Selection of DDX5 as a novel internal control for Q-RT-PCR from microarray data using a block bootstrap re-sampling scheme
© Su et al; licensee BioMed Central Ltd. 2007
Received: 11 January 2007
Accepted: 01 June 2007
Published: 01 June 2007
The development of microarrays permits us to monitor transcriptomes on a genome-wide scale. To validate microarray measurements, quantitative-real time-reverse transcription PCR (Q-RT-PCR) is one of the most robust and commonly used approaches. The new challenge in gene quantification analysis is how to explicitly incorporate statistical estimation in such studies. In the realm of statistical analysis, the various available methods of the probe level normalization for microarray analysis may result in distinctly different target selections and variation in the scores for the correlation between microarray and Q-RT-PCR. Moreover, it remains a major challenge to identify a proper internal control for Q-RT-PCR when confirming microarray measurements.
Sixty-six Affymetrix microarray slides using lung adenocarcinoma tissue RNAs were analyzed by a statistical re-sampling method in order to detect genes with minimal variation in gene expression. By this approach, we identified DDX5 as a novel internal control for Q-RT-PCR. Twenty-three genes, which were differentially expressed between adjacent normal and tumor samples, were selected and analyzed using 24 paired lung adenocarcinoma samples by Q-RT-PCR using two internal controls, DDX5 and GAPDH. The percentage correlation between Q-RT-PCR and microarray were 70% and 48% by using DDX5 and GAPDH as internal controls, respectively.
Together, these quantification strategies for Q-RT-PCR data processing procedure, which focused on minimal variation, ought to significantly facilitate internal control evaluation and selection for Q-RT-PCR when corroborating microarray data.
Microarrays, by making use of the sequence resources created in genomic projects, are a powerful technology capable of measuring the expression levels of thousands of genes simultaneously and have dramatically expedited comprehensive understanding of gene expression profiles for disease development. For example, microarray technology has been used to compare gene expression profiles between normal and diseased cells and this has led to dramatic advances in the understanding of cellular processes at the molecular level . Several microarray platforms are currently available. The short-oligonucleotide-based Affymetrix GeneChip® arrays utilize multiple probes for each gene with an automated control for the experimental process from hybridization to quantification and thus provide reliable and comparable data . The multiple probe sets for each gene are typically scattered across the surface of the Affymetrix microarrays. Variations in intensity from probe to probe or chip to chip for samples need to be resolved to obtain a reliable level of expression. Various statistical algorithms are available for probe-cell level normalization and expression-value summary.
Researchers are still confronted with challenging questions after completing the expression profiling and these include how to validate and standardize the data processing using proper statistical analysis. Quantitative-real time-reverse transcription PCR (Q-RT-PCR) is widely used and is a sensitive and robust technique for the detection and quantification of often rare mRNA targets . Q-RT-PCR has also become one of the gold standards for both pathogen detection and gene expression studies and is the method of choice for corroborating microarray data . In this study, the Q-RT-PCR system is based on the detection of the fluorescent activity and quantification of the TaqMan® probe, which undergoes cleavage in proportion to the amount of PCR product formed [5, 6]. By recording the amount of fluorescence emission at each cycle, it is possible to monitor the PCR reaction during the exponential phase where the first significant increase occurs and the amount of PCR product correlates to the initial amount of target template.
