Transcriptomic and proteomic profiling of two porcine tissues using high-throughput technologies
© Hornshøj et al; licensee BioMed Central Ltd. 2009
Received: 14 July 2008
Accepted: 19 January 2009
Published: 19 January 2009
The recent development within high-throughput technologies for expression profiling has allowed for parallel analysis of transcriptomes and proteomes in biological systems such as comparative analysis of transcript and protein levels of tissue regulated genes. Until now, such studies of have only included microarray or short length sequence tags for transcript profiling. Furthermore, most comparisons of transcript and protein levels have been based on absolute expression values from within the same tissue and not relative expression values based on tissue ratios.
Presented here is a novel study of two porcine tissues based on integrative analysis of data from expression profiling of identical samples using cDNA microarray, 454-sequencing and iTRAQ-based proteomics. Sequence homology identified 2.541 unique transcripts that are detectable by both microarray hybridizations and 454-sequencing of 1.2 million cDNA tags. Both transcript-based technologies showed high reproducibility between sample replicates of the same tissue, but the correlation across these two technologies was modest. Thousands of genes being differentially expressed were identified with microarray. Out of the 306 differentially expressed genes, identified by 454-sequencing, 198 (65%) were also found by microarray. The relationship between the regulation of transcript and protein levels was analyzed by integrating iTRAQ-based proteomics data. Protein expression ratios were determined for 354 genes, of which 148 could be mapped to both microarray and 454-sequencing data. A comparison of the expression ratios from the three technologies revealed that differences in transcript and protein levels across heart and muscle tissues are positively correlated.
We show that the reproducibility within cDNA microarray and 454-sequencing is high, but that the agreement across these two technologies is modest. We demonstrate that the regulation of transcript and protein levels across identical tissue samples is positively correlated when the tissue expression ratios are used for comparison. The results presented are of interest in systems biology research in terms of integration and analysis of high-throughput expression data from mammalian tissues.
High-throughput quantitative profiling of transcripts and proteins is a widely used approach for studying biological processes. As a consequence, technologies develop rapidly in order to improve quality, increase the throughput and reduce the cost of expression profiling. Currently, transcript profiling technologies include DNA microarray , Serial Analysis of Gene Expression (SAGE) , Massive Parallel Signature Sequencing (MPSS)  and recently the sequencing technologies from 454 Life Sciences (now Roche)  and Solexa (now Illumina). Hybridization-based microarray technologies have been the dominating method for transcript profiling and are characterized by their ability to globally profile gene expression in large numbers of tissue samples. We recently developed and applied the cDNA microarray technology in porcine studies of gene expression in multiple samples of diseased  and healthy  tissues. Microarray expression profiles are extracted from signal intensities reflecting the amount of hybridized mRNA to spotted DNA whereas the above mentioned sequencing-based technologies provide expression levels that are absolute values computed as the number of transcripts observed for individual genes. At the protein level, introduction of the iTRAQ-tagging approach, has allowed simultaneous quantitative comparison of individual protein levels in multiple tissue samples . Currently the iTRAQ-based proteomics technology is not able to fully characterize entire proteomes , which is a limiting factor in global comparative studies of transcript and protein expression. Comparative analysis of protein expression in pig tissues using iTRAQ-based tagging was recently reported . In comparison to SAGE, MPSS and Solexa, 454-sequencing has increased the sequence length to a minimum of 110 bp. The ability for transcript profiling across multiple tissue samples has been reported for most high-throughput sequencing-based technologies, but has been limited to single tissue profiling for 454-sequencing [10, 11].
As new high-throughput technologies emerge and develop, more comparative expression studies across technologies and across transcriptomes and proteomes have been reported. At the transcript level, these studies have been dominated by comparisons of SAGE data with either Affymetrix short oligonucleotide microarrays [12–22] or cDNA microarrays [13, 14, 16, 21, 23, 24]. A few studies have compared long oligonucleotide microarrays with SAGE [25, 26] and MPSS [25, 27]. In one study, several commercial oligonucleotide-based platforms were compared with MPSS . As demonstrated by many previous studies, computation and evaluation of Pearson's or Spearman correlation coefficients allows for comparison of transcript-based expression profiles across technologies [13, 15, 17–22, 25, 28]. Comparison of transcript and protein profiling has been used in studies of various mammalian tissues including high productivity Chinese Hamster Ovary (CHO) cells , murine stem cell populations  and recombinant NS0 cells . Studies in yeast have also been reported that integrate transcript and protein expression [32–35]. Integrative studies of transcriptomic and proteomic profiles by means of 454-sequencing and iTRAQ-based proteomics have not been reported. The reported levels of correlation in gene expression across technologies have been fairly inconsistent. The observed discrepancies between transcript-based technologies have been suggested to result from errors in SAGE tag-to-gene mapping, errors in microarray probe-to-gene mapping [19, 22] and alternative polyadenylation . The correlation between quantitative proteomics data and microarray data may be affected by alternative splicing . Studies of the correlation between the regulation of transcriptomes and proteomes have mostly used direct comparisons of absolute measurements within single tissues where the differences across genes in translational efficiencies, turn-over rates and half life will have great impact on the level of agreement. On the other hand, it seems appropriate to assume that, when comparing the transcript and protein level of a gene in two tissues, the tissue with the highest transcript level will also be the tissue with the highest protein level. Hence, tissue expression ratios should be more compatible measurements when analyzing the relationship between transcript and protein abundances.
