中文摘要：

核小体定位对真核生物基因表达调控发挥着重要作用。前期基于核小体核心及连接区域的k-mer频次分布偏好性，构建了位置权重矩阵算法，并在酿酒酵母基因组内较好地预测了核小体占据率。利用该理论模型，以1bp碱基为步长、147bp碱基为窗口，用该算法计算了酵母1号、3号、14号染色体上核小体形成能力强、中、弱各3条长度为147bp的DNA序列，将这些片段克隆到重组质粒中，大量扩增回收9条标记biotin分子的目的序列。同时分别表达纯化了组蛋白H2A、H2B、H3和H4，复性后装配形成组蛋白八聚体结构。利用盐透析方法将9条DNA序列在体外组装形成核小体结构，经biotin标记检测后计算了反应过程的吉布斯自由能，对比了9条目的序列形成核小体的亲和力大小。研究发现，9条序列中有5条序列与理论预测完全符合，4条序列与理论预测不完全一致。实验结果与该算法预测的核小体定位结果基本一致，表明该理论模型能够有效预测酿酒酵母基因组核小体占据水平。

英文摘要：

Nucleosome positioning plays a very key role in gene expression in eukaryotes. A position-corre- lation scoring function algorithm on the bias of k-mer frequency in linker sequences to distinguish nucleo- some vs linker sequences has been constructed, which achieved a good performance in the Saccharomyces cerevisiae genome. According to the calculation results using this model with 1 bp a step, 147 bp a window in the genome, the sequences with top, medium and lower abilities to form nucleosome were selected from each of Chromosome 1, 3, and 14 of Saccharomyces cerevisiae respectively. In order to abundantly amplify the biotin labeling sequences, these target sequences were integrated to recombinant plasmids by molecular cloning methods. Meanwhile, the H2A, H2B, H3 and H4 proteins were expressed and purified, and constitut- ed to form octamer. Then Gradient salt dialysis was used to assemble nucleosome structure on the sequences in vitro. And Gibbs free energy in the reaction was calculated after biotin labeling test, which was employed to compare the affinities of the target sequences to octamer. Consequently, it shows that results of five se- quences are consistent with the theoretical prediction, which suggests that this model can be used to predict nucleosome occupancy in Saccharomyces cerevisiae genome efficiently.