An appropriate internal control for Q-RT-PCR should be expressed stably across all data samples and if this is true, measurement of genes relative to the internal control will reflect the real gene expression. It implies that a reference gene should have a small variance and a sufficient intensity when applied as an appropriate internal control. Moreover, most published studies have focused on the identification of reference genes that can be used to normalize expression of a gene across patient samples or tissue types rather than within one specific type of tissue or cell line [7, 8]. Generally speaking, housekeeping genes, such as ACTB (actin, β), GAPDH (glyceraldehyde-3-phosphate dehydrogenase), and 18S ribosomal RNA, are commonly employed in Q-RT-PCR analysis [9–11]. However, several studies have also demonstrated that the gene expression patterns of many commonly used internal controls may vary as a result of tissue type, experimental conditions or pathological state [12–15]. The "perfect" control gene for all Q-RT-PCR does not exist because variability in Q-RT-PCR data can also stem from differences in the expression of the reference gene, for example GAPDH and ATCB, on which the expression of all the other genes is based . Although 18S ribosomal RNA has been shown to be a reliable control in many studies [7, 8, 17], it does not undergo reverse transcription when using oligo (dT) primers and is inappropriate for use when such primers are used. Szabo et al. developed statistical models to assist in identifying appropriate housekeeping genes as Q-RT-PCR normalization controls in one or multiple types of tissue samples . However, their rigorous approach heavily relies on an assumption that there is a multivariate normal distribution for the microarray expression levels and may not fit a practical situation, especially without a large number of arrays. In addition, their models are only applicable to the analysis of random samples, not paired samples collected from each patient as in this study.
We aimed to address two unanswered questions associated with microarray target selections for Q-RT-PCR validation. Firstly, it is not certain which gene or genes can serve as better internal controls for Q-RT-PCR simply because there is no perfect internal control . Secondly, a major challenge when scoring the correlation between microarray and Q-RT-PCR measurements remains unsettled because different probe level normalization methods may result in different correlations. In this study, we propose a statistical re-sampling method to display the variation pattern or to calculate the inter-quartile range (IQR) and the variance of gene expression levels that are associated with different probe level normalization methods. We utilized the block bootstrap re-sampling technique to circumvent the within-block dependence of Affymetrix microarray data when using paired adjacent normal and tumor samples from lung adenocarcinoma patients. Moreover, we employed box plot results for lung adenocarcinoma gene expression and identified DDX5 as a novel internal control for Q-RT-PCR. DDX5 is a highly conserved member of the DEAD box family and is known to be a RNA helicase that is involved in both pre-mRNA and pre-rRNA processing . Twenty-three genes, which were differentially expressed between adjacent normal and tumor samples, were further selected for Q-RT-PCR analysis and were examined by microarray analysis with several probe level quantile normalization methods using either DDX5 or GAPDH as internal controls. No matter which probe level quantile normalization was used for comparison, DDX5 was a better internal control than GAPDH for lung cancer datasets.
Results and Discussion
Identification of a novel internal control through variation in gene expression levels
Summary of probe set characteristics of 10 well-known internal controls for Q-RT-PCR in Affymetrix HG-U133A chip
Probe set ID
Column # in Fig. 2A
200801 _x at
The copy number of the individual housekeeping gene chosen for relative quantification should be in a similar range to that of the target gene to make comparative quantification possible . Further analysis indicates that we can identify a series of internal control candidates, which have characteristics of small variance in different microarray intensity intervals (Additional file 1). These potential internal controls are presented in different intensity ranges in order to appropriately normalize different target genes. Despite the fact that these potential internal controls may exhibit a greater variance than DDX5 or other internal controls listed in Figure 2A, these potential internal controls have much smaller variance than ACTB and GAPDH (Additional file 1, right portion). This finding further supports a view that the intensities of normalized microarray data and the copy numbers of Q-RT-PCR detections in gene expression patterns could be examined in a similar range. Our approach may provide a method to identify potential internal controls to be in a similar range of expression as the selected target genes.
To prioritize the potential internal control for the lung cancer microarray data, two major public accessible lung cancer microarray datasets, which also used Affymetrix chips, namely the Boston and the Ann Arbor datasets [23, 24], were included for data comparison. Eleven candidates, after the exclusion of ABCF1, BHLHB2 and LAPTM4A, were available in both Boston and Ann Arbor datasets and were included in the analysis. Figure 2B and 2C show the results of basic bootstrapping using these two datasets of unpaired design. All 11 candidates exhibited less variation than most well known internal controls, suggesting that all 11 candidates have potential to serve as an internal control, at least for lung cancer. To finalize the target for further empirical validation, these potential controls were sorted in the order of increasing gene expression intensities and decreasing IQR, respectively. As a result, CLTC, DDX5 and MSN were found to exhibit sufficient intensity. However, DDX5 gave the smallest variation among the three. Therefore, in this study, we chose DDX5 for further characterization.