Here we report a comparative study of three high-throughput technologies for multi-sample expression profiling using tissue samples from pig heart and skeletal muscle. We compare transcript profiles from 454-sequencing and cDNA microarray based on relative expression within tissues and expression ratios across tissues. Furthermore, we analyze the relationship between transcript and protein regulation between the two tissues by direct comparison of expression ratios from cDNA microarray, 454-sequencing and iTRAQ-based proteomics.
Expression profiles for genes overlapping between technologies
Summary of 454-sequencing and overlap to microarray cDNAs via pig UniGene sequence IDs
Unique UniGene sequence count with array overlap
Total, N ≥ 1
Reproducibility of gene expression within and across 454-sequencing and cDNA microarray
Level of agreement between 454-sequencing and cDNA microarray in detecting differentially expressed genes
Correlation between heart-muscle expression ratios across transcript-based technologies and iTRAQ-based proteomics
In this study, cDNA microarray and 454-sequencing have been compared for transcript expression profiling in porcine heart and skeletal muscle tissues. Protein abundance data for identical tissue samples was generated using iTRAQ-based proteomics to analyze the relationship within genes between transcript and protein ratios across tissues.
A total of 2.541 genes were detectable by both transcript-based technologies. The degree of overlapping genes detectable by both cDNA microarray and 454-sequencing is affected by the nature of the two technologies. With cDNA microarray, spot intensities are provided for all genes printed on the microarray slides, regardless of whether the genes are expressed or not. In case of 454-sequencing, expression of a gene can be measured if the gene is transcribed and if the sequencing depth of the experiment is sufficiently high. For example, low expressed genes will require that more sequences are generated in order to be captured. Thus, the degree of overlap will depend largely on the genes being expressed in the particular tissues being examined. In order to compare t-tests for differential expression with cDNA microarray we restricted the 454-sequencing data set to genes with expression values in all six tissue samples, which reduced the number of genes and thereby the number of overlapping genes. This restriction may have removed informative data, but was necessary to allow comparison.
The reproducibility in profiling gene expression across replicates of the same tissue sample is high for both transcript-based technologies with correlation values at 0.85 for 454-sequencing and 0.98 for cDNA microarray. The expression correlation between heart and skeletal muscle within cDNA microarray is relatively high (0.67) in comparison to the corresponding correlation observed for 454-sequencing (0.23), which suggests a larger tissue difference. We speculate that the main reason for the difference between these two transcript-based technologies is the differences in the dynamic range of the expression profiles. Thus, the distribution of the RA values from 454-sequencing is considerably wider than that observed for microarray, indicating that sensitivity for detecting expression difference across tissue samples is highest for 454-sequencing. Whether the larger number of genes detected to be different across skeletal muscle and heart by 454-sequencing is reflecting real biological differences between these two tissues is uncertain. The correlation values across 454-sequencing and microarray at 0.25 and 0.51 for heart and muscle respectively was relatively low, but considerably higher than the correlation value at 0.13 across the two transcript-based technologies and across the two tissues. This means that even though the technologies are quite different they still measure the respective expression profiles within heart and muscle to be most similar.
