| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
First published online May 27, 2005; 10.1104/pp.104.055673 Plant Physiology 138:923-934 (2005) © 2005 American Society of Plant Biologists Local Coexpression Domains of Two to Four Genes in the Genome of Arabidopsis1,[w]Applied Bioinformatics, Plant Research International, NL–6700 AA Wageningen, The Netherlands (X.-Y.R., M.W.E.J.F., J.-P.N.); Laboratory of Molecular Biology, Plant Sciences Group, Wageningen University and Research Centre, NL–6703 HA Wageningen, The Netherlands (X.-Y.R., W.J.S.); and Centre for BioSystems Genomics, NL–6708 PB, Wageningen, The Netherlands (W.J.S., J.-P.N.)
Expression of genes in eukaryotic genomes is known to cluster, but cluster size is generally loosely defined and highly variable. We have here taken a very strict definition of cluster as sets of physically adjacent genes that are highly coexpressed and form so-called local coexpression domains. The Arabidopsis (Arabidopsis thaliana) genome was analyzed for the presence of such local coexpression domains to elucidate its functional characteristics. We used expression data sets that cover different experimental conditions, organs, tissues, and cells from the Massively Parallel Signature Sequencing repository and microarray data (Affymetrix) from a detailed root analysis. With these expression data, we identified 689 and 1,481 local coexpression domains, respectively, consisting of two to four genes with a pairwise Pearson's correlation coefficient larger than 0.7. This number is approximately 1- to 5-fold higher than the numbers expected by chance. A small (5%–10%) yet significant fraction of genes in the Arabidopsis genome is therefore organized into local coexpression domains. These local coexpression domains were distributed over the genome. Genes in such local domains were for the major part not categorized in the same functional category (GOslim). Neither tandemly duplicated genes nor shared promoter sequence nor gene distance explained the occurrence of coexpression of genes in such chromosomal domains. This indicates that other parameters in genes or gene positions are important to establish coexpression in local domains of Arabidopsis chromosomes.
The combination of DNA sequence and expression data has revealed the existence of chromosomal domains of similarly expressed genes in several genomes, such as in yeast (Saccharomyces cerevisiae; Cohen et al., 2000  ), fly (Drosophila melanogaster; Spellman and Rubin, 2002), worm (Caenohabditis elegans; Roy et al., 2002; Lercher et al., 2003), human (Homo sapiens; Caron et al., 2001; Lercher et al., 2002; Versteeg et al., 2003), and more recently in the genome of the plant Arabidopsis (Arabidopsis thaliana; Birnbaum et al., 2003; Williams and Bowles, 2004). These analyses have focused on coexpression (Cohen et al., 2000; Spellman and Rubin, 2002; Lercher et al., 2003; Williams and Bowles, 2004), high expression (Caron et al., 2001), and so-called localized expression domains (Birnbaum et al., 2003), defined as spatial and/or temporal chromosomal domains of coordinated induction and repression in gene expression. Chromosomal domains (or clusters or regions) of similarly expressed (or coexpressed or coregulated or correlated) genes have been identified using sliding windows of either a given sequence length (number of nucleotides; Lercher et al., 2002) or of a given number of genes (Spellman and Rubin, 2002; Lercher et al., 2003; Williams and Bowles, 2004). Major experimental differences exist in the size of the window used for analysis and therefore the fraction of the genome evaluated as chromosomal domain. To determine the similarity between different expression profiles, the Pearson's correlation coefficient (R) was used and the average of all pairwise Rs over the expression values across experiments or tissues was evaluated over all windows and chromosomes (Cohen et al., 2000; Spellman and Rubin, 2002; Lercher et al., 2003).