Q-RT-PCR validation and comparison with microarray
Summary of the correlations between microarray and Q-RT-PCR analyzed by Pearson's and Kendall's τ correlations.
Kendall's τ Correlation
The expression pattern comparisons of DDX5 in other microarray datasets
In summary, we adapted block bootstrap for using with a paired design to circumvent the within-pair dependence. This proposed re-sampling strategy together with the use of a box plot provides a useful distribution-free statistical procedure for exploratory data analysis. This systematic analysis procedure focused on identifying genes with minimal variation in their microarray data, which facilitates the essential internal control selection steps prior to Q-RT-PCR analysis. Finally, systematic microarray and Q-RT-PCR analyses reveal that the proposed re-sampling technique of block bootstrap suits paired design experiments and adequately detects genes with minimal variation in a microarray dataset.
A total of 66 samples were used for microarray analysis, including paired adjacent normal-tumor samples from 27 patients underwent surgery for lung cancer at the Taipei Veterans General Hospital, two tissue mixtures from the Taichung Veterans General Hospital (one was adjacent normal lung mixtures and the other was lung adenocarcinoma mixtures), two commercial human normal lung tissues (Clontech (Catalog No. 636524) and Stratagene (Catalog No. 735020)), one immortalized, nontumorigenic human bronchial epithelial cell line (NL-20 (ATCC® No. CRL-2503™)) and 7 lung cancer cell lines (A-549 (ATCC® No. CCL-185™), NCI-H1299 (ATCC® No. CRL-5803™), NCI-H661 (ATCC® No. HTB-183™), CL1-0 , CL1-1 , CL1–5 , and CL1–5-F4 ).
RNA extraction and reverse transcription
We used the total RNA samples for Q-RT-PCR analyses. RNA preparation and analysis were performed according to the previous study . Briefly, the quality of the total RNA for microarray analysis was determined using Spectra Max Plus (Molecular Devices) and had an OD260/OD280 ratio ranging from 1.9 to 2.1. RNA was subjected to reverse transcription with random hexamer primers. To hydrolyze contaminating DNA in the RNA preparations, RNA was incubated with amplification-grade DNase I (Life Technologies, Gaithersburg, MD). After incubating the reaction mixture, the reaction was stopped by heating at 65°C. After DNase treatment, the RNA was subjected to reverse transcription reaction by the ThermoScript™ RT-PCR system (Life Technologies) and cDNAs were then used in the Q-RT-PCR.
Summary table of Assays-on-Demand ID of 23 genes and 2 internal controls for Q-RT-PCR.
