recent discovery that, in vitro, introns are excised in a lariat form where the 5' end of the intron is joined via a 2'-5'-phosphodiester linkage to an A residue (branchpoint acceptor) close to the 3' end of the intron, suggests that internal intron sequences may nonetheless be important for splicing. Indeed, in yeast nuclear genes, the internal sequence 5'-TACTAAC-3' (or close homologue) is essential for splicing in vivo. A proposed consensus sequence for branchpoints in mammalian introns is 5'-CT(A/G)A(C/T)-3'. This sequence resembles the essential yeast internal sequence. Are branchpoints involved in the splicing of introns of higher eukaryotes in vivo? We show here that a branchpoint sequence from a human globin gene (5'-CTGACTCTCTCTG-3') greatly enhances the efficiency of splicing of a 'synthetic' intron in HeLa cells. A mutated branchpoint sequence, 5'-CTCCTCTCTCTG-3', in which the branchpoint acceptor nucleotide A has been deleted and the neighbouring purine G mutated to a C, does not exhibit this enhancing capability. We conclude that branchpoints have an important function in the splicing process in vivo.     

利用条件随机场实现DNA剪接位点的预测

为解决传统的基因识别算法主要关注编码区的整体特性,而并不着重考虑个别位点的信息,因此难以准确地识别出剪接位点的缺点,提出了基于条件随机场的剪接位点预测方法,条件随机场能够更好的处理标记数据之间的依赖关系,并且能够避免数据标记偏置的问题。实验结果表明基于条件随机场的剪接位点预测方法是一种合适的方法,能够取得更好的效果。     

Correction of RNA aberrant splice increases foreign gene expression in transgenic mice

Transgenic mice with mammary gland secreting human granulocyte colony stimulating factor (G-CSF) were produced using mice whey acid protein gene promoter. It was found that there was very low expression level in mammary gland. Human G-CSF cDNA was obtained by RT-PCR from transgenic mice mammary gland. Sequence analysis showed that this G-CSF gene deleted the 4th exon, and compared with human G-CSF genomic DNA, there were donor and acceptor splice sites in the deletion fragment. It was considered that the 3rd and 4th introns also delete in G-CSF fragment. The transgenic construct was corrected by deleting the 3rd and 4th introns to construct the minigene, which was used to produce transgenic mice by microinjection. Northern blot showed that G-CSF expression using the new construct increased 5.4 times as that before in transgenic mice. The results suggested that it was possible that RNA aberrant splice result in low expression in transgenic mice.