In such genome-wide analyses, four different types of gene organization may account for high coexpression without giving evidence for the presence of chromosomal domains. These four types are: (1) overlapping genes (Cohen et al., 2000
The molecular mechanisms responsible for coordinated expression of neighboring genes are not well understood (Hurst et al., 2004
We here present the identification and analysis of local coexpression domains in the Arabidopsis genome. Local coexpression domains are here defined as chromosomal regions where physically adjacent genes have high correlated expression across all experiments. This definition focuses on the behavior of neighboring genes. Using the Munich Information Center for Protein Sequences (MIPS) Arabidopsis genome annotation (Schoof et al., 2002
Chromosomal Coexpression Maps Reveal Local Coexpression Domains
The combination of the MIPS annotation of the Arabidopsis genome with the available MPSS and MA data resulted in a collection of 16,144 gene pairs with MPSS expression values and 18,443 pairs with MA expression values that could be analyzed. A more detailed description of the data sets generated is given in Table I and in "Materials and Methods." For visualization purposes, the overall whole-genome coexpression data were plotted in chromosomal coexpression maps as introduced by Cohen et al. (2000)
The number of local coexpression domains in the Arabidopsis genome and the number of genes involved were calculated. Two genes were considered to be adjacent, so present in a local coexpression domain, if their rank IDs (see "Materials and Methods") were consecutive with a difference of one and their pairwise R exceeded 0.7. Notably tandemly duplicated genes are known to influence coexpression statistics (Zhu, 2003 ; Hurst et al., 2004). Therefore, the occurrence of tandemly duplicated pairs was determined with pairwise protein BLAST using a cutoff of E < 2 x 10–1 (Lercher et al., 2002, 2003; Fukuoka et al., 2004; Williams and Bowles, 2004). This criterion has a false error rate of about 10% (Lercher et al., 2002; Williams and Bowles, 2004). In both the MPSS and MA expression data sets, only about 12% of all adjacent pairs (1,928 for MPSS and 2,278 for MA) are tandemly duplicated (Table I), of which only 11% to 14% are coexpressed (216 for MPSS and 319 for MA). This implies that in either expression data set only approximately 20% of the coexpressed pairs consist of tandemly duplicated genes. Only a minority of 11% to 14% of all tandemly duplicated gene pairs in the Arabidopsis genome are coexpressed (with R > 0.7), reflecting gene divergence after duplication (Williams and Bowles, 2004). As about 5% to 9% from all adjacent pairs excluding the tandemly duplicated pairs (689 for MPSS and 1,481 for MA) are coexpressed, a tandemly duplicated pair is about 2-fold (that is, 11%–14% relative to 5%–9%) more likely to be coexpressed than a nontandemly duplicated adjacent pair. Further analyses of the subpopulation of tandemly duplicated gene pairs do not indicate that a particular transcriptional orientation of the tandemly duplicated genes has a significantly higher inclination to be coexpressed (data not shown). In subsequent analyses, the subpopulations were evaluated with and without tandemly duplicated genes. The results are summarized in Table II. Depending on the expression data set considered, 5% to 9% of all nonduplicated gene pairs consist of coexpressed neighboring pairs (689 for MPSS and 1,481 for MA). These pairs tend to be spread throughout the genome (Fig. 2; MPSS data). The MA data set gave similar results (data not shown). Only 58 coexpressed pairs were common between the MPSS and MA sets out of 11,144 total common pairs (excluding tandemly duplicated pairs). These common coexpressed pairs are also widespread throughout the genome (Fig. 2).
In addition to the number of coexpressed gene pairs (duplets), the number of coexpressed triplets, quadruplets, and pentaplets in the Arabidopsis genome was determined (Table II; Fig. 2) using the strict criterion of highly correlated expression (R > 0.7) of all members in a multiplet. Triplet and quadruplet coexpression domains were considerably rarer (Table II), whereas coexpressed pentaplets did not occur in either the MPSS or the MA data set. To evaluate the significance of the observed numbers of the local coexpression domains in Arabidopsis, these numbers were compared with the numbers of pairs, triplets, and quadruplets obtained from randomized sets using the cumulative binomial distribution (Cohen et al., 2000 ). Such comparisons indicated that in all cases examined, local coexpression domains ranging from two to four genes occur in the Arabidopsis genome significantly more often than expected by chance alone (Table II). Excluding tandem duplicates, coexpressed adjacent genes also occurred significantly more often than in random sets (Table II). Tandem duplicates are therefore not an important explanation for the occurrence of local coexpression domains in the Arabidopsis genome.
