The importance of tissue specificity for RNA-seq: highlighting the errors of composite structure extractions
© Johnson et al.; licensee BioMed Central Ltd. 2013
Received: 10 May 2013
Accepted: 17 August 2013
Published: 28 August 2013
A composite biological structure, such as an insect head or abdomen, contains many internal structures with distinct functions. Composite structures are often used in RNA-seq studies, though it is unclear how expression of the same gene in different tissues and structures within the same structure affects the measurement (or even utility) of the resulting patterns of gene expression. Here we determine how complex composite tissue structure affects measures of gene expression using RNA-seq.
We focus on two structures in the honey bee (the sting gland and digestive tract) both contained within one larger structure, the whole abdomen. For each of the three structures, we used RNA-seq to identify differentially expressed genes between two developmental stages, nurse bees and foragers. Based on RNA-seq for each structure-specific extraction, we found that RNA-seq with composite structures leads to many false negatives (genes strongly differentially expressed in particular structures which are not found to be differentially expressed within the composite structure). We also found a significant number of genes with one pattern of differential expression in the tissue-specific extraction, and the opposite in the composite extraction, suggesting multiple signals from such genes within the composite structure. We found these patterns for different classes of genes including transcription factors.
Many RNA-seq studies currently use composite extractions, and even whole insect extractions, when tissue and structure specific extractions are possible. This is due to the logistical difficultly of micro-dissection and unawareness of the potential errors associated with composite extractions. The present study suggests that RNA-seq studies of composite structures are prone to false negatives and difficult to interpret positive signals for genes with variable patterns of local expression. In general, our results suggest that RNA-seq on large composite structures should be avoided unless it is possible to demonstrate that the effects shown here do not exist for the genes of interest.
KeywordsRNA-seq Tissue specificity Genomics
RNA-seq is revolutionizing the study of gene expression. RNA-seq has been shown to be quantitatively accurate over a larger range of expression levels than previous methods, such as microarrays, while also being more effective at identifying genes that show low expression levels [1–7]. RNA-seq is also leading to major breakthroughs in the study of functional RNAs and gene regulation [7–10]. Studies of large scale patterns of expression of microRNAs have shown the fundamental roles these molecules play in regulating transcripts, while studies of long non-coding RNAs have revealed an unforeseen depth of functional roles for these genes in gene regulation and epigenetics [11–16].
While the technical and experimental logistics of how best to use RNA-seq are being addressed in a variety of contexts [17–20], one question that has received relatively little attention is the extent to which structure specific extractions are necessary for an accurate determination of gene expression. We use the terms “structure” and “organ” interchangeably, as we are referring to biological structures with distinct functions within a larger whole. In insects, for example, structures (organs) such as the fat body and segmental ganglia are contained within the abdominal body segment. We refer to large structures, such as the abdomen, with many internal structures as “composite structures”. Essentially, although efforts are under way to develop procedures for effective isolation of structures for RNA extraction (even particular cell types within complex tissue) [21–25], there is little experimental support for composite structure extractions actually being difficult to interpret or prone to error. This has led to widespread use of composite structures in RNA-seq studies of small organisms such as insects [26–31].
Although many studies are using composite extractions, there are potential problems with this approach. First, if genes are expressed in many different structures within the composite structure, then signals of gene expression from the different organs may interfere with one another. For example, if the structure of interest is small relative to the size of the rest of the structure (a gland within a whole larva, for example), then a strong difference in gene expression within the gland may be washed out by different patterns of expression elsewhere. Second, incomplete homogenization of the composite structure during extraction may lead to little tissue from particular structures actually being extracted. Hence, there are straightforward reasons to suspect that RNA-seq from composite-structures such as body segments may cause false negatives and difficult to interpret patterns of differential expression.
Here we explore the necessity of using structure (organ) specific extractions for RNA-seq using three honey bee structures. We focus on two structures, the sting gland and the digestive tract, that occur within one body segment, the whole abdomen. The sting gland is a relatively small structure within the larger composite structure that can be predicted to expresses a large number of specialized genes. These include the many venom proteins that make up honey bee venom, along with the enzymatic machinery used to produce and modify these proteins [32, 33]. Exploring patterns of differential gene expression in sting gland specific extractions and comparing them to patterns of expression in the whole abdomen extractions can address the question of how much error can be expected when a composite extraction is conducted but the focal tissue is a small specialized structure within it. The digestive tract is a large structure not thought to be highly specialized (though many digestive enzymes undoubtedly show tissue specific expression patterns). By comparing patterns of gene expression in digestive tract specific extractions to patterns in the whole abdomen we can explore how much error is to be expected when the focal structure is a sizeable portion of the composite structure. Both comparisons will shed light on whether contrasting patterns of gene expression in different structures and false negatives in general are a problem for RNA-seq using composite structures.
