> z|ycB bjbj 4xx+
)))8a]L)TVSiii0!0!0!_SaSaSaSaSaSaS$rVY>SQ0!,!0!0!0!SiiS///0!8iiN/0!_S//6EqHi@h)%FNS0TFRY%DRY8qHqHRYH00!0!/0!0!0!0!0!SS,+0!0!0!T0!0!0!0!RY0!0!0!0!0!0!0!0!0!
: HSupplemental Material for
The Condition-Dependent Transcriptional Landscape of Burkholderia pseudomallei
Wen Fong Ooi1*, Catherine Ong2*, Tannistha Nandi1, Jason F. Kreisberg1, Hui Hoon Chua1, Guangwen Sun3, Yahua Chen4, Claudia Mueller5, Laura Conejero6, Majid Eshaghi1, Roy Moh Lik Ang1, Jianhua Liu1, Bruno W. Sobral7,8, Sunee Korbsrisate9, Yunn Hwen Gan4, Richard W. Titball5, Gregory J. Bancroft6, Eric Valade10,11, Patrick Tan1,12,13,#
1Genome Institute of Singapore, Singapore, Republic of Singapore,
2DMERI@DSO, DSO National Laboratories, Singapore, Republic of Singapore,
3School of Applied Science, Republic Polytechnic, Singapore, Republic of Singapore,
4Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Republic of Singapore,
5College of Life and Environmental Sciences, University of Exeter, Exeter, United Kingdom,
6Department of Immunology and Infection, Faculty of Infectious & Tropical Diseases, London School of Hygiene & Tropical Medicine, London, United Kingdom,
7Virginia Bioinformatics Institute at Virginia Tech, Blacksburg, United States of America,
8Current address: Nestle Institute of Health Sciences, Campus EPFL, Lausanne, Switzerland,
9Department of Immunology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand,
10Institut de Recherche Biomdicale des Armes / CRSSA, La Tronche, France,
11Ecole du Val-de-Grce, Paris, France,
12Duke-NUS Graduate Medical School, Singapore, Republic of Singapore,
13Cancer Science Institute of Singapore, National University of Singapore, Singapore, Republic of Singapore
* Equal Contribution
# Address Correspondence to HYPERLINK "mailto:tanbop@gis.a-star.edu.sg"tanbop@gis.a-star.edu.sg
Genome Institute of Singapore, 60 Biopolis Street, Genome, Singapore 138672
Telephone #: +65-64788182
Fax #: +65-64789003
The file includes:
Supplemental Methods
Identification of TARs
Experimental validation of transcripts
Enrichment analysis of silent genes
Relative expression of genes and ncRNAs
Differential expression analysis
Evaluating the significance of overlap between detectible proteins and transcripts
Optimizing the clustering granularity
Correlations between ncRNAs and clusters
Transforming the weighted coexpression network for approximating a scale-free Network
Clustering of conditions based on condition-dependent profiles
Motif confirmation by BioProspector
Computing network distance between disrupted genes in T3SS3 and other differentially expressed genes
Computing the weighted clustering coefficients
Motility assays
Bp capsule electron microscopy
Identifying in-vitro conditions showing similar expression profiles to mice lung infection
Supplemental References
SUPPLEMENTAL METHODS
Identification of TARs
We performed TAR identification using a moving window binomial approach ADDIN EN.CITE Li201186868617Li, J.Zhu, L.Eshaghi, M.Liu, J.Karuturi, K. M.Computational & Systems Biology, Genome Institute of Singapore, 60 Biopolis Street, (S)138672, Singapore. karuturikm@gis.a-star.edu.sg.Deciphering transcription factor binding patterns from genome-wide high density ChIP-chip tiling array dataBMC ProcBMC ProcS85 Suppl 22011/05/1120111753-6561 (Electronic)
1753-6561 (Linking)21554766http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=215547661753-6561-5-S2-S8 [pii]
10.1186/1753-6561-5-S2-S8eng[1]. Briefly, the probe intensity data for each condition was log transformed to follow a Gaussian distribution. We defined a base threshold as c fold MAD (median absolute deviation) and the probability, EMBED Equation.3 that a single probe passes the base threshold as:
EMBED Equation.3
where EMBED Equation.3 QUOTE (i = 1,., n) = median centered log transformed data for probe i, and n = total number of probes. The region EMBED Equation.3 represents the binomial sequence with signal probability EMBED Equation.3 QUOTE for each QUOTE EMBED Equation.3 , where w is the predefined half window size. EMBED Equation.3 was defined as the probability that EMBED Equation.3 is classified as signal within the signal region EMBED Equation.3 . A region EMBED Equation.3 was defined as a signal region if
EMBED Equation.3
where is the p-value cutoff of the binomial test. Parameters were optimized against known TARs, leading to selection of EMBED Equation.3 QUOTE , QUOTE EMBED Equation.3 and QUOTE EMBED Equation.3 for further analysis.