Apart from tandem duplications, gene orientation and gene distance could also explain the occurrence of local coexpression domains. If promoter sharing is an important mechanism for coexpression in the Arabidopsis genome, divergently transcribed gene pairs (
The closer two genes are, the higher the likelihood may be that their promoters influence each other, irrespective of gene orientation. Gene distance is here defined as the distance in nucleotides from the 5' start ATG of one gene to the 5' start ATG of the next gene. Thus defined, gene distance covers the distance between one coding sequence and promoter region for tandemly transcribed genes. For the other two gene orientations, this definition of gene distance results in the length of either the promoter sequence (in the case of divergently transcribed genes) or the inclusion of two coding regions (in the case of convergently transcribed genes). As a consequence, gene distance will favor divergently transcribed genes in the shorter distances and tandemly and convergently transcribed pairs in the larger distances. The subsequent distance analysis also distinguishes between gene orientations. All adjacent gene pairs (excluding the tandemly duplicated gene pairs) were sorted by gene distance and divided into consecutive bins of 1,000 pairs. This way, any influence of unequal numbers of pairs in distance bins was prevented. For each 1,000-pair bin, the number of tandemly, divergently, and convergently transcribed adjacent pairs was counted and plotted against the average gene distance (Fig. 4A). The average distance was calculated by averaging the gene distance of all pairs in each 1,000-pair bin. In the same way, the number of coexpressed adjacent pairs in each orientation group was counted and plotted (Fig. 4B). In both cases, it can be concluded that in the shorter gene distance classes divergently transcribed pairs occur more often than tandemly and convergently transcribed pairs, whereas in the larger gene distance classes divergently transcribed genes occur less often. Interestingly, coexpressed adjacent genes in any orientation could occur even at a gene distance as large as 12 kb. To be able to compare the relative occurrence of the orientation groups among all the 1,000-pair bins, the fraction of coexpressed pairs was plotted for each orientation group (Fig. 4C). The fractions of coexpressed pairs stay similar among three orientation groups and also stay similar over large gene distance range. Basically identical results were obtained for the MA data set (Fig. 4, D–F). Similar results were obtained using distance bins of 1-kb intervals or intergenic distances (data not shown). These results show that over a large gene distance range, the relative fraction of coexpressed pairs does not depend on gene distance, irrespective of gene orientation. Therefore, also gene distance and/or gene orientation are not important explanations for the occurrence of local coexpression domains in the Arabidopsis genome.
Genes in Local Coexpression Domains Scatter over Functional Categories
Having estimated the number of local coexpression domains in the Arabidopsis genome, the nature of the genes involved in such chromosomal domains was analyzed. The Arabidopsis Information Resource (TAIR)'s Gene Ontology (GO) using the high-level ontology terms known as GOslims developed for plants (Berardini et al., 2004
For different organisms, it has been demonstrated that appreciable numbers of genes in a genome occur in clusters characterized by correlated expression. Averaging coexpression over size-based or gene number-based windows showed that about 20% of the Drosophila genome resides in coexpression clusters (Spellman and Rubin, 2002 ). Within the Arabidopsis genome, such window-based coexpression clusters may consist of up to 20 genes (Birnbaum et al., 2003; Zhu, 2003; Williams and Bowles, 2004), while some evidence from quantitative trait loci studies suggested that clusters may be much larger (Khavkin and Coe, 1997). These data support the notion of higher-level genome organization that may range over distances up to several megabases (Hurst et al., 2004). Yet the concept of large coexpression clusters in such studies is based on a loose definition of the term cluster or chromosomal domain and associated terms such as neighboring. The process of summing and averaging may obscure local effects and underrate the presence and/or role of individual genes with different expression levels or expression patterns in large clusters. Therefore, it is perhaps not surprising that coexpression clusters are often associated with the activity of housekeeping (Lercher et al., 2002, 2003; Roy et al., 2002) or highly expressed (Caron et al., 2001; Versteeg et al., 2003) genes. Previous experience with transgene expression data indicated that the particular position of a single gene in a genome affects the expression of that gene. Depending on chromosomal context, two physically neighboring transgenes could be made to show correlated expression (Mlynarova et al., 2002). Therefore, we have here taken a very rigorous approach to the concept of cluster and analyzed the coexpression characteristics of genes that are physically adjacent in the genome according to genome annotation data.