Results and discussion
Transcriptome characterizations based on structure-specific or composite extractions
For genes that were present in the whole abdomen, but not present in the sting gland, the pattern was different than that for genes present in the sting gland but not the abdomen (Figure 2). Here the mode class of genes had high expression in the abdomen (not low expression as for genes present in the sting gland and missing from the abdomen). The genes present in the abdomen, but missing from the sting gland transcriptome, are therefore presumably highly expressed specialized genes not expressed in the sting gland. The sting gland itself, for example, has many venom proteins with high expression levels that presumably would not be expressed in other structures in the abdomen. Genes with equally specialized functions are likely found in other structures as well.
Figure 3 shows the results for the same type of analysis for the digestive tract and the whole abdomen (genes and expression values in Additional file 2: Table S2). The pattern here is quite different from that with respect to the sting gland versus the whole abdomen, presumably due to the sting gland being a tiny structure within the whole abdomen, and the digestive tract being quite a large structure. In this case, genes that were present in the digestive tract, but missing from the whole abdomen, are all genes with low expression levels (there are no highly expressed genes in the digestive tract missing from the abdomen). Hence, the issue of missing key genes may not be a problem for large focal structures within composite structures, as the composite extraction contained all the genes except those with very low expression in the focal tissue. This is true for transcriptome characterization, but as we will see in the next section does not hold for RNA-seq. The opposite comparison, genes that were found in the whole abdomen but missing from the digestive tract showed a pattern more like that found for the sting gland versus the abdomen comparison. In short, many genes with high expression levels in the abdomen were not found in the digestive tract presumably because they exhibit specialized functions elsewhere in the abdomen.
Errors in diagnosing differentially expressed genes due to composite structure extraction
The major hypothesized problem associated with composite structure extraction and RNA-seq is contrasting gene expression patterns for the same genes in different structures interfering with a determination of differential expression. For example, when a researcher conducts an RNA-seq study on a whole body segment and finds that a gene is not differentially expressed, does it mean that it is not differentially expressed anywhere in the body segment or does it mean that there is no overall difference in expression level when summing the inputs from all the internal organs? Essentially, it is easy to imagine that a gene that is strongly differentially expressed in one organ may be determined to be non-differentially expressed in a much larger composite structure due to expression of the gene elsewhere washing out the signal from the small structure. Hence, there could be a strong false negative problem with composite extractions. Likewise, it is easy to imagine that a gene with contrasting differential expression patterns in different structures may give a strong signal of being differentially expressed in one direction that is representative of a strong signal from one structure washing out several opposite signals from other structures. Hence, although a determination of differential expression in a composite extraction is not a false positive, it can be difficult to interpret.
With respect to genes potentially showing the opposite pattern of differential expression in a composite structure relative to a smaller structure within it, 25 such genes were identified in the sting gland, and 7 in the digestive tract (Figure 4, very similar results for DESeg and EdgeR shown in Additional file 3: Figure S1). These contrasting signals of differential expression between tissue specific and composite tissue analyses suggests there may be a rich pattern of variation of function across tissues for such genes. This would make interpreting expression patterns for those genes quite difficult in composite extractions. This result suggests again that composite tissue extractions can be difficult to interpret and might best be avoided whenever possible.
Making sequencing libraries from composite structures, such as body segments, and even whole insects is commonplace in RNA-seq studies [26–31]. Our results suggest that this practice can lead to false negatives for genes that show even strong patterns of differential expression in particular structures (organs). For genes that show complex patterns of variable expression in different structures across the organism, it is further likely that composite extractions are of little utility other than for identifying genes as candidates of interest. This is because it is difficult to infer the direction of differential expression in composite structures when more than one pattern may be present. Finally, our study focused on the abdomen, which is a relatively simple composite structure in comparison to a whole larval insect. It is likely that the types of errors documented here associated with identifying differentially expressed genes will be magnified in studies using whole organisms. Hence, given that structure specific extractions are usually possible for even very small structures [34–37], they should be conducted whenever possible.
Colonies and collection of bees
Honey bees were kept according to standard beekeeping practices at the bee biology facility at UC Davis. Two full size colonies were used in the study. All colonies were healthy and had been undisturbed for at least a month prior to collection of bees. Nurses were collected by identifying bees with their head and thorax completely in an open brood cell for at least three seconds [38, 39]. A further check to ensure that bees identified in this way were nurses was to check the developmental state of the Hypopharyngeal Glands (HP Glands) of each nurse at the start of each dissection. Nurses have large HP glands, as this gland produces the brood food fed by nurses to larva [40, 41]. Foragers, in contrast, have smaller, often yellowish, HP glands, as this gland produces digestive enzymes, not brood food, in bees in the foraging caste . Only bees with large white HP glands were kept as nurses. Foragers were collected as they entered the nest. Only foragers with pollen loads were used in the study. Immediately after collection, bees of both castes were placed into 50 ml centrifuge tubes and placed in the -80°C freezer.