Experimental validation of transcripts
Transcribed regions were PCR-amplified using a MJ Research DNA engine Dyad Peltier thermocycler (Bio-Rad Laboratories, USA) in 100ul reactions containing 5 units of Platinum Taq DNA polymerase (Invitrogen, USA), 200uM dNTPs, 2mM MgCl2, 200nM primers, and 2ul template cDNA. Antisense transcripts were validated using strand-specific PCR. For data analysis, the absolute value of the Ct difference was plotted, between either the Forward or Reverse primers, and the RT reaction lacking any primers.
Enrichment analysis of silent genes
We evaluated the significance of overlap between 468 silent Sanger genes and (a) 2,040 genes coding for hypothetical proteins, (b) 1,346 genes not conserved in other B. pseudomallei strains by performing a hypergeometric test.
Relative expression of genes and ncRNAs
We calculated the relative expression, QUOTE EMBED Equation.3 of each gene/ncRNA, EMBED Equation.3 in a condition, EMBED Equation.3 as:
EMBED Equation.3
where QUOTE EMBED Equation.3 is the gene/ncRNA expression from all conditions; EMBED Equation.3 is the average expression; MAD is a function of computing the median absolute deviation.
Differential expression analysis
Changes in expression under one condition (denoted as T; e.g. pH 4) with respect to another condition (denoted as R; e.g. pH 7) were quantified in fold change and were calculated as follows ADDIN EN.CITE Yang200220920920917Yang, Y. H.Speed, T.Department of Statistics and Program in Biostatistics, 367 Evans Hall, 3860, University of California, Berkeley, California 94720-3860, USA.Design issues for cDNA microarray experimentsNat Rev GenetNat Rev Genet579-88382002/08/03Gene Expression ProfilingGene Expression RegulationHumansOligonucleotide Array Sequence Analysis/*methods*Research DesignTime Factors2002Aug1471-0056 (Print)
1471-0056 (Linking)12154381http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=1215438110.1038/nrg863
nrg863 [pii]eng[2]: Firstly, each condition was normalized to a common reference. The common reference (Rc) in our study is Bp K96243 grown to stationary phase in rich media (Luria Bertani broth, K9LBS). All common references were labeled with Cy5. Log-transformed fold changes (a.k.a. log ratio) were then computed as the difference between the two normalized values: EMBED Equation.3 . The absolute fold change is the untransformed log ratio. Genes or ncRNAs with absolute fold changes > 2 were considered biologically significant.
Evaluating the significance of overlap between detectible proteins and transcripts
We evaluated the significance of overlap between detectible proteins and transcripts at early stationary phase ADDIN EN.CITE Wongtrakoongate201119519519517Wongtrakoongate, P.Roytrakul, S.Yasothornsrikul, S.Tungpradabkul, S.Department of Biochemistry, Faculty of Science, Mahidol University, Bangkok 10400, Thailand. pexcotung@gmail.comA proteome reference map of the causative agent of melioidosis Burkholderia pseudomalleiJ Biomed BiotechnolJ Biomed Biotechnol53092620112011/10/01Bacterial Proteins/genetics/*metabolismBurkholderia pseudomallei/genetics/*metabolismElectrophoresis, Gel, Two-Dimensional/methodsHumansMelioidosis/*microbiologyProtein Processing, Post-Translational/geneticsProteome/genetics/*metabolismReference StandardsSpectrometry, Mass, Matrix-Assisted Laser Desorption-Ionization/methodsStress, Physiological/genetics20111110-7251 (Electronic)
1110-7243 (Linking)21960737http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=21960737318064110.1155/2011/530926eng[3] using a randomization procedure. A new set of detectible proteins was created through random selection from all protein-coding genes. The new set of proteins was then compared with the detectible transcripts. The steps were iteratively performed for 10,000 times, and the empirical p-value was then computed.
