基因表达聚类分析技术的现状与发展

随着多个生物基因组测序的完成、DNA芯片技术的广泛应用,基因表达数据分析已成为后基因组时代的研究热点.聚类分析能将功能相关的基因按表达谱的相似程度归纳成类,有助于对未知功能的基因进行研究,是目前基因表达分析研究的主要计算技术之一.已有多种聚类分析算法用于基因表达数据分析,各种算法因其着眼点、原理等方面的差异,而各有其优缺点.如何对各种聚类算法的有效性进行分析、并开发新型的、适合于基因表达数据分析的方法已是当务之急.     

Mathematical modeling of the temperature field distribution in insect winter clusters

A model is proposed for the distribution of temperature fields and mechanisms of thermoregulation in an insect (honeybee) cluster self-organizing for protection against long-term cooling. The winter cluster is an ordered system of heat generation, accumulation, and dissipation. Estimates are obtained for the influence of ambient temperature that promotes clustering and of the number of clustered insects on the efficiency of cold protection. Described are the dynamic links between the cluster structures that ensure minimal energy expenditures in metabolism and thermoregulation; this allows the cluster to withstand the long unfavorable exposure to temperature varying far beyond the limits of viability for an individual insect.     

Genetic diversity and dispersal of Phragmites australis in a small river system

Even though the reed, Phragmites australis, is an extensively studied wetland species, little is known about reproduction and dispersal modes within and among reed populations at the scale of small river systems. Using microsatellite analysis of 189 individuals from three adjacent river catchments in the Czech Republic, we elucidated the role of the river corridors in the dispersal of P. australis. Using Bayesian clustering of individuals, we found that 19% of clusters were distributed only along one river, which implied dispersal by water (or by wind) along river corridors, whereas 38% of clusters were widely distributed and were likely the product of wind long-distance dispersal among rivers. Intensive exchange of propagules among river systems is further demonstrated by only 6% of total variance being attributed to the variance among rivers in the AMOVA-analysis. Spatial autocorrelation analysis revealed a decreasing pattern up to 5–10 km and no clear pattern over longer distances. This gives an evidence for pollen and seed dispersal at short distances (up to 1 km), whereas most likely only seed dispersal at longer distances up to 10 km. We found five multilocus genotypes distributed in two different populations. The distances between populations with the same genotype ranged from 0.5 to 10.8 km. This can be interpreted as long-distance vegetative dispersal.     

Application of a clustering-based peak alignment algorithm to analyze various DNA fingerprinting data

DNA fingerprinting analysis such as amplified ribosomal DNA restriction analysis (ARDRA), repetitive extragenic palindromic PCR (rep-PCR), ribosomal intergenic spacer analysis (RISA), and denaturing gradient gel electrophoresis (DGGE) are frequently used in various fields of microbiology. The major difficulty in DNA fingerprinting data analysis is the alignment of multiple peak sets. We report here an R program for a clustering-based peak alignment algorithm, and its application to analyze various DNA fingerprinting data, such as ARDRA, rep-PCR, RISA, and DGGE data. The results obtained by our clustering algorithm and by BioNumerics software showed high similarity. Since several R packages have been established to statistically analyze various biological data, the distance matrix obtained by our R program can be used for subsequent statistical analyses, some of which were not previously performed but are useful in DNA fingerprinting studies.     

Molecular Diversity and Assessment of Biological Characteristics of Greek Colletotrichum lindemuthianum Populations

Colletotrichum lindemuthianum  isolates collected in Greece were characterized by the temperature effect on their biological characteristics (mycelial growth, sporulation and spore germinability) and by molecular diversity revealed by RAM and ERIC–BOX PCR analysis. The temperature effect on the assessed biological characteristics resulted in a similar classification according to the origin and virulence patterns of isolates. Colletotrichum lindemuthianum isolates originating from the areas of Nevrokopi and Vrodou showed better adaptation at the lower temperatures exposure (12 and 18°C) compared to isolates originating from the Municipality of Hrisoupolis, which showed better adaptation at the highest temperature tested (24°C). Molecular diversity was detected using RAM and ERIC–BOX PCR primers. Both methods revealed, in a similar way ( r  = 0.58, P = 0.05), two main clusters of isolates, in agreement with previous findings using RAPD and RFLP analysis. The majority of the tested isolates were grouped in the same main cluster (29 out of 35 Greek isolates for both methods) underlying high levels of genotypic similarities between Greek populations of C. lindemuthianum . This study, an extension of previous research, provides further information on population diversity of C. lindemuthianum required for developing more efficient control strategies of bean anthracnose disease.     