activator of S phase kinase
BUB1 budding uninhibited by benzimidazoles 1 homolog beta (yeast)
cell division cycle associated 8
centaurin, delta 2
chemokine (C-X-C motif) ligand 5
cytochrome P450, family 27, subfamily A, polypeptide 1
chromosome 10 open reading frame 3
hypothetical protein FLJ20530
hypothetical protein FLJ20605
Duffy blood group
frizzled homolog 4(Drosophila)
glycoprotein A repetitions predominant
matrix metalloproteinase 9 (gelatinase B, 92 kDa gelatinase, 92 kDa type IV collagenase)
macrophage scavenger receptor 1
S100 calcium binding protein A2
serine (or cysteine) proteinase inhibitpr, clade A (alpha-1 antiproteinase, antitrypsin), member 3
SRY(sex determining region Y)-box 4
steroid-5-alpha-reductase, alpha polypeptide 1 (3-oxo-5 alpha-steroid delta 4-dehydrogenase
lung type-I cell membrane-associated glycoprotein
TEK tyrosine kinase, endothelial (venous malformations, multiple cutaneous and mucosal)
T-LAK cell-originated protein kinase
trophinin associated protein (tastin)
DEAD (Asp-(Glu-Ala-Asp) box polypeptide 5
Microarray experiments, normalization and Wilcoxon signed-rank test
Protocols, reagents for hybridization, washing and staining followed previous methods  and the Affymetrix's instructions . Labeled cDNA was hybridized to the Affymetrix GeneChip Test 3 Array to verify quality prior to hybridize to the Affymetrix Human Genome U133A Array. The data discussed in this publication have been deposited in NCBIs Gene Expression Omnibus (GEO)  and are accessible through GEO Series accession number GSE7670. The images were transformed into text files as intensity information using the MAS5.0 (Microarray Suite software 5.0) developed by Affymetrix [38, 39]. We used three array processing methods to produce and normalize the Affymetrix expression signals for the transcripts based on corresponding probe pairs of oligonucleotides. MAS5.0 as provided by Affymetrix was used to carry out the probe-pairs adjustment. Both MAS5.0 and RMA , Robust Multichip Average via quantile adjustment, are commonly used. Another algorithm GC-RMA  is similar to RMA, but pools the probes with comparable numbers of G-C bonds to achieve a stable mismatch adjustment. Based on paired adjacent normal and tumor samples of 27 patients, we used the Wilcoxon signed-rank test  to identify differentially expressed genes. In contrast to Student's t-test inappropriately used in some studies , the signed-rank test is distribution-free and adjusts for the paired design.
To assess the variation in the microarray expression level for a specified gene in an experiment of moderate sample size and possibly including paired samples, we designed a block bootstrapping procedure to analyze the microarray data from 66 lung samples, containing 27 pairs of patient adjacent normal-tumor samples and 12 un-paired samples. Bootstrapping is a breakthrough statistical approach using a computationally intensive re-sampling technique and it allows complex problems to be solved in which the accuracy of a devised statistical procedure can not be analytically evaluated [44, 45]. Block bootstrapping was originally named for re-sampling methods in dependent cases, especially time series data . The basic bootstrap generates artificial samples that allow the making of an inference of interest through re-sampling the original data with replacement in which all observations are assumed to be mutually independent and from the same distribution. To guarantee the structure of independence in bootstrap re-sampling, we employed the concept of blocking for the paired data by treating each individual patient, mixture tissue or cell line as a block. By selecting an observation within the block with equal probability when combined with all the other un-paired samples, we obtained an independently re-sampled dataset. We then created a bootstrap sample by randomly sampling the blocks in the dataset, and computed the bootstrap replicates of the relevant summary statistics of the expression levels. Repeating this bootstrap re-sampling scheme sufficient times, such as 1,000, we then used the averages of these bootstrap replicates to reveal the variation in expression summaries corresponding to a specific gene across the microarrays. Appropriate internal controls can be selected by ranking the variations in gene expression.
Correlations of microarray and Q-RT-PCR data
To explicate the correlation between microarray and Q-RT-PCR in this study of paired design, we calculated the differences between the log-scaled measurements of the Q-RT-PCR and microarraydata from the tumor and adjacent normal tissues of 24 patients. The other 3 paired samples did not have sufficient materials for further Q-RT-PCR analysis. Pearson's and Kendall's τ correlation coefficients were then tested at a significant level of 0.05. Summarization of the expression levels and normalization for microarray data were conducted using GeneSpring® 7.3 (Silicon Genetics, Redwood City, CA). The computer programs for the block bootstrapping method and correlation using R 2.1.1  are presented in the Additional File 2.