Optimizing the clustering granularity
Granularity of clustering was optimized against two parameters: (a) structural efficiency and (b) biological and functional coherence. Structurally efficient clustering, which achieves balance between the density and the size of clusters, was measured by using the tool, CLMINFO that is part of the MCL package ADDIN EN.CITE Enright200223232317Enright, A. J.Van Dongen, S.Ouzounis, C. A.Computational Genomics Group, The European Bioinformatics Institute, EMBL Cambridge Outstation, Cambridge CB10 1SD, UK. anton@ebi.ac.ukAn efficient algorithm for large-scale detection of protein familiesNucleic Acids ResNucleic Acids Res1575-843072002/03/28*AlgorithmsAmino Acid Sequence*Databases, ProteinGenome, HumanHumansInternetMolecular Sequence DataProteins/*geneticsSequence AlignmentSequence Homology, Amino AcidTranscription Factor TFIIBTranscription Factors/genetics2002Apr 11362-4962 (Electronic)
0305-1048 (Linking)11917018http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=11917018101833eng[4]. On the other hand, functional coherence of each cluster EMBED Equation.3 , EMBED Equation.3 QUOTE was measured by the fraction of gene pairs sharing identical Riley functional categories ADDIN EN.CITE Wuchty200725252517Wuchty, S.Ipsaro, J. J.Northwestern Institute on Complexity, Northwestern University, Evanston, Illinois 60208, USA. s-wuchty@northwestern.eduA draft of protein interactions in the malaria parasite P. falciparumJ Proteome ResJ Proteome Res1461-70642007/02/16AnimalsEvolution, MolecularMalaria/*parasitologyPlasmodium falciparum/*metabolismProtein Interaction MappingProteome/analysis/*metabolismProteomicsProtozoan Proteins/analysis/*metabolism2007Apr1535-3893 (Print)
1535-3893 (Linking)17300188http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=1730018810.1021/pr0605769eng[5], EMBED Equation.3 :
EMBED Equation.3
given the ith cluster contains QUOTE EMBED Equation.3 of annotated gene pairs. We assessed the functional coherence of clusters containing at least 60% of the interactions annotated in Riley functional categories, n. The functional coherence score generated from every cluster was summarized using the average coherence, QUOTE EMBED Equation.3 :
EMBED Equation.3
The scores for efficiency and coherence were normalized and compared. The optimal granularity was determined as the function of maximizing the scores of the two properties.
Correlations between ncRNAs and clusters
We measured mean Pearsons correlations between all pairs of ncRNAs and the expression clusters. The summarized correlation values were tested for significance using a Z-test, under the assumption that the correlations are normally distributed.
Transforming the weighted coexpression network for approximating a scale-free Network
We approximated the weighted coexpression network as a scale-free network by applying the fitting index EMBED Equation.3 on the linear regression of a spectrum of weights, EMBED Equation.3 and corresponding density, QUOTE EMBED Equation.3 in the logarithmic scale ADDIN EN.CITE Min201221212117Min, J. L.Nicholson, G.Halgrimsdottir, I.Almstrup, K.Petri, A.Barrett, A.Travers, M.Rayner, N. W.Magi, R.Pettersson, F. H.Broxholme, J.Neville, M. J.Wills, Q. F.Cheeseman, J.Allen, M.Holmes, C. C.Spector, T. D.Fleckner, J.McCarthy, M. I.Karpe, F.Lindgren, C. M.Zondervan, K. T.The Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, United Kingdom. jlmin@well.ox.ac.ukCoexpression network analysis in abdominal and gluteal adipose tissue reveals regulatory genetic loci for metabolic syndrome and related phenotypesPLoS GenetPLoS Genete1002505822012/03/032012Feb1553-7404 (Electronic)
1553-7390 (Linking)22383892http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=22383892328558210.1371/journal.pgen.1002505
PGENETICS-D-11-00968 [pii]eng[6]. The EMBED Equation.3 in a perfect scale-free network is 1.0. So, the EMBED Equation.3 in the coexpression network was optimized for being greater than 0.9. Therefore, the weights between any two genes, QUOTE EMBED Equation.3 and EMBED Equation.3 were iteratively transformed using a soft power function, EMBED Equation.3 QUOTE with increasing QUOTE EMBED Equation.3 values. The iteration stopped at EMBED Equation.3 QUOTE when EMBED Equation.3 QUOTE , thus formed an approximate scale-free network satisfying a power law: EMBED Equation.3 QUOTE .