The number of differentially expressed genes identified in relation to the number of tested genes was much higher for cDNA microarray (47%; 5.888 out of 12.563) than for 454-sequencing (10%; 306 out of 2.954). Because it is not the same set of genes being tested, it will of course not lead to the same list of identified differentially expressed genes. However, microarray tends to predict more genes that are differentially expressed. Some might be false positives, even though P-values were corrected for multiple testing. If expression ratios across any given two tissues are close to one, they are vulnerable to random shifts between up-regulated, not-regulated and down-regulated genes. This has previously been suggested as a reason for discrepancy between cDNA- and oligonucleotide-based microarray platforms in expression ratios . Cardiac and skeletal muscles represent two different tissues, although they can both be classified as muscle type tissues. The fact that only about 10% of the monitored genes with 454-sequencing are differentially expressed may not be surprising due to the similarity of these two tissues. Thus, genes that are related to basic functions in muscle tissue cells would be expected to be expressed in both tissues. A relatively low number of differentially expressed genes between similar tissue samples has previously been observed in a study of rhabdomyosarcoma, fetal and skeletal muscle tissues, where 403 tags out of 41084 unique tags were identified as differentially expressed, although this study was based on the SAGE technology . The fact that cDNA microarray has a lower dynamic range suggests that this technology is less sensitive to minor differences in gene expression across tissues than the sequencing-based approach. However, microarray detects more genes than 454-sequencing that are differentially expressed across cardiac and skeletal muscle. One explanation may be that if the variation in expression across sample replicates within tissue is considerably lower than the variation across tissues, significantly regulated genes can still be detected. This is supported by the high reproducibility observed for microarray within tissues in figure 2. Of the 306 genes identified as being differentially expressed with 454-sequencing, 198 was also identified by microarray and 160 were regulated in the same direction. Out of 198 genes found to be significantly regulated between the two tissues, 38 genes are regulated in opposite direction with the two transcript-based technologies. This may also be due to random shifts between positive and negative values in the log2 ratio in either of the two technologies. Other reasons for the disagreement in detection of differential expression could be gene mapping errors or alternative splicing events.
To investigate the relationship between the regulation of transcript and protein levels across tissues, the expression ratios across heart and skeletal muscle from microarray, 454-sequencing and iTRAQ-based proteomics were compared. Transcript and protein ratios across heart and skeletal muscle were shown to be positively correlated for 148 genes detectable by all three technologies. The correlation appear to be slightly more positive for the comparison with 454-sequencing (0.53) than for the comparison with cDNA microarray (0.49). The slope is 0.6 for cDNA microarray and 1.4 for 454-sequencing when ratios are compared against iTRAQ-based ratios. The positive correlation between transcript and protein ratios suggests that for most genes, a relative tissue difference across a given tissue pair in transcript level leads to a similar difference in the protein level. A list of all 148 genes with expression ratios for the three technologies and indications of concordance was generated. Genes, where the transcript-protein relationship deviates from this linearity might be interesting to study further, since they could be examples of differential regulation of either translation or mRNA- or protein-turnover among the tissues. Alternatively, translation of mRNAs from genes that generate high levels of transcripts might be limited by the availability of RNA-binding factor or the capacity of the translational machinery. Deviations may also result from gene annotation errors [19, 22] or alternative splicing events taking place differentially across the two tissues, which are not equally detected by the transcript-based technologies and the iTRAQ-based technology . Discrepancies arising from alternative splicing may potentially be verified with real time PCR. Future improvement of the proteomics-based technologies should also increase the depth of comparative studies of transcripts and proteins profiles. While the analysis described here allows for the detection of genes with alternative regulation in different tissues, further experiments are needed in order to understand the biological basis of these differences.
In this study we have analyzed the reproducibility of expression data within and between microarray and 454-sequencing technologies. Furthermore, we have integrated expression data from both transcriptomic and proteomic profiling to analyze the gene regulation across two porcine tissues by comparing tissue expression ratios of transcripts and proteins. Both transcript-based technologies displayed a high degree of reproducibility within technology, but the reproducibility across these two technologies was modest. The majority of the differentially regulated genes identified by 454-sequencing was also found by the cDNA microarray platform. Most interesting was the comparison of data from both transcript-based technologies with relative expression values from iTRAQ-based proteomics. Integrative analysis revealed that the regulation of transcript and protein levels across the two tissues is positively correlated for most genes using tissue expression ratios for comparison. Some genes without transcript-protein concordance were identified, which may arise from annotation errors or differential regulation of translation, turnover or alternative splicing. The results presented here should be of high importance for integration and analysis of high-throughput expression data, in particular for studies of the regulation of transcript and protein abundances in mammalian tissues.
Tissue samples of heart and skeletal muscle
Tissue samples of heart (HEA) and skeletal muscle (Longissimus dorsi; LDO) were prepared by pooling equal amounts of tissue sampled from five healthy Hampshire gilts at age four to six months. Each pool was divided into six sub-samples, three for each tissue named HEA1, HEA2, HEA3, LDO1, LDO2 and LDO3. The exact same six tissue sub-samples were used for expression profiling with cDNA microarray, 454-sequencing and iTRAQ-based proteomics. A reference sample for the cDNA microarray experiment and iTRAQ-based proteomics was constructed by combining equal amounts of tissue from the six sub-samples.
