基因芯片癌与癌旁组织分析中药品混合对结果的影响

基因芯片技术已经广泛地应用到生物科学与医学研究的多个领域中.由于实验费用的制约,在使用芯片数目较少的情况下,一些研究者倾向于把经过同一种处理的多个个体的RNA进行混合,希望能消除个体差异.但应用一个研究肿瘤的具体例子,比较单个个体与多个个体的RNA混合后得到的基因芯片的结果表明,多个样品的RNA混合后得到的结果和单个样品得到的结果,无论从正确率、假阳性、假阴性来比较,在统计学上都没有明显差异,因此：结合多个因素考虑,建议进行基因芯片分析的RNA样品不用进行混合.     

滚环DNA扩增的原理、应用和展望

滚环DNA扩增(Rolling circle DNA amplification,RCA)是一种等温信号扩增方法,其线性扩增倍数为10^5,指数化扩增能力大于10^9,产生的扩增产物连接在固相支持物(如玻片、微孔板等)表面的DNA引物或抗体上(适宜作生物芯片研究)。RCA是一种适合在芯片上(on-chip)进行信号扩增的新技术,它既能提供研究分析的敏感性和特异性,又能保持立体分析的多元性。RCA亦是一种痕量的分子检测方法,可用于极其微量的生物大分子和生物标志的检测与研究。,     

DNA芯片--世纪之交的技术革命

ＤＮＡ 芯片是以硅、玻璃等材料作固相载片,应用计算机、半导体、光化学合成等微加工技术,将数十万个寡核苷酸片段高密度集成于１ｃｍ ２ 左右的芯片上．它可以作为微反应进行ＰＣＲ、ＬＣＲ等反应,尤其是芯片上的探针列阵可用于检测人类遗传病的基因变异,加速人类基因组的研究及为临床医学开辟新路．ＤＮＡ 芯片有望成为研究生命信息的一种新的有力工具．     

基于支持向量机的肿瘤基因数据挖掘

针对两种类别的肿瘤分类问题,首先运用信噪比方法筛选出表达水平发生显著性变化的特征基因,然后采用支持向量机作为分类器进行肿瘤分类,通过对两种类别的白血病DNA微阵列数据进行计算,达到了97.1%的分类准确度.     

Chronic quercetin exposure affects fatty acid catabolism in rat lung

Dietary quercetin intake is suggested to be health promoting, but this assumption is mainly based on mechanistic studies performed in vitro. Previously, we identified rat lung as a quercetin target tissue. To assess relevant in vivo health effects of quercetin, we analyzed mechanisms of effect in rat lungs of a chronic (41 weeks) 1% quercetin diet using whole genome microarrays. We show here that fatty acid catabolism pathways, like beta-oxidation and ketogenesis, are up-regulated by the long-term quercetin intervention. Up-regulation of genes (Hmgcs2, Ech1, Acox1, Pcca, Lpl and Acaa2) was verified and confirmed by quantitative real time PCR. In addition, free fatty acid levels were decreased in rats fed the quercetin diet, confirming that quercetin affects fatty acid catabolism. This in vivo study demonstrates for the first time that fatty acid catabolism is a relevant process that is affected in rats by chronic dietary quercetin. Received 6 July 2006; received after revision 29 August 2006; accepted 28 September 2006     

Discovering the cellular-localized functional modules and modular interactions in response to liver cancer

In this paper, we firstly identify the functional modules enriched with differentially expressed genes （DEGs） and characterized by biological processes in specific cellular locations, based on gene ontology （GO） and microarray data. Then, we further define and filter disease relevant signature modules according to the ranking of the disease discriminating abilities of the pre-selected functional modules. At last, we analyze the potential way by which they cooperate towards human disease. Application of the proposed method to the analysis of a liver cancer dataset shows that, using the same false discovery rate （ FDR ） threshold, we can find more biologically meaningful and detailed processes by using the cellular localization information. Some biological evidences support the relevancy of our biological modules to the disease mechanism.     