Clustering of conditions based on condition-dependent profiles
We performed hierarchical clustering on the condition-dependent profiles using the hclust function in R, using the average linkage method and Pearsons correlation as the similarity measure, which was subsequently modified to dissimilarity by subtracting from 1. To assess the significance of clustering, bootstrap assessment was conducted with varying fractions of genes (multiscale, 50% to 140%) using Pvclust ADDIN EN.CITE Suzuki200611911911917Suzuki, R.Shimodaira, H.Department of Mathematical and Computing Sciences, Tokyo Institute of Technology 2-12-1 Ookayama, Meguro-ku, Tokyo 152-8552, Japan. ryota.suzuki@is.titech.ac.jpPvclust: an R package for assessing the uncertainty in hierarchical clusteringBioinformaticsBioinformatics1540-222122006/04/06Algorithms*Cluster AnalysisComputational Biology/*methodsGene Expression ProfilingGene Expression Regulation, NeoplasticHumansInternetModels, StatisticalOligonucleotide Array Sequence Analysis/*methodsProbabilityProgramming LanguagesSoftwareUncertainty2006Jun 151367-4803 (Print)
1367-4803 (Linking)16595560http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=16595560btl117 [pii]
10.1093/bioinformatics/btl117eng[7]. Groups of conditions with EMBED Equation.3 QUOTE were considered statistically robust.
Motif confirmation by BioProspector
All motifs identified by MEME ADDIN EN.CITE Bailey199439393917Bailey, T. L.Elkan, C.Department of Computer Science and Engineering, University of California at San Diego, La Jolla 92093-0114, USA.Fitting a mixture model by expectation maximization to discover motifs in biopolymersProc Int Conf Intell Syst Mol BiolProc Int Conf Intell Syst Mol Biol28-3621994/01/01AlgorithmsAnimalsBiopolymers/*chemistryHumans*Models, TheoreticalSequence Analysis19941553-0833 (Print)
1553-0833 (Linking)7584402http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=7584402eng[8] were confirmed using BioProspector ADDIN EN.CITE Liu200131313117Liu, X.Brutlag, D. L.Liu, J. S.Stanford Medical Informatics, 251 Campus Dr. X215, Stanford University, Stanford, CA 94305-5479, USA. xliu@smi.stanford.eduBioProspector: discovering conserved DNA motifs in upstream regulatory regions of co-expressed genesPac Symp BiocomputPac Symp Biocomput127-382001/03/27Bacillus subtilis/genetics/metabolismBase SequenceBinding SitesCarrier ProteinsConserved SequenceCyclic AMP Receptor Protein/metabolismDNA/*genetics/metabolismEscherichia coli/genetics/metabolismGene Expression Profiling/statistics & numerical data*Genes, RegulatorMarkov ChainsModels, GeneticOligonucleotide Array Sequence Analysis/statistics & numerical dataSaccharomyces cerevisiae/genetics/metabolismSequence Alignment/statistics & numerical data*SoftwareTATA Boxrap1 GTP-Binding Proteins/metabolism20011793-5091 (Print)11262934http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=11262934eng[9]. BioProspector parameters included (i) motif width corresponds to the MEME-discovered motif width, (ii) the motif may occur on only some input sequences, and (iii) single strand motif search.
Computing network distance between disrupted genes in T3SS3 and other differentially expressed genes
Two mutants were constructed by knocking out bsaN (BPSS1546) and bprC (BPSS1520). Genes with altered expression greater than two fold as a consequence were identified. Their distances from the mutated gene were calculated based on the transformed co-expression MIS. To determine if the observed distances are significantly shorter than chance, a randomization procedure was conducted by measuring the distances of a set of randomly selected genes, in which the total number of genes selected is the same as the differentially expressed genes. The observed and the random distances were compared using a one-tailed Wilcoxon rank sum test. The procedure was repeated for 10, 000 times. The p-values resulted from all iterations were then corrected using Benjamini and Hochberg method.
Computing weighted clustering coefficients
The weighted clustering coefficient (WCC) of each gene, EMBED Equation.3 was computed as follow ADDIN EN.CITE Zhang200534343417Zhang, B.Horvath, S.Department of Human Genetics, University of California at Los Angeles, USA. binzhang.ucla@gmail.comA general framework for weighted gene co-expression network analysisStat Appl Genet Mol BiolStat Appl Genet Mol BiolArticle1742006/05/0220051544-6115 (Electronic)
1544-6115 (Linking)16646834http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=1664683410.2202/1544-6115.1128eng[10]:
EMBED Equation.3
The weight QUOTE EMBED Equation.3 is the power-transformation of mutual information (see Transforming the weighted coexpression network for approximating a scale-free Network ). The WCC of the entire network was computed as the average WCC of all genes.