This work was supported in part by grants from the National Health Research Institutes, National Science Council (Program for Interdisciplinary Research Project: NSC95-2627-B-400-002) and Department of Health (DOH94-TD-G-111-013) to Dr. Chi-Ying F. Huang, a grant from the National Science Council (NSC95-3112-B-400-009) to Dr. Jacqueline Whang-Peng, and a grant from the National Science Council (NSC 95-3112-B-001-018-Y) to Dr. Chen-Hsin Chen. The authors thank the technical supports provided by Microarray & Gene Expression Analysis Core Facility of the National Yang-Ming University VGH Genome Research Center (VYMGC), and the Genetic Statistic Unit of the Advanced Bioinformatics Core. Both Cores are supported by National Research Program for Genomic Medicine (NRPGM), National Science Council, Taiwan.
- Jain KK: Applications of biochips: from diagnostics to personalized medicine. Curr Opin Drug Discov Devel. 2004, 7 (3): 285-289.PubMedGoogle Scholar
- Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ: High density synthetic oligonucleotide arrays. Nat Genet. 1999, 21 (1 Suppl): 20-24. 10.1038/4447.PubMedView ArticleGoogle Scholar
- Bustin SA: Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays. J Mol Endocrinol. 2000, 25 (2): 169-193. 10.1677/jme.0.0250169.PubMedView ArticleGoogle Scholar
- Yun JJ, Heisler LE, Hwang, Wilkins O, Lau SK, Hyrcza M, Jayabalasingham B, Jin J, McLaurin J, Tsao MS, Der SD: Genomic DNA functions as a universal external standard in quantitative real-time PCR. Nucleic Acids Res. 2006, 34 (12): e85-10.1093/nar/gkl400.PubMed CentralPubMedView ArticleGoogle Scholar
- Lee LG, Connell CR, Bloch W: Allelic discrimination by nick-translation PCR with fluorogenic probes. Nucleic Acids Res. 1993, 21 (16): 3761-3766. 10.1093/nar/21.16.3761.PubMed CentralPubMedView ArticleGoogle Scholar
- Livak KJ, Flood SJ, Marmaro J, Giusti W, Deetz K: Oligonucleotides with fluorescent dyes at opposite ends provide a quenched probe system useful for detecting PCR product and nucleic acid hybridization. PCR Methods Appl. 1995, 4 (6): 357-362.PubMedView ArticleGoogle Scholar
- Aerts JL, Gonzales MI, Topalian SL: Selection of appropriate control genes to assess expression of tumor antigens using real-time RT-PCR. Biotechniques. 2004, 36 (1): 84-6, 88, 90-1.PubMedGoogle Scholar
- Kim BR, Nam HY, Kim SU, Kim SI, Chang YJ: Normalization of reverse transcription quantitative-PCR with housekeeping genes in rice. Biotechnol Lett. 2003, 25 (21): 1869-1872. 10.1023/A:1026298032009.PubMedView ArticleGoogle Scholar
- Suzuki T, Higgins PJ, Crawford DR: Control selection for RNA quantitation. Biotechniques. 2000, 29 (2): 332-337.PubMedGoogle Scholar
- Thellin O, Zorzi W, Lakaye B, De Borman B, Coumans B, Hennen G, Grisar T, Igout A, Heinen E: Housekeeping genes as internal standards: use and limits. J Biotechnol. 1999, 75 (2-3): 291-295. 10.1016/S0168-1656(99)00163-7.PubMedView ArticleGoogle Scholar
- Lin YS, Su LJ, Yu CT, Wong FH, Yeh HH, Chen SL, Wu JC, Lin WJ, Shiue YL, Liu HS, Hsu SL, Lai JM, Huang CY: Gene expression profiles of the aurora family kinases. Gene Expr. 2006, 13 (1): 15-26.PubMedView ArticleGoogle Scholar