Differential expression analysis and regulatory network reconstruction for genes associated with muscle growth and adipose deposition in obese and lean pigs

During the growth and development of skeletal muscle cells and adipose cells, the regulatory mechanism of micro-effect polygenes determines porcine meat quality, carcass characteristics and other relative quantitative traits. Obese and lean type pig breeds show obvious differences in muscle growth and adipose deposition; however, the molecular mechanism underlying this phenotypic variation remains unknown. We used pathway-focused oligo microarray studies to examine the expression changes of 140 genes associated with muscle growth and adipose deposition in longissimus dorsi muscle at six growth stages （birth, 1, 2, 3, 4 and 5 months） of Landrace （a leaner, Western breed） and Taihu pigs （a fatty, indigenous, Chinese breed）. Variance analysis （ANOVA） revealed that differences in the expression of 18 genes in Landrace pigs and three genes in Taihu pigs were very significant （FDR-adjusted permutation, P 〈 0.01） and differences for 22 genes in Landrace pigs and seven genes in Taihu pigs were significant （FDR-adjusted permutation, P 〈 0.05） among six growth stages. Clustering analysis revealed a high level of significance （FDR-adjusted, P 〈 0.01） for four gene expression patterns, in which genes that strongly up-regulated were mainly associated with the positive regulation of myofiber formation and fatty acid biogenesis and genes that strongly down-regulated were mainly associated with the inhibition of cell proliferation and positive regulation of fatty acid β-oxidation. Based on a dynamic Bayesian network （DBN） model, gene regulatory networks （GRNs） were reconstructed from time-series data for each pig breed. These two GRNs initially revealed the distinct differences in physiological and biochemical aspects of muscle growth and adipose deposition between the two pig breeds; from these results, some potential key genes could be identified. Quantitative real-time RT-PCR （QRT-PCR） was used to verify the microarray data for five modulated genes, and a good correlation between     

Gene association study with SVM, MLP and cross validation for diagnosis of diseases

Gene association study is one of the major challenges of biochip technology both for gene diagnosis where only a gene subset is responsible to some diseases, and for treatment of curse of dimensionality which occurs especially in DNA microarray datasets where there are more than thousands of genes and only a few number of experiments (samples). This paper presents a gene selection method by training linear support vector machine (SVM)/nonlinear MLP (multi-layer perceptron) classifiers and testing them with cross validation for finding a gene subset which is optimal/suboptimal for diagnosis of binary/multiple disease types. Genes are selected with linear SVM classifier for the diagnosis of each binary disease types pair and tested by leave-one-out cross validation; then, genes in the gene subset initialized by the union of them are deleted one by one by removing the gene which brings the greatest decrease of the generalization power, for samples, on the gene subset after removal, where generalization is measured by training MLPs with leave-one-out and leave-4-out cross validations. The proposed method was tested with experiments on real DNA microarray MIT data and NCI data. The result shows that it outperforms conventional SNR method in separability of the data with expression levels on selected genes. For real DNA microarray MIT/NCI data, which is composed of 7129/2308 effective genes with only 72/64 labeled samples belonging to 2/4 disease classes, only 11/6 genes are selected to be diagnostic genes. The selected genes are tested by classification of samples on these genes with SVM/MLP with leave-one-out/both leave-one-out and leave-4-out cross validations. The result of no misclassification indicates that the selected genes can be really considered as diagnostic genes for the diagnosis of the corresponding diseases.     

Hammerhead ribozyme design and application

The emerging knowledge about RNA-based enzymes has already had great impact on our concept of evolutionary history, making the ‘RNA world’ more likely. It may well have an equally important impact on the diagnostic and therapeutic practices of human and veterinary medicine in the next decade. We are not quite there yet. This review addresses the design and application of hammerhead ribozymes, two aspects of a conserved and most commonly studied and used enzymatically active entity among the RNA enzymes. The emerging picture is one of great diversity. There is at this stage no general cell model nor a clearly preferable ribozyme structure. Each and every cell line (and tissue) may be unique in that they vary with respect to structural requirements for optimal uptake, activity and stability of ribozymes. We may have seen only the tip of the iceberg when it comes to RNA-based enzymes and their roles in biology and medicine. Received 3 June 1998; received after revision 28 July 1998; accepted 28 July 1998     

用于微阵列数据癌症分类的演化硬件多分类器

针对单分类器识别率低、稳定性差的问题,提出了一种用于微阵列数据分类的演化硬件多分类器选择性集成方法.首先把经过预处理的原始训练集随机划分为训练集和验证集;然后通过对训练集的学习获得基于演化硬件的基分类器;再用验证集评价基分类器的性能,选择其中一部分较好的基分类器组成最终的分类系统;最后用独立的测试集验证系统的性能.试验结果表明,对急性白血病和结肠癌数据集的识别率分别为95.42％、88.33％,与其他的模式识别方法具有可比性;同时在识别率相当的情况下,该方法的硬件代价远低于全集成的演化硬件多分类器.