Motility assays
Motility phenotypes of wild type (Bp008) and quorum sensing mutant (Bp008::pmlI) were determined using motility tubes (BIOMERIEUX, Marcy lEtoile). Strains were grown during 24 hours on a TSA plate, and inoculated into tubes. Tubes were observed after 24hrs and 48hrs. Motility was assessed qualitatively by examining the circular swarm formed by growing motile cells.
Bp capsule electron microscopy
Electron microscopy was performed using protocols modified from Puthucheary et al., 1996 ADDIN EN.CITE Puthucheary199613313313317Puthucheary, S. D.Vadivelu, J.Ce-Cile, C.Kum-Thong, W.Ismail, G.Department of Medical Microbiology, Faculty of Medicine, University of Malaya, Kuala Lumpur.Short report: Electron microscopic demonstration of extracellular structure of Burkholderia pseudomalleiAm J Trop Med HygAm J Trop Med Hyg313-45431996/03/01Burkholderia pseudomallei/chemistry/*ultrastructureCell Wall/ultrastructureGlycoproteins/analysis/*ultrastructureHumansMelioidosis/*microbiologyMicroscopy, ElectronPolysaccharides/analysis/*ultrastructure1996Mar0002-9637 (Print)
0002-9637 (Linking)8600773http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=8600773eng[11]. Bacterial cultures were harvested from hypoosmotic broth (50mM NaCl) in enriched CO2 after 48 hours. Pelleted cells were washed with 0.15M cacodylate buffer, pH7.0 before fixation in 25% glutaraldehyde (Sigma, USA), 0.2M cacodylate buffer, pH 7.0 (Amersham, GE Healthcare, United Kingdom), and 0.15% (w/v) aqueous ruthenium red (Fluka, USA) solution (ratio 1:6:7) for 2hrs at room temperature. Cells were washed in 0.15M cacodylate buffer and post-fixed with 4% osmium tetroxide (EMS, Euromedex, France), 0.2M cacodylate buffer, pH7.0, and 0.15% (w/v) ruthenium red mixed together in equal volumes for 3hrs at 4C in the dark. The preparation was then incubated in a waterbath at 40C and cells were embedded in 3% agarose (Sigma, USA). Finally, the preparation was embedded in spur resin (EMS, Euromedex, France), double-stained with uranyl acetate and lead citrate (EMS, Euromedex, France), prior to transmission electron microscopy using a JEOL JEM1010 (Japan) microscope.
Identifying the in-vitro conditions showing similar expression profiles to mice lung infection
The similarity of mice lung infection profile to other in vitro physio-chemical profiles (also known as a reference database) was determined by using a nonparametric, rank-based pattern matching strategy based on the Kolmogorov-Smirnov statistic ADDIN EN.CITE ADDIN EN.CITE.DATA [12]. The analysis began with a query signature and assessed its similarity to each expression profile in the reference database. The query signature is a list of genes with altered expression in mice lung infection. Detailed implementation is reported in ADDIN EN.CITE ADDIN EN.CITE.DATA [12].
SUPPLEMENTAL REFERENCES
ADDIN EN.REFLIST 1. Li J, Zhu L, Eshaghi M, Liu J, Karuturi KM (2011) Deciphering transcription factor binding patterns from genome-wide high density ChIP-chip tiling array data. BMC Proc 5 Suppl 2: S8.
2. Yang YH, Speed T (2002) Design issues for cDNA microarray experiments. Nat Rev Genet 3: 579-588.
3. Wongtrakoongate P, Roytrakul S, Yasothornsrikul S, Tungpradabkul S (2011) A proteome reference map of the causative agent of melioidosis Burkholderia pseudomallei. J Biomed Biotechnol 2011: 530926.
4. Enright AJ, Van Dongen S, Ouzounis CA (2002) An efficient algorithm for large-scale detection of protein families. Nucleic Acids Res 30: 1575-1584.
5. Wuchty S, Ipsaro JJ (2007) A draft of protein interactions in the malaria parasite P. falciparum. J Proteome Res 6: 1461-1470.