Total-RNA extractions were carried out from each sample using the RNeasy Maxi Kit from Qiagen. Alexa Flour-labelled cDNA was synthesized from 20 μg of total-RNA using SuperScript Plus Direct cDNA Labeling System from Invitrogen. The reference sample was labelled with Alexa 555 and the individual tissue samples were labelled with Alexa 647. Each of the six labeled tissue samples was co-hybridized with the labelled reference sample to three 27 k pig cDNA microarray slides representing approximately 20 k genes. Microarray cDNA platform development was based on a large EST sequence resource established as part of the Sino-Danish Pig Genome Sequencing Project [40, 41]. Detailed description of the pig cDNA microarray platform can be found at NCBI's Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo)  using the accession ID GPL3585. Following hybridization, washing and drying, the slides were scanned and the output images were analyzed. Using the limma package  from Bioconductor http://www.bioconductor.org, the log2-transformed median intensity ratios of Alexa-647 to Alexa-555 were normalized within-slide using the printtip-loess method . The raw and normalized cDNA microarray gene expression data set was submitted to GEO at NCBI http://www.ncbi.nlm.nih.gov/geo and given the accession ID GSE10122.
Preparation of cDNA libraries and 454-sequencing
From the same total-RNA extraction batches used for the microarray experiment, poly A+ RNA was purified from ~200 μg total-RNA using the Oligotex mRNA Mini Kit from Qiagen. The poly A+ RNA isolates were used as templates in synthesis of high quality double-stranded cDNA with oligo(dT)12–18-priming of the first strand cDNA (SuperScriptTM Doubled-Stranded cDNA Synthesis Kit from Invitrogen). The final purification of the synthesized cDNA was performed on MinElute PCR Purification column from Qiagen, and an average cDNA fragment size of 300–400 bp was obtained by applying 45 psi (= 3 bar) of nitrogen for 50–60 seconds on 2.7–4.7 μg starting material using a nebulizer. Subsequently, the distribution of fragments was profiled on a BioAnalyzer DNA 1000 LabChip (Agilent Technologies). As part of the library preparation, the ends of the sheared cDNA fragments were polished by treatment with T4 DNA Polymerase and T4 Polynucleotide Kinase prior to ligation of sequencing adaptors as previously described. The fragments were immobilized on streptavidin beads, following nick repair with Bst DNA Polymerase, and the sstDNA libraries were eluted by alkali denaturation. Finally, the qualities and quantities of the libraries were assayed on Bioanalyzer RNA 6000 Pico LabChip (Agilent Technologies). A titration run, based on DNA/bead ratios of 4 and 16 copies per bead (cpb) in the emulsion-based clonal amplification, revealed the optimal experimental set-up for the large sequencing run using enriched beads. Each of the six libraries was sequenced on a full 70 × 75 PicoTiterPlate (PTP). In total 1.253.361 sequences were generated and submitted to the Short Read Archive (SRA) of NCBI and assigned the accession ID SRA000267.
Eight samples (including two reference samples) were processed in parallel. Total protein fraction was purified from 200 mg tissues as previously described . The supernatant was stored at 80°C until use. Tryptic digestion (100 μg protein from each sample) and tagging with iTRAQ reagents (Applied Biosystems, Forster City, CA, USA) was performed as previously described . The two reference samples were labelled with mass-tag 114, and then pooled and split for normalization. Samples were combined in a 1:1:1:1 ratios into two parallel 4-plexed samples, with each 4-plex containing a common reference sample, as well as both heart and skeletal muscle samples. Inclusion of a common reference sample in every 4-plex sample allowed comparison of the protein expression ratios across the different 4-plexes. Fifty (50) μg of iTRAQ-labelled peptides were separated by 2D-HPLC (Agilent Technologies, Palo Alto, CA, USA) according to detailed descriptions . The eluted peptides were sprayed through a nanospray needle (Fused Silica Emitters, OD 360 μm, ID 75 μm, Proxeon Biosystems, Odense, Denmark) directly into the Q-star XL mass spectrometer (Applied Biosystems, Forster City, CA, USA). The raw data files were searched with the Protein Pilot 1.0 software (Applied Biosystems) using the ParagonTM algorithm for protein grouping and confidence scoring, and searched against a database of proteins predicted from pig UniGene release 30 downloaded at trEST . There was no processing (e.g. smoothing) of the raw data files prior to database searching. The database allowed for iTRAQ reagent labels at N-terminal residues, internal K and Y residues, and MMTS-labelled cysteine as fixed modifications, deamidation, O-phosphorylation (STY) and oxidation (M) as variable modifications and one missed cleavage. Confidence of protein identification was selected according to a 95% confidence and a minimum of two peptides identified per protein.