Whole-genome chromosomal coexpression maps indicate the existence of numerous cases of local coexpression (Fig. 1), as was also shown in yeast (Cohen et al., 2000
From all nontandemly duplicated pairs in the MPSS data set, 4.9% shows coexpression. They are distributed over the whole genome (Fig. 2). Although this is a low percentage, randomization assays indicate that the number is significantly higher than to be expected by chance alone (Table II). There is a small yet significant fraction of the Arabidopsis genome that shows correlated expression between neighboring genes. Enlarging such local clusters by looking for series of consecutive genes that are correlated in all pairwise combinations reveals that there are few areas in the Arabidopsis genome that consist of more than two (up to four) genes (Table II; Fig. 2) with highly correlated expression. The size of these local coexpression domains is in agreement with local cluster sizes observed in yeast (Cohen et al., 2000 Over the whole genome, the two expression data sets show areas that have different coexpression patterns (Fig. 1) and in total only 58 coexpressed pairs were shared between both data sets (Fig. 2). These differences in coexpression and low number of shared pairs are likely to reflect the biological differences between the data sets. The MA data are well defined root cells and tissues, while the MPSS data concern more broad tissues and organs. Such biological differences will influence correlations in gene activity. Any expression data set will present a fixed average of expression over the sampled cells, tissues, organs, and experiments that should be taken into account when comparing such data sets.
To understand the possible causes for coexpression, the role of shared promoters and/or short gene distances was analyzed. The population of divergently transcribed genes does not contain a higher proportion of coexpressed genes compared to tandemly or convergently transcribed genes (Fig. 3; Table III). Promoter sharing is therefore not a likely explanation for the presence of local coexpression domains in the Arabidopsis genome, unlike the situation in the yeast genome (Cohen et al., 2000
Previous studies suggested that clustering of functionally related genes may occur in all metazoans (including yeast, fly, worm, and human; Cohen et al., 2000
In our coexpression analyses of expression data, different libraries from either MPSS or MA data were combined irrespective of the biological characteristics of the material assayed. Therefore, the analyses have revealed the gene pairs that show stringent coexpression under a range of different (biological) conditions, cells, and/or tissue types. Combining more and different data sets, such as the data in various Arabidopsis expression repositories now available at TAIR (Rhee et al., 2003
True neighboring pairs can form local coexpression domains of two to four genes irrespective of gene orientation or gene distance. Having eliminated such configuration factors, a role of either the gene sequence itself or the DNA sequences surrounding these genes is suggested. In the transgenic situation, it was shown before that the expression of two unrelated genes became correlated when their surrounding DNA was supplied with a chromatin boundary element (Mlynarova et al., 2002
Defining local coexpression domains as genome areas with physically neighboring genes showing tight coexpression, we have here shown that the Arabidopsis genome contains a small yet significant number of coexpression domains that range from two to four genes. Neither tandemly duplicated genes nor divergently transcribed promoter regions nor short gene distances explain such local coexpression of adjacent genes. Either gene sequence or the surrounding DNA sequences are of importance for the coexpression pattern of such neighboring genes. Our study and the further unraveling of the relationships between local and global coexpression domains in relationship to surrounding DNA, gene regulation, and chromosome structure will help to gain understanding of the molecular mechanisms that establish local chromosomal domains of genes with high coexpression characteristics.