6. Min JL, Nicholson G, Halgrimsdottir I, Almstrup K, Petri A, et al. (2012) Coexpression network analysis in abdominal and gluteal adipose tissue reveals regulatory genetic loci for metabolic syndrome and related phenotypes. PLoS Genet 8: e1002505.
7. Suzuki R, Shimodaira H (2006) Pvclust: an R package for assessing the uncertainty in hierarchical clustering. Bioinformatics 22: 1540-1542.
8. Bailey TL, Elkan C (1994) Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc Int Conf Intell Syst Mol Biol 2: 28-36.
9. Liu X, Brutlag DL, Liu JS (2001) BioProspector: discovering conserved DNA motifs in upstream regulatory regions of co-expressed genes. Pac Symp Biocomput: 127-138.
10. Zhang B, Horvath S (2005) A general framework for weighted gene co-expression network analysis. Stat Appl Genet Mol Biol 4: Article17.
11. Puthucheary SD, Vadivelu J, Ce-Cile C, Kum-Thong W, Ismail G (1996) Short report: Electron microscopic demonstration of extracellular structure of Burkholderia pseudomallei. Am J Trop Med Hyg 54: 313-314.
Ohiuv
- . ? B U V e f z { ̼{{{{{{{{{{{{{{{{{{{hCh<CJH*OJQJaJhCh<CJOJQJaJhj@Vh5CJOJQJaJ"hj@VhJG56CJOJQJaJhj@VhJG5CJOJQJaJhj@Vhql'CJ OJQJhj@Vhx86CJ OJQJhj@Vh(pCJ OJQJhj@VhCJ OJQJ0i F
wl/
{
UVk5$dha$gdhdhG$gd<gd<$dha$gd<$da$gdJG$a$gd F
G
wxlm/
1
{
}
UJLNOν{ocohj@Vh05OJQJhj@Vhb5OJQJ)hj@Vhh0JB*CJOJQJaJphhj@Vhhjhj@VhhUhj@VhhCJOJQJaJh@lCJOJQJaJ hCh<CJOJPJQJaJ#hCh<CJH*OJPJQJaJhCh<CJOJQJaJhCh<CJH*OJQJaJ&5JKLMNObw$w$
&FnWdx^n`Wa$gd,$
&FnWdx7$8$H$^n`Wa$gd,
&FnWdx^n`Wgd,$d<^a$gdDN$dha$gdhOabovw$Y]wbcuv78|pg[Khj@Vh5;5CJOJQJaJhj@Vh&q&6OJQJh /6OJQJhj@VhQ6OJQJhw!6OJQJhe6 6OJQJhj@VhX6OJQJhj@Vh06OJQJhj@Vh6OJQJh46OJQJhj@Vhfp6OJQJhj@Vhx865CJOJQJaJhj@VhDN5CJOJQJaJhj@Vh905OJQJhj@Vh5OJQJ\$Cx$d<^a$gd5;$d<^a$gdDN$
&FnWdx^n`Wa$gdEh$
&FnWdx^n`Wa$gd,
&FnWdx^n`Wgd,$
&FnWdx^n`Wa$gdX./viXiXFX7hj@Vh(dCJOJQJaJ#hj@Vh(dOJQJaJmHnHu!jhj@Vh(dOJQJUaJhj@Vh(dOJQJaJhj@Vh(dCJOJQJhj@VhDN56OJQJhj@VhDNCJOJQJaJhj@Vh>>>>tiWFt7jhj@Vh"OJQJU!jhj@Vh}oEHOJQJU#j1 U
hj@Vh}oOJQJUVhj@VhHvOJQJjhj@VhHvOJQJU!jhj@Vh}oEHOJQJU#jZ1 U