Identification of IDs for genes detected across expression technologies
The microarray cDNA gene expression data and the 454-sequencing data was linked together by mapping all microarray cDNA sequences and 454-based sequences to NCBI's pig UniGene database release 28 and then linking the data by their shared UniGene IDs. The sequences were mapped to UniGene IDs by BlastN and collecting only IDs for the first hit with score at or above 100. For correct microarray cDNA mappings it was required for each UniGene Id that the corresponding UniGene sequence was able to identify the exact same microarray cDNA sequence as the first hit when it was compared back to a database of all microarray cDNA sequences. In total, 12.563 out of 26.877 microarray cDNA sequences were mapped to 12.563 UniGene IDs and 647.093 out of 1.253.361 454-sequences were mapped to 18.624 UniGene IDs. In total, 2.541 genes were identified with expression profiles from both cDNA microarray and 454-sequencing by using the overlapping UniGene IDs to merge the two data sets together. For iTRAQ-based proteomics, a database of predicted pig proteins from trEST based on translations of the latest release (30) of UniGene sequences was downloaded and searched . Using this database, proteins corresponding to 356 UniGene IDs were identified and 148 of these could be directly linked to 454-sequencing, whereas 202 IDs could be linked to the microarray cDNAs. The total overlap between all three technologies consisted of 148 UniGene IDs.
Computation of relative expression values for 454-sequencing and microarray
The expression profiles from 454-sequencing was achieved by counting the number of transcripts per gene, in this case represented by the number of BlastN query 454-sequences for each target UniGene ID in each of the six tissue samples. The minimum number of sequence counts in a tissue sample was set to one resulting in a total number of 2.954 UniGene IDs for which gene expression could be detected by 454-sequencing in all six samples. To calculate within-tissue relative expression profiles for each of the six tissue samples, taken into account the variation in total raw sequence counts, Relative Abundance (RA) values were computed for each gene (UniGene ID) in each tissue sample. The RA value for a given gene in a given tissue sample was computed as the 454-sequence count for that gene divided by the total 454-sequence count for the 2.541 genes in the whole tissue sample. To be able to make direct comparison of 454-based RA values with microarray-based expression profiles for the six tissue samples, corresponding microarray RA values were also computed using the same approach. The microarray RA value for a given gene was computed as previously proposed [6, 37] based on the raw median signal intensity for Alexa 647 divided by the total signal intensity for Alexa 647 in that tissue sample. Three microarray slides were used per tissue sample so an average RA value was calculated for each tissue sample and used for comparison. Computation of RA values was based the 2.541 genes detected by both 454-sequencing and cDNA microarray.
Computation of expression correlation values
For comparison of relative expression profiles across 454-sequencing and cDNA microarray we computed tissue sample correlations for examining the reproducibility within and across technologies. Twelve RA vectors, six RA vectors from 454-sequencing and the six RA vectors from cDNA microarray, corresponding to the six tissue samples were paired in 12 × 12 = 144 possible combinations and the Pearson's correlation coefficient between the vectors in each combination was computed. The correlation values were grouped into nine bins corresponding to the three tissue comparisons (heart versus heart, heart versus skeletal muscle and skeletal muscle versus skeletal muscle) and the three technology comparisons (within 454-sequencing, within cDNA microarray and across 454-sequencing/cDNA microarray). Average correlations were computed for these nine bins and used to evaluate the reproducibility within and across 454-sequencing and cDNA microarray. For comparing across-tissue mRNA expression ratios from the two transcript-based technologies against iTRAQ-based protein ratios we first prepared log2-transformed values of the expression ratios between heart and skeletal muscle for all three technologies. These values were organized in three vectors and pair-wise Pearson's correlation coefficients were calculated for the two comparisons 454-sequencing versus iTRAQ-based proteomics and cDNA microarray versus iTRAQ-based proteomics. In the comparison of expression ratios across transcripts and proteins a straight line was fitted for the data points described by the standard equation Ax+By+C = 0, where -A/B is the slope and C is the intersection between the line and the y-axis. Distances for all data points (m, n) to the fitted line were calculated based on the standard formula d = |Am+Bn+C|/√(A2+B2). These distances were used to evaluate the concordance of genes across transcript and protein ratios.
Identification of differentially expressed genes
Statistical analysis was carried out in the R computing environment version 2.6.0 using the limma package  version 2.11.14 from Bioconductor. To identify differentially expressed genes between heart and skeletal muscle, normalized log2-transformed ratio between tissue and reference were used for cDNA microarray and log2-transformed RA values for 454-sequencing. The empirical Bayes method was applied and the P-values were adjusted for multiple testing using the false discovery rate ("fdr"). A 5% significance level (adjusted P-values ≤ 0.05) threshold was applied for differentially expressed genes.