- Bereta J, Bereta M: Stimulation of glyceraldehyde-3-phosphate dehydrogenase mRNA levels by endogenous nitric oxide in cytokine-activated endothelium. Biochem Biophys Res Commun. 1995, 217 (1): 363-369. 10.1006/bbrc.1995.2785.PubMedView ArticleGoogle Scholar
- Gibbs PJ, Cameron C, Tan LC, Sadek SA, Howell WM: House keeping genes and gene expression analysis in transplant recipients: a note of caution. Transpl Immunol. 2003, 12 (1): 89-97. 10.1016/S0966-3274(03)00010-8.PubMedView ArticleGoogle Scholar
- Hamalainen HK, Tubman JC, Vikman S, Kyrola T, Ylikoski E, Warrington JA, Lahesmaa R: Identification and validation of endogenous reference genes for expression profiling of T helper cell differentiation by quantitative real-time RT-PCR. Anal Biochem. 2001, 299 (1): 63-70. 10.1006/abio.2001.5369.PubMedView ArticleGoogle Scholar
- Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F: Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002, 3 (7): RESEARCH0034-10.1186/gb-2002-3-7-research0034.PubMed CentralPubMedView ArticleGoogle Scholar
- Moore DJ, Chambers JK, Wahlin JP, Tan KB, Moore GB, Jenkins O, Emson PC, Murdock PR: Expression pattern of human P2Y receptor subtypes: a quantitative reverse transcription-polymerase chain reaction study. Biochim Biophys Acta. 2001, 1521 (1-3): 107-119.PubMedView ArticleGoogle Scholar
- Ginzinger DG: Gene quantification using real-time quantitative PCR: an emerging technology hits the mainstream. Exp Hematol. 2002, 30 (6): 503-512. 10.1016/S0301-472X(02)00806-8.PubMedView ArticleGoogle Scholar
- Szabo A, Perou CM, Karaca M, Perreard L, Quackenbush JF, Bernard PS: Statistical modeling for selecting housekeeper genes. Genome Biol. 2004, 5 (8): R59-10.1186/gb-2004-5-8-r59.PubMed CentralPubMedView ArticleGoogle Scholar
- Wilson BJ, Bates GJ, Nicol SM, Gregory DJ, Perkins ND, Fuller-Pace FV: The p68 and p72 DEAD box RNA helicases interact with HDAC1 and repress transcription in a promoter-specific manner. BMC Mol Biol. 2004, 5: 11-10.1186/1471-2199-5-11.PubMed CentralPubMedView ArticleGoogle Scholar
- NetAffx™ Analysis Center. [https://www.affymetrix.com/site/login/login.affx]
- geNorm. [http://medgen.ugent.be/~jvdesomp/genorm]
- Jung M, Spthmann J, Kalbe A, Wankenbauer W, Ebenbichler C, Jung K: Housekeeping gene sets facilitate the search for a suitable reference gene for relative quantification. Biochemica. 2002, 4: 9-11.Google Scholar
- Beer DG, Kardia SL, Huang CC, Giordano TJ, Levin AM, Misek DE, Lin L, Chen G, Gharib TG, Thomas DG, Lizyness ML, Kuick R, Hayasaka S, Taylor JM, Iannettoni MD, Orringer MB, Hanash S: Gene-expression profiles predict survival of patients with lung adenocarcinoma. Nat Med. 2002, 8 (8): 816-824.PubMedGoogle Scholar
- Bhattacharjee A, Richards WG, Staunton J, Li C, Monti S, Vasa P, Ladd C, Beheshti J, Bueno R, Gillette M, Loda M, Weber G, Mark EJ, Lander ES, Wong W, Johnson BE, Golub TR, Sugarbaker DJ, Meyerson M: Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses. Proc Natl Acad Sci U S A. 2001, 98 (24): 13790-13795. 10.1073/pnas.191502998.PubMed CentralPubMedView ArticleGoogle Scholar
- Ploner A, Miller LD, Hall P, Bergh J, Pawitan Y: Correlation test to assess low-level processing of high-density oligonucleotide microarray data. BMC Bioinformatics. 2005, 6 (1): 80-10.1186/1471-2105-6-80.PubMed CentralPubMedView ArticleGoogle Scholar