hj@Vh}oOJQJUVjhj@Vh}oOJQJUhj@Vh}oOJQJhj@VhDNOJQJhj@VhOJQJmHnHuhj@VhOJQJjhj@Vh5QOJQJUhj@Vh5QOJQJ>>>>!>p>z>{>|>DDDDDDDDDDDDEEδΊm\ΊJ#j1 U
hj@Vh}oOJQJUV!j9hj@Vh}oEHOJQJU#j1 U
hj@Vh}oOJQJUVhj@VhHvOJQJjhj@VhHvOJQJUhj@VhOJQJmHnHuhj@VhOJQJjhj@Vh6"BOJQJUhj@VhxOJQJhj@VhDNOJQJjhj@Vh"OJQJUjhj@Vh"EHUhj@Vh"OJQJE
EEEE.E/E6E7E8E9E:EMENEOEPEQEwE{EEEEEEEEE^F_FfFԹudYYQYYԹhyOJQJhj@Vh!OJQJ!jhj@Vh}oEHOJQJU#j2 U
hj@Vh}oOJQJUVhj@Vh}oOJQJjhj@Vh}oOJQJUjhj@Vh"EHUhj@Vh"OJQJjhj@Vh"OJQJUhj@VhDN6OJQJhj@VhDNOJQJjhj@VhHvOJQJU!jKhj@Vh}oEHOJQJUfFgFhFiFjF}F~FFFFFFFFFFHGIGrGfHhHɾؐɾ~mɐaSH=hj@VhXOJQJhj@VhYOJQJhj@VhX56OJQJhj@Vh5OJQJ!jhj@VhEHOJQJU#j+2 U
hj@VhOJQJUVhj@VhDNOJQJ!jhj@VhEHOJQJU#j2 U
hj@VhOJQJUVhj@Vh}oOJQJjhj@Vh}oOJQJUjhj@Vh"OJQJUhj@Vh"OJQJjhj@Vh"EHUFFHGIGrGgHhHH2T3TrT]]]ylzlloop$ddd[$\$a$gdP\,ddd[$\$gdDN$ddd[$\$a$gdX$ddd[$\$a$gdDN$ddd[$\$a$gd}ohHH%I&I9I:I;I*S?S@SASTSUSVSWSXSzS{SSSSSSSSSSSSSSɾɾ~m`ɾN#j=3 U
hj@VhOJQJUVjhj@Vh"EHU!jkhj@VhEHOJQJU#j,3 U
hj@VhOJQJUVhj@VhDNOJQJ!jqhj@VhEHOJQJU#j 3 U
hj@VhOJQJUVhj@VhOJQJjhj@VhOJQJUjױhj@Vh"EHUhj@Vh"OJQJjhj@Vh"OJQJUSSSSSSSS
TT!T"T#T$T%T,T-T.T0T2T3TrTTTVŸЭߢrЭfXLhj@VhDN6OJQJhj@VhDN56OJQJhj@VhDN5OJQJj<hj@Vh"EHU!j9hj@Vh&cEHOJQJU#jf3 U
hj@Vh&cOJQJUVhj@VhOJQJhj@VhDNOJQJjhj@Vh"EHUhj@Vh"OJQJjhj@Vh"OJQJUjhj@VhOJQJU!j
hj@VhEHOJQJUVVd]e]h]i]]]]]]]]]]]]]]]]]ʻ~sfs~ZLALh56OJQJhj@VhDN56OJQJhj@VhDN5OJQJjEnhj@Vh"EHUhj@Vh"OJQJjhj@Vh"OJQJU!jkhj@Vh&cEHOJQJU#j{3 U
hj@Vh&cOJQJUVhj@VhOJQJjhj@VhOJQJUhj@VhDNOJQJhj@VhOJQJmHnHuhj@VhOJQJjhj@Vh6"BOJQJU]]]^^^cccccckkkkxllllllmmmm m$m&m.mѾ܌qcXLLLLhj@VhDN6OJQJh /56OJQJhj@Vh_56OJQJhj@VhOJQJmHnHuhj@VhOJQJjhj@Vh6"BOJQJU'hj@VhCJOJQJaJmHnHuhj@VhCJOJQJaJ%jhj@VhenCJOJQJUaJhj@VhenOJQJhj@VhDNOJQJhj@VhDN56OJQJhf"56OJQJ.mmm)n5nnn8o:ooopSpTpgphpipjpppDuEuIuJuLurcXcHchj@VhOJQJmHnHuhj@VhOJQJjhj@Vh6"BOJQJU!jhj@VhjEHOJQJU#j5 U
hj@VhjOJQJUVhj@VhxOJQJjhj@VhxOJQJUhj@VhDN56OJQJhj@VhDN6OJQJhj@Vh8MOJQJhj@Vh8OJQJhj@Vhy$OJQJhj@Vh&OJQJhj@VhDNOJQJpLudueuvvwxxỹ̓,}~fʇ$da$gd$d<a$gd $ddd[$\$a$gdDN$ddd[$\$a$gdx$ddd[$\$a$gdDNLuMu`uaubucudupuquxuyuzu{u|uuuuuuvvvw巨vdSH:hj@VhDN56OJQJhj@Vh~OJQJ!jhj@VhjEHOJQJU#j6 U