The authors wish to acknowledge the support from the Danish Meat Association and the Danish Ministry of Agriculture and Fisheries. The construction of the genome-wide cDNA microarrays was made possible by the work of the Sino-Danish Porcine Genome Consortium.
- Schena M, Shalon D, Davis RW, Brown PO: Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science. 1995, 270: 467-470. 10.1126/science.270.5235.467.View ArticlePubMedGoogle Scholar
- Velculescu VE, Zhang L, Vogelstein B, Kinzler KW: Serial analysis of gene expression. Science. 1995, 270: 484-487. 10.1126/science.270.5235.484.View ArticlePubMedGoogle Scholar
- Brenner S, Johnson M, Bridgham J, Golda G, Lloyd DH, Johnson D, et al: Gene expression analysis by massively parallel signature sequencing (MPSS) on microbead arrays. Nat Biotechnol. 2000, 18: 630-634. 10.1038/76469.View ArticlePubMedGoogle Scholar
- Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, et al: Genome sequencing in microfabricated high-density picolitre reactors. Nature. 2005, 437: 376-380.PubMed CentralPubMedGoogle Scholar
- Hedegaard J, Skovgaard K, Mortensen S, Sorensen P, Jensen TK, Hornshoj H, et al: Molecular characterisation of the early response in pigs to experimental infection with Actinobacillus pleuropneumoniae using cDNA microarrays. Acta Vet Scand. 2007, 49: 11-10.1186/1751-0147-49-11.PubMed CentralView ArticlePubMedGoogle Scholar
- Hornshoj H, Conley LN, Hedegaard J, Sorensen P, Panitz F, Bendixen C: Microarray Expression Profiles of 20.000 Genes across 23 Healthy Porcine Tissues. PLoS ONE. 2007, 2: e1203-10.1371/journal.pone.0001203.PubMed CentralView ArticlePubMedGoogle Scholar
- Wiese S, Reidegeld KA, Meyer HE, Warscheid B: Protein labeling by iTRAQ: a new tool for quantitative mass spectrometry in proteome research. Proteomics. 2007, 7: 340-350. 10.1002/pmic.200600422.View ArticlePubMedGoogle Scholar
- Bitton DA, Okoniewski MJ, Connolly Y, Miller CJ: Exon level integration of proteomics and microarray data. BMC Bioinformatics. 2008, 9: 118-10.1186/1471-2105-9-118.PubMed CentralView ArticlePubMedGoogle Scholar
- Danielsen M, Hornshoj H, Siggers RH, Jensen BB, van Kessel AG, Bendixen E: Effects of bacterial colonization on the porcine intestinal proteome. J Proteome Res. 2007, 6: 2596-2604. 10.1021/pr070038b.View ArticlePubMedGoogle Scholar
- Cheung F, Haas BJ, Goldberg SM, May GD, Xiao Y, Town CD: Sequencing Medicago truncatula expressed sequenced tags using 454 Life Sciences technology. BMC Genomics. 2006, 7: 272-10.1186/1471-2164-7-272.PubMed CentralView ArticlePubMedGoogle Scholar
- Emrich SJ, Barbazuk WB, Li L, Schnable PS: Gene discovery and annotation using LCM-454 transcriptome sequencing. Genome Res. 2007, 17: 69-73. 10.1101/gr.5145806.PubMed CentralView ArticlePubMedGoogle Scholar
- Evans SJ, Datson NA, Kabbaj M, Thompson RC, Vreugdenhil E, De Kloet ER, et al: Evaluation of Affymetrix Gene Chip sensitivity in rat hippocampal tissue using SAGE analysis. Serial Analysis of Gene Expression. Eur J Neurosci. 2002, 16: 409-413. 10.1046/j.1460-9568.2002.02097.x.View ArticlePubMedGoogle Scholar
- Griffith OL, Pleasance ED, Fulton DL, Oveisi M, Ester M, Siddiqui AS, et al: Assessment and integration of publicly available SAGE, cDNA microarray, and oligonucleotide microarray expression data for global coexpression analyses. Genomics. 2005, 86: 476-488. 10.1016/j.ygeno.2005.06.009.View ArticlePubMedGoogle Scholar
- Hackam AS, Qian J, Liu D, Gunatilaka T, Farkas RH, Chowers I, et al: Comparative gene expression analysis of murine retina and brain. Mol Vis. 2004, 10: 637-649.PubMedGoogle Scholar