- Benjamini Y, Hochberg Y: Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Statist Soc B. 1995, 57 (1): 289-300.Google Scholar
- GNF Genome Informatics Applications & Datasets . [http://wombat.gnf.org/index.html]
- Stanford Microarray Database. [http://genome-www5.stanford.edu/]
- Lung Adenocarcinoma. [http://genome-www.stanford.edu/lung_cancer/adeno/index.shtml]
- Liver Cancers. [http://genome-www.stanford.edu/hcc/index.shtml]
- The Human Cell Cycle Data from Hela Cells . [http://genome-www.stanford.edu/Human-CellCycle/Hela/]
- Whitfield ML, Sherlock G, Saldanha AJ, Murray JI, Ball CA, Alexander KE, Matese JC, Perou CM, Hurt MM, Brown PO, Botstein D: Identification of genes periodically expressed in the human cell cycle and their expression in tumors. Mol Biol Cell. 2002, 13 (6): 1977-2000. 10.1091/mbc.02-02-0030..PubMed CentralPubMedView ArticleGoogle Scholar
- Chu YW, Yang PC, Yang SC, Shyu YC, Hendrix MJ, Wu R, Wu CW: Selection of invasive and metastatic subpopulations from a human lung adenocarcinoma cell line. Am J Respir Cell Mol Biol. 1997, 17 (3): 353-360.PubMedView ArticleGoogle Scholar
- Chen JJ, Peck K, Hong TM, Yang SC, Sher YP, Shih JY, Wu R, Cheng JL, Roffler SR, Wu CW, Yang PC: Global analysis of gene expression in invasion by a lung cancer model. Cancer Res. 2001, 61 (13): 5223-5230.PubMedGoogle Scholar
- Su LJ, Hsu SL, Yang JS, Tseng HH, Huang SF, Huang CY: Global gene expression profiling of dimethylnitrosamine-induced liver fibrosis: from pathological and biochemical data to microarray analysis. Gene Expr. 2006, 13 (2): 107-132.PubMedView ArticleGoogle Scholar
- Affymetrix Technical Documentation. [http://www.affymetrix.com/support/technical/manuals.affx]
- Gene Expression Omnibus. [http://www.ncbi.nlm.nih.gov/geo/]
- Hubbell E, Liu WM, Mei R: Robust estimators for expression analysis. Bioinformatics. 2002, 18 (12): 1585-1592. 10.1093/bioinformatics/18.12.1585.PubMedView ArticleGoogle Scholar
- Microarray Suite Software - Support Materials. [http://www.affymetrix.com/support/technical/byproduct.affx?product=mas]
- Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP: Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003, 4 (2): 249-264. 10.1093/biostatistics/4.2.249.PubMedView ArticleGoogle Scholar
- Wu Z, Irizarry RA: Stochastic models inspired by hybridization theory for short oligonucleotide arrays. J Comput Biol. 2005, 12 (6): 882-893. 10.1089/cmb.2005.12.882.PubMedView ArticleGoogle Scholar
- Lehmann EL: Nonparametrics: Statistical Methods Based on Ranks. 1975, San Francisco , Holden-Day, Inc.Google Scholar
- Contag SA, Gostout BS, Clayton AC, Dixon MH, McGovern RM, Calhoun ES: Comparison of gene expression in squamous cell carcinoma and adenocarcinoma of the uterine cervix. Gynecol Oncol. 2004, 95 (3): 610-617. 10.1016/j.ygyno.2004.08.021.PubMedView ArticleGoogle Scholar
- Efron B: Bootstrap methods: another look at the jacknife. Ann Stat,. 1979, 7: 1-26.View ArticleGoogle Scholar
- Davison AC, Hinkley DV: Bootstrap Methods and Their Application. 1997, Cambridge University Press.View ArticleGoogle Scholar
- Lahiri SN: Resampling Methods for Dependent Data. Springer-Verlag, New York. 2003Google Scholar
- R Development Core Team: R: A language and environment for statistical computing. 2006, Vienna, Austria, R Foundation for Statistical Computing