hj@VhjOJQJUVhj@VhjOJQJjhj@VhjOJQJUj hj@Vh"EHUhj@Vh"OJQJjhj@Vh"OJQJUhj@VhDNOJQJ!jhj@VhjEHOJQJU#j5 U
hj@VhjOJQJUVhj@Vh&cOJQJjhj@Vh&cOJQJUwwwwxx y
yyyyaygynyoyOP˃̃̓,ck"#23GĶtiZijhj@VhOJQJUhj@VhOJQJ*hj@VhDNOJQJhj@VhDNH*OJQJhj@VhOJQJmHnHujhj@Vh6"BOJQJUhj@Vh(dOJQJhj@Vh&q&56OJQJhj@VhDN56OJQJhj@VhjE6OJQJhj@VhDN6OJQJhj@Vh4uOJQJhj@VhDNOJQJGHIJNOMN]^rstuyz~ϿϴϩީϿϴxhXhE%jhj@Vh6"BCJOJQJUaJhj@Vh=R5CJOJQJaJhj@Vhp75CJOJQJaJhj@VhN55CJOJQJaJhj@VhDNCJOJQJaJ#jHhj@VhOJQJUhj@VhOJQJhj@VhDNOJQJhj@VhOJQJmHnHujhj@Vh6"BOJQJUjhj@VhOJQJU#jhj@VhOJQJUٍyuh}fh@lmHnHuh,hjh,hUh,zjh,zUhj@Vh5;5CJOJQJaJhj@Vhv`CJOJQJaJhj@VhWnaCJOJQJaJU'hj@VhCJOJQJaJmHnHu%jhj@Vh6"BCJOJQJUaJhj@Vhp7CJOJQJaJ*5}ٍ dgd,$d<a$gd5;$d<a$gd $da$gd$da$gd12. Lamb J, Crawford ED, Peck D, Modell JW, Blat IC, et al. (2006) The Connectivity Map: using gene-expression signatures to connect small molecules, genes, and disease. Science 313: 1929-1935.
PAGE \* MERGEFORMAT 1
$d<a$gd5;$a$21h:p] . A!n"#n$%Dd
b
c$A??3"`?2D:902 D`!:902:Rxcdd`` $X bbd12,(ㆫaZ(Wcgb ԀРc@
UXRY7S?$L ZZǰ8֖Ue֕JKf3@- p]c!I)$5dP+CX7>3@`2`7FDd
b
c$A??3"`?2T?Q4#
>`!T?Q4#
@@2xRKQ]*PjN.pjZj\H.Z\:Z&^O)84KK^" ق>0,:h(J}%Ŕ]'w|QmҚ"M
|B`f+z"qNgI36A[M?+YV9a7lZ]qޜ툄Ý3(;$ϙ5?u;2WNuT_֞p
U<6^˙.%}2>17-a|
Ĺt\IEågLmPo,?PDd
hb
c$A??3"`?2_,0DY;`!3,0DY@R|xcdd``> $X bbd12,(ㆫar`j`3H1g`Yˀh3 憪aM,,He`HI&meabM-VK-WMcAakѲ'f망e0c7L
j#ps}VTrA>d`p021)W2Ë'q Cbd`Mu*Dd
,j)
3A",)&)
12?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnoprstuvwx{~Root Entry F6m}?Data
qWordDocument4ObjectPoolrd`6m_1436951702;F``Ole
CompObjfObjInfo
!$'*-./038=BGLQTW\adehmrwz}
FMicrosoft Equation 3.0DS EquationEquation.39qi
,\i+
p
FMicrosoft Equation 3.0DS EquationEquation.39qEquation Native )_1426664864Fd`d`Ole
CompObj
fObjInfo
Equation Native _1426664933 Fd`d`Ole
f`iQh
p=#{xi
|xi
e"cMAD}n
FMicrosoft Equation 3.0DS EquationEquation.39qCompObj
fObjInfoEquation Native 6_1426664889F``i$Vh
xi
FMicrosoft Equation 3.0DS EquationEquation.39qE@ltTh
{xi
}i"wi+wOle
CompObjfObjInfoEquation Native asidRoot wsp:val="00303990"/>*