- Haverty PM, Hsiao LL, Gullans SR, Hansen U, Weng Z: Limited agreement among three global gene expression methods highlights the requirement for non-global validation. Bioinformatics. 2004, 20: 3431-3441. 10.1093/bioinformatics/bth421.View ArticlePubMedGoogle Scholar
- Iacobuzio-Donahue CA, Ashfaq R, Maitra A, Adsay NV, Shen-Ong GL, Berg K, et al: Highly expressed genes in pancreatic ductal adenocarcinomas: a comprehensive characterization and comparison of the transcription profiles obtained from three major technologies. Cancer Res. 2003, 63: 8614-8622.PubMedGoogle Scholar
- Ibrahim AF, Hedley PE, Cardle L, Kruger W, Marshall DF, Muehlbauer GJ, et al: A comparative analysis of transcript abundance using SAGE and Affymetrix arrays. Funct Integr Genomics. 2005, 5: 163-174. 10.1007/s10142-005-0135-4.View ArticlePubMedGoogle Scholar
- Ishii M, Hashimoto S, Tsutsumi S, Wada Y, Matsushima K, Kodama T, et al: Direct comparison of GeneChip and SAGE on the quantitative accuracy in transcript profiling analysis. Genomics. 2000, 68: 136-143. 10.1006/geno.2000.6284.View ArticlePubMedGoogle Scholar
- Li S, Li YH, Wei T, Su EW, Duffin K, Liao B: Too much data, but little inter-changeability: a lesson learned from mining public data on tissue specificity of gene expression. Biol Direct. 2006, 1: 33-10.1186/1745-6150-1-33.PubMed CentralView ArticlePubMedGoogle Scholar
- Lu J, Lal A, Merriman B, Nelson S, Riggins G: A comparison of gene expression profiles produced by SAGE, long SAGE, and oligonucleotide chips. Genomics. 2004, 84: 631-636. 10.1016/j.ygeno.2004.06.014.View ArticlePubMedGoogle Scholar
- Romualdi C, De Pitta C, Tombolan L, Bortoluzzi S, Sartori F, Rosolen A, et al: Defining the gene expression signature of rhabdomyosarcoma by meta-analysis. BMC Genomics. 2006, 7: 287-10.1186/1471-2164-7-287.PubMed CentralView ArticlePubMedGoogle Scholar
- van Ruissen F, Ruijter JM, Schaaf GJ, Asgharnegad L, Zwijnenburg DA, Kool M, et al: Evaluation of the similarity of gene expression data estimated with SAGE and Affymetrix GeneChips. BMC Genomics. 2005, 6: 91-10.1186/1471-2164-6-91.PubMed CentralView ArticlePubMedGoogle Scholar
- Husson H, Manavalan P, Akmaev VR, Russo RJ, Cook B, Richards B, et al: New insights into ADPKD molecular pathways using combination of SAGE and microarray technologies. Genomics. 2004, 84: 497-510. 10.1016/j.ygeno.2004.03.009.View ArticlePubMedGoogle Scholar
- Nacht M, Ferguson AT, Zhang W, Petroziello JM, Cook BP, Gao YH, et al: Combining serial analysis of gene expression and array technologies to identify genes differentially expressed in breast cancer. Cancer Res. 1999, 59: 5464-5470.PubMedGoogle Scholar
- Gowda M, Venu RC, Raghupathy MB, Nobuta K, Li H, Wing R, et al: Deep and comparative analysis of the mycelium and appressorium transcriptomes of Magnaporthe grisea using MPSS, RL-SAGE, and oligoarray methods. BMC Genomics. 2006, 7: 310-10.1186/1471-2164-7-310.PubMed CentralView ArticlePubMedGoogle Scholar
- Kim HL: Comparison of oligonucleotide-microarray and serial analysis of gene expression (SAGE) in transcript profiling analysis of megakaryocytes derived from CD34+ cells. Exp Mol Med. 2003, 35: 460-466.View ArticlePubMedGoogle Scholar
- Chen J, Agrawal V, Rattray M, West MA, St Clair DA, Michelmore RW, et al: A comparison of microarray and MPSS technology platforms for expression analysis of Arabidopsis. BMC Genomics. 2007, 8: 414-10.1186/1471-2164-8-414.PubMed CentralView ArticlePubMedGoogle Scholar
- Liu F, Jenssen TK, Trimarchi J, Punzo C, Cepko CL, Ohno-Machado L, et al: Comparison of hybridization-based and sequencing-based gene expression technologies on biological replicates. BMC Genomics. 2007, 8: 153-10.1186/1471-2164-8-153.PubMed CentralView ArticlePubMedGoogle Scholar