Data Retrieval and Processing
The Arabidopsis (Arabidopsis thaliana) genome annotation from the March 2003 version of MIPS (Schoof et al., 2002
The MA expression data were obtained from the on-line supplementary material of a Science article (Birnbaum et al., 2003 In case of physically overlapping genes in either data set, the smaller one of the overlapping genes was removed from the data set, by which both gene and rank ID orders were maintained. For this reason, 39 and 34 genes were removed from the MPSS and the MA data set, respectively. The resulting data sets used for analysis consisted of 20,002 genes with MPSS expression data and 21,906 genes with MA expression data. Two genes were considered to be adjacent when their rank IDs were consecutive with a difference of one and when the genome sequence had no long stretches of N's in-between. In six cases (three on Chr1 and three on Chr2), the genome sequence was interrupted by a stretch of 60 or 120 N's. These genes were included in the subsequent analyses. Adjacent genes were considered per chromosome. With these criteria, 16,144 adjacent gene pairs were identified in the MPSS data set. These pairs comprised 19,151 genes with expression values. A total of 851 (that is, the difference between 20,002 and 19,151) genes in the MPSS data set had no neighbors with expression data. In the MA data set, 18,443 adjacent gene pairs in the MA data set were identified, comprising 21,255 genes and 651 isolated genes without expressed neighboring genes (Table I).
Tandemly duplicated genes were identified by local pairwise protein BLAST (BLASTP 2.2.6 [April 9, 2003]; Altschul et al., 1997
R was calculated between all adjacent pairs (duplets) of genes using the expression data from all libraries available in each data set. If R was higher than 0.7, the gene pair concerned was considered to be coexpressed. The value of R > 0.7 is generally considered a rule-of-thumb threshold (for example, see http://bbc.botany.utoronto.ca/affydb/BAR_instructions) and is used in various analyses (Cohen et al., 2000
The number of coexpressed adjacent pairs was counted. To evaluate the statistical significance of these numbers, they were compared with the number of coexpressed pairs from 100 randomizations of the population of expressed genes using the cumulative binomial distribution (Cohen et al., 2000
Adjacent gene pairs were separated into tandemly, divergently, and convergently transcribed pairs according to their relative direction of transcription. The number of coexpressed pairs in each orientation group was expressed as percentage relative to the total number of adjacent pairs in that group. Random pairs were made by randomly picking two nonadjacent genes from the list of expressed genes represented in pairs, analyzed for their orientation, and compared with the real genome using a variant of the two-sample t test for proportions for determining the significance of a difference between two population proportions (Ott and Longnecker, 2001 For gene distance, the length in nucleotides from the 5' start of one gene to the 5' start of the next gene was used. The data sets excluding the tandemly duplicated gene pairs were analyzed. This way, there were 14,216 pairs in the MPSS-tandemly duplicated data set and 16,165 pairs in MA-tandemly duplicated data set. For each data set, gene pairs were sorted based on gene distance and bins of 1,000 pairs were taken and analyzed, excluding the last bin with less than 1,000 pairs. Per 1,000-pair bin, gene distance was calculated as the average over all 1,000 pairs. In total, 14 bins of 1,000 pairs for the MPSS-tandemly duplicated data set and 16 bins of 1,000 pairs in MA data set were analyzed. Within each 1,000-pair bin, the numbers of tandem, divergent, and convergent pairs were determined, as well as the numbers of coexpressed pairs within each orientation group. To be able to compare bins, the fraction of coexpressed pairs relative to the total number of pairs in each orientation group in each bin was calculated.
TAIR's GOslim, the GO developed for plants (Berardini et al., 2004 Upon request, all novel materials described in this publication will be made available in a timely manner for noncommercial research purposes, subject to the requisite permission from any third-party owners of all or parts of the material. Obtaining any permissions will be the responsibility of the requestor.
We thank Nayelli Marsch-Martinez and Harrie Verhoeven (Bioscience, Plant Research International, Wageningen, The Netherlands), Blake Meyers (University of Delaware), Jack Leunissen (Bioinformatics, Wageningen University, The Netherlands), and Huanming Yang (Beijing Genomics Institute, Beijing) as well as Roeland van Ham, Paulien Adamse, Oscar Vorst, and other members of the Applied Bioinformatics cluster of Plant Research International for helpful comments, suggestions, data, and discussions. Received October 26, 2004; returned for revision February 27, 2005; accepted February 28, 2005.