Reduced genetic diversity, increased isolation and multiple introductions of invasive giant hogweed in the western Swiss Alps

The giant hogweed ( Heracleum mantegazzianum ) has successfully invaded 19 European countries as well as parts of North America. It has become a problematic species due to its ability to displace native flora and to cause public health hazards. Applying population genetics to species invasion can help reconstruct invasion history and may promote more efficient management practice. We thus analysed levels of genetic variation and population genetic structure of H. mantegazzianum in an invaded area of the western Swiss Alps as well as in its native range (the Caucasus), using eight nuclear microsatellite loci together with plastid DNA markers and sequences. On both nuclear and plastid genomes, native populations exhibited significantly higher levels of genetic diversity compared to invasive populations, confirming an important founder event during the invasion process. Invasive populations were also significantly more differentiated than native populations. Bayesian clustering analysis identified five clusters in the native range that corresponded to geographically and ecologically separated groups. In the invaded range, 10 clusters occurred. Unlike native populations, invasive clusters were characterized by a mosaic pattern in the landscape, possibly caused by anthropogenic dispersal of the species via roads and direct collection for ornamental purposes. Lastly, our analyses revealed four main divergent groups in the western Swiss Alps, likely as a consequence of multiple independent establishments of H. mantegazzianum .     

Ceramide-enriched membrane domains—Structure and function

Membrane lipids seem to be organized and not randomly distributed in the cell membrane. In particular, sphingolipids seem to interact with cholesterol in the outer leaflet of the cell membrane resulting in the formation of distinct membrane domains, i.e. rafts. The generation of ceramide within rafts alters their biophysical properties and results in the formation of large ceramide-enriched membrane platforms. These platforms serve to cluster receptor molecules and to organize intracellular signalling molecules to facilitate signal transduction via a receptor upon stimulation. Thus, ceramide-enriched membrane domains amplify not only receptor-, but also stress-mediated signalling events. Although many receptors cluster, the molecular mechanisms mediating this important and general event in signal transduction need to be identified.     

Genetic diversity and gene flow in a Caribbean tree <Emphasis Type="Italic">Pterocarpus officinalis</Emphasis> Jacq.: a study based on chloroplast and nuclear microsatellites

We analysed the molecular diversity of Pterocarpus officinalis, a tree species distributed in Caribbean islands, South and Central America to quantify the genetic variation within island, to assess the pattern of differentiation and infer levels of gene flow; with the overall goal of defining a strategy of conservation. Two hundred two individuals of 9 populations were analysed using three chloroplast and six nuclear microsatellite markers. The observed heterozygosity varied markedly among the populations for nuclear (H _Onuc = 0.20–0.50) and chloroplast microsatellites (H _cp = 0.22–0.68). The continental population from French Guyana showed a higher value of H _Onuc than island populations, and the differences were significant in some cases. The fixation index F _IS ranged from −0.043 to 0.368; a significant heterozygote deficit was detected in 7 populations. The heterozygosity excess method suggested that two populations in Guadeloupe have undergone a recent bottleneck. Global and pairwise F _ST were high for both nuclear (F _STnuc = 0.29) and chloroplast microsatellites (F _STcp = 0.58). The neighbour-joining tree based on both markers, presented a differentiation pattern that can be explained by the seed dispersal by flotation and marine stream. The comparison of Bayesian approach and the method based on allelic frequency demonstrate a very limited number of migrants between populations.     

基于信息论k—modes聚类法的基因表达数据分析

k-均值聚类算法是一种广泛应用于基因表达数据聚类分析中的迭代变换算法,它通常用距离法来表示基因间的关系,但不能有效的反应基因间的相互依赖的关系。为此,提出基于信息论的k-modes聚类算法,克服了以上缺点。另外,还引入了伪F统计量,一方面,可以对空间中有部分重叠的点进行有效的分类;另一方面,可以给出最佳聚类数目,从而弥补了k-modes聚类法的不足。使其成为一种非常有效的算法,从而达到较优的聚类效果。