- Nissom PM, Sanny A, Kok YJ, Hiang YT, Chuah SH, Shing TK, et al: Transcriptome and proteome profiling to understanding the biology of high productivity CHO cells. Mol Biotechnol. 2006, 34: 125-140. 10.1385/MB:34:2:125.View ArticlePubMedGoogle Scholar
- Unwin RD, Smith DL, Blinco D, Wilson CL, Miller CJ, Evans CA, et al: Quantitative proteomics reveals posttranslational control as a regulatory factor in primary hematopoietic stem cells. Blood. 2006, 107: 4687-4694. 10.1182/blood-2005-12-4995.View ArticlePubMedGoogle Scholar
- Seth G, Philp RJ, Lau A, Jiun KY, Yap M, Hu WS: Molecular portrait of high productivity in recombinant NS0 cells. Biotechnol Bioeng. 2007, 97: 933-951. 10.1002/bit.21234.View ArticlePubMedGoogle Scholar
- Futcher B, Latter GI, Monardo P, McLaughlin CS, Garrels JI: A sampling of the yeast proteome. Mol Cell Biol. 1999, 19: 7357-7368.PubMed CentralView ArticlePubMedGoogle Scholar
- Griffin TJ, Gygi SP, Ideker T, Rist B, Eng J, Hood L, et al: Complementary profiling of gene expression at the transcriptome and proteome levels in Saccharomyces cerevisiae. Mol Cell Proteomics. 2002, 1: 323-333. 10.1074/mcp.M200001-MCP200.View ArticlePubMedGoogle Scholar
- Gygi SP, Rochon Y, Franza BR, Aebersold R: Correlation between protein and mRNA abundance in yeast. Mol Cell Biol. 1999, 19: 1720-1730.PubMed CentralView ArticlePubMedGoogle Scholar
- Washburn MP, Koller A, Oshiro G, Ulaszek RR, Plouffe D, Deciu C, et al: Protein pathway and complex clustering of correlated mRNA and protein expression analyses in Saccharomyces cerevisiae. Proc Natl Acad Sci USA. 2003, 100: 3107-3112. 10.1073/pnas.0634629100.PubMed CentralView ArticlePubMedGoogle Scholar
- Sperisen P, Iseli C, Pagni M, Stevenson BJ, Bucher P, Jongeneel CV: trome, trEST and trGEN: databases of predicted protein sequences. Nucleic Acids Res. 2004, 32: D509-D511. 10.1093/nar/gkh067.PubMed CentralView ArticlePubMedGoogle Scholar
- Liao BY, Zhang J: Evolutionary conservation of expression profiles between human and mouse orthologous genes. Mol Biol Evol. 2006, 23: 530-540. 10.1093/molbev/msj054.View ArticlePubMedGoogle Scholar
- Lee M, Xiang CC, Trent JM, Bittner ML: Performance characteristics of 65-mer oligonucleotide microarrays. Anal Biochem. 2007, 368: 70-78. 10.1016/j.ab.2007.05.010.PubMed CentralView ArticlePubMedGoogle Scholar
- Schaaf GJ, Ruijter JM, van Ruissen F, Zwijnenburg DA, Waaijer R, Valentijn LJ, et al: Full transcriptome analysis of rhabdomyosarcoma, normal, and fetal skeletal muscle: statistical comparison of multiple SAGE libraries. FASEB J. 2005, 19: 404-406.PubMedGoogle Scholar
- Gorodkin J, Cirera S, Hedegaard J, Gilchrist MJ, Panitz F, Jorgensen C, et al: Porcine transcriptome analysis based on 97 non-normalized cDNA libraries and assembly of 1,021,891 expressed sequence tags. Genome Biol. 2007, 8: R45-10.1186/gb-2007-8-4-r45.PubMed CentralView ArticlePubMedGoogle Scholar
- Jorgensen FG, Hobolth A, Hornshoj H, Bendixen C, Fredholm M, Schierup MH: Comparative analysis of protein coding sequences from human, mouse and the domesticated pig. BMC Biol. 2005, 3: 2-10.1186/1741-7007-3-2.PubMed CentralView ArticlePubMedGoogle Scholar
- Barrett T, Troup DB, Wilhite SE, Ledoux P, Rudnev D, Evangelista C, et al: NCBI GEO: mining tens of millions of expression profiles–database and tools update. Nucleic Acids Res. 2007, 35: D760-D765. 10.1093/nar/gkl887.PubMed CentralView ArticlePubMedGoogle Scholar
- Smyth GK: Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004, 3: Article 3-Google Scholar
- Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J, et al: Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res. 2002, 30: e15-10.1093/nar/30.4.e15.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.