1 This work was supported by the Dutch Organization for Scientific Research (in the framework of the project Wageningen Phytoinformatics: The Added Value from Plants) and by the Centre of BioSystems Genomics, which is part of the Netherlands Genomics Initiative/Netherlands Organization for Scientific Research. 
[w] The online version of this article contains Web-only data. Article, publication date, and citation information can be found at www.plantphysiol.org/cgi/doi/10.1104/pp.104.055673. * Corresponding author; e-mail janpeter.nap{at}wur.nl; fax 31–317–418094.
Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402
Bell AC, West AG, Felsenfeld G (2001) Insulators and boundaries: versatile regulatory elements in the eukaryotic genome. Science 291: 447–450
Berardini TZ, Mundodi S, Reiser L, Huala E, Garcia-Hernandez M, Zhang PF, Mueller LA, Yoon J, Doyle A, Lander G, et al (2004) Functional annotation of the Arabidopsis genome using controlled vocabularies. Plant Physiol 135: 745–755
Birnbaum K, Shasha DE, Wang JY, Jung JW, Lambert GM, Galbraith DW, Benfey PN (2003) A gene expression map of the Arabidopsis root. Science 302: 1956–1960
Blanc G, Wolfe KH (2004) Functional divergence of duplicated genes formed by polyploidy during Arabidopsis evolution. Plant Cell 16: 1679–1691 Boulikas T (1995) Chromatin domains and prediction of MAR sequences. Int Rev Cytol 162A: 279–388[Web of Science][Medline]
Caron H, van Schaik B, van der Mee M, Baas F, Riggins G, van Sluis P, Hermus MC, van Asperen R, Boon K, Voute PA, et al (2001) The human transcriptome map: clustering of highly expressed genes in chromosomal domains. Science 291: 1289–1292 Cohen BA, Mitra RD, Hughes JD, Church GM (2000) A computational analysis of whole-genome expression data reveals chromosomal domains of gene expression. Nat Genet 26: 183–186[CrossRef][Web of Science][Medline]
Edgar R, Domrachev M, Lash AE (2002) Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 30: 207–210 Fukuoka Y, Inaoka H, Kohane IS (2004) Inter-species differences of co-expression of neighboring genes in eukaryotic genomes. BMC Genomics 5: 4[CrossRef][Medline]
Gollub J, Ball CA, Binkley G, Demeter J, Finkelstein DB, Hebert JM, Hernandez-Boussard T, Jin H, Kaloper M, Matese JC, et al (2003) The Stanford Microarray Database: data access and quality assessment tools. Nucleic Acids Res 31: 94–96 Hurst LD, Pal C, Lercher MJ (2004) The evolutionary dynamics of eukaryotic gene order. Nat Rev Genet 5: 299–310[CrossRef][Web of Science][Medline] Khavkin E, Coe E (1997) Mapped genomic locations for developmental functions and QTLs reflect concerted groups in maize (Zea mays L.). Theor Appl Genet 95: 343–352[CrossRef]
Lee HK, Hsu AK, Sajdak J, Qin J, Pavlidis P (2004) Co-expression analysis of human genes across many micro array data sets. Genome Res 14: 1085–1094
Lercher MJ, Blumenthal T, Hurst LD (2003) Co-expression of neighboring genes in Caenorhabditis elegans is mostly due to operons and duplicate genes. Genome Res 13: 238–243 Lercher MJ, Urrutia AO, Hurst LD, Cohen BA, Mitra RD, Hughes JD, Church GM (2002) Clustering of housekeeping genes provides a unified model of gene order in the human genome. Nat Genet 31: 180–183[CrossRef][Web of Science][Medline]
Meyers BC, Lee DK, Vu TH, Tej SS, Edberg SB, Matvienko M, Tindell LD (2004) Arabidopsis MPSS. An online resource for quantitative expression analysis. Plant Physiol 135: 801–813 Mlynarova L, Jansen RC, Conner AJ, Stiekema WJ, Nap JP (1995) The MAR-mediated reduction in position effect can be uncoupled from copy number-dependent expression in transgenic plants. Plant Cell 7: 599–609[Abstract] Mlynarova L, Loonen A, Heldens J, Jansen RC, Keizer P, Stiekema WJ, Nap JP (1994) Reduced position effect in mature transgenic plants conferred by the chicken lysozyme matrix-associated region. Plant Cell 6: 417–426[Abstract]
Mlynarova L, Loonen A, Mietkiewska E, Jansen RC, Nap JP (2002) Assembly of two transgenes in an artificial chromatin domain gives highly coordinated expression in tobacco. Genetics 160: 727–740 Obayashi T, Okegawa T, Sasaki-Sekimoto Y, Shimada H, Masuda T, Asamizu E, Nakamura Y, Shibata D, Tabata S, Takamiya K, et al (2004) Distinctive features of plant organs characterized by global analysis of gene expression in Arabidopsis. DNA Res 11: 11–25[Abstract] Ott RL, Longnecker M (2001) An Introduction to Statistical Methods and Data Analysis, Chapter 10, Categorical Data, Ed 5. Duxbury, Pacific Grove, CA, pp 482–485
Rhee SY, Beavis W, Berardini TZ, Chen G, Dixon D, Doyle A, Garcia-Hernandez M, Huala E, Lander G, Montoya M, et al (2003) The Arabidopsis Information Resource (TAIR): a model organism database providing a centralized, curated gateway to Arabidopsis biology, research materials and community. Nucleic Acids Res 31: 224–228 Roy PJ, Stuart JM, Lund J, Kim SK (2002) Chromosomal clustering of muscle-expressed genes in Caenorhabditis elegans. Nature 418: 975–979[Medline]
Schoof H, Zaccaria P, Gundlach H, Lemcke K, Rudd S, Kolesov G, Arnold R, Mewes HW, Mayer KF (2002) MIPS Arabidopsis thaliana Database (MAtDB): an integrated biological knowledge resource based on the first complete plant genome. Nucleic Acids Res 30: 91–93
Smith LI (2002) A Tutorial on Principal Components Analysis. http://kybele.psych.cornell.edu/ Spellman PT, Rubin GM (2002) Evidence for large domains of similarly expressed genes in the Drosophila genome. J Biol 1: 5[CrossRef][Medline]
Versteeg R, van Schaik BD, van Batenburg MF, Roos M, Monajemi R, Caron H, Bussemaker HJ, van Kampen AH (2003) The human transcriptome map reveals extremes in gene density, intron length, GC content, and repeat pattern for domains of highly and weakly expressed genes. Genome Res 13: 1998–2004
Williams EJ, Bowles DJ (2004) Co-expression of neighboring genes in the genome of Arabidopsis thaliana. Genome Res 14: 1060–1067
Wu CH, Huang H, Arminski L, Castro-Alvear J, Chen Y, Hu ZZ, Ledley RS, Lewis KC, Mewes HW, Orcutt BC, et al (2002) The Protein Information Resource: an integrated public resource of functional annotation of proteins. Nucleic Acids Res 30: 35–37 Zhu T (2003) Global analysis of gene expression using GeneChip microarrays. Curr Opin Plant Biol 6: 418–425[CrossRef][Web of Science][Medline]
Zimmermann P, Hirsch-Hoffmann M, Hennig L, Gruissem W (2004) GENEVESTIGATOR. Arabidopsis microarray database and analysis toolbox. Plant Physiol 136: 2621–2632 This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ASPB Publications | PLANT PHYSIOLOGY® | THE PLANT CELL | |
|---|---|---|---|
TOP
ABSTRACT
RESULTS
gene A gene B
) should be overrepresented in the subpopulation of coexpressed pairs, compared to coexpressed pairs that are tandemly (gene A
(p1*(1 – p1)/n1 + p2*(1 – p2)/n2), with p1 and p2 the two sample proportions, n1 and n2 the two sample sizes, under the condition that n1*p1, n1*(1 – p1), n2*p2 and n2*(1 – p2) are all larger than 5. When |z| > 2.575, the two sample proportions are considered to be significantly different at the 1% level (P < 0.01). The z value is converted to a P value using standard normal tables. 
edelman/Psych-465-Spring-2003/PCA-tutorial.pdf