Dissection, extractions, and sequencing
Additional file 7: Figure S3 shows the dry mass of each structure used in the study for comparative purposes. Dissections and RNA extractions were carried out one after the other to minimize degradation of RNA during dissection. Dissections were conducted by removing individual bees from the freezer and placing them into a petri dish with 50% ethanol under a dissection scope. As soon as the bee thawed, the dissection began. Dissections were completed within 3-5 minutes. For the sting gland, which is designed to detach from the adult bee when it stings, the process of dissection is simple. The stinger is grasped with forceps and pulled. The entire sting gland with associated venom sac pulls away from the body. Thirty stingers were pooled for each biological sample. All individuals pooled into one sample were from the same colony. Hence, 30 individuals from colony 1 and thirty individuals from colony 2 were used in the study. As each stinger was removed it was immediately homogenized in Trizol. This was repeated for each subsequent dissection. After all 30 stingers were dissected and homogenized, total RNA was extracted according to the manufacturer’s recommendations.
For the digestive tract a similar procedure was used with a few modifications. First the digestive tract is a large structure, so only 3 could be extracted per 1.5 ml microtube. Hence, 10 tubes total were used (3 individual bee digestive tracts per tube). After all 30 individual samples were homogenized, the samples were pooled in one 50 ml centrifuge tube and vortexed. Then one sample of 1 ml was taken from the total for extraction in a 1.5 ml tube. For the abdomen, which is too large for complete extraction in one microtube, 30 bee abdomens were first ground in liquid N2 using a mortar and pestle. 75 mg of the resulting fine powder was then taken for RNA extraction with Trizol.
RNA quality was checked with the Bioanalyzer 2100 and libraries were made using Illumina’s TruSeq v2 kit according to the manufacturer’s recommendations. 100 bp paired end sequencing was then performed on the HiSeq 2000 machine. The raw fastq files from this study are available at the NCBI SRA archive (SRP020361).
Quality control and RNA-seq analyses
Low quality bases and adapter contamination were removed with the fastx toolkit and the cutadapt software packages . Tophat (v2.04) was used for aligning reads to the Apis mellifera genome  (v4, the most recent officially published version). HTSeq was used for quantifying the number of reads mapping to each gene. NOISEQ, EdgeR and DESeq were used to determine differential expression [19, 45, 46]. For NOISeq, RPKM normalization was used along with a 0.8 p cutoff (the recommended cut-off level). For EdgeR and DESeq, an adjusted p value (FDR) < 0.05 was used to call differentially expressed genes. All analyses made use of 2 biological samples and 12 million quality controlled paired end reads. Expression levels within biological replicates for the same tissue were highly correlated (mean: 98.3%, range 97.1% -99.6%).
Identification of transcription factors
All Drosophila melanogaster genes with the GO term “sequence specific DNA binding” were downloaded from flybase and blasted against all genes in the official gene set of Apis mellifera. Genes with a hit (e < 10-20) to one of the Drosophila transcription factors that had a functional domain involved in DNA biding were kept. Overall, 462 genes passed this filter (Additional file 8: Table S5). While the resulting list is not exhaustive, in that there are surely many more Apis transcription factors, it is a large sample of transcription factors that should be broadly representative of this class of genes.
We thank Billy Synk and W. Cameron Jasper for help collecting honey bees. This work was funded by the University of California, Davis. The work described here complies with the laws of the United States of America. We thank three anonymous referees for careful reading of the manuscript.
- Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B: Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods. 2008, 5: 621-628. 10.1038/nmeth.1226.View ArticlePubMed
- Wang Z, Gerstein M, Snyder M: RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet. 2009, 10: 57-63. 10.1038/nrg2484.PubMed CentralView ArticlePubMed
- Marioni JC, Mason CE, Mane SM, Stephens M, Gilad Y: RNA-seq: an assessment of technical reproducibility and comparison with gene expression arrays. Genome Res. 2008, 18: 1509-1517. 10.1101/gr.079558.108.PubMed CentralView ArticlePubMed
- Hawkins RD, Hon GC, Ren B: Next-generation genomics: an integrative approach. Nat Rev Genet. 2010, 11: 476-486.PubMed CentralPubMed
- van Bakel H, Nislow C, Blencowe BJ, Hughes TR: Most “dark matter” transcripts are associated with known genes. Plos Biology. 2010, 8: 21-View Article
- Hallegger M, Llorian M, Smith CWJ: Alternative splicing: global insights. Febs J. 2010, 277: 856-866. 10.1111/j.1742-4658.2009.07521.x.View ArticlePubMed
- Pepke S, Wold B, Mortazavi A: Computation for ChIP-seq and RNA-seq studies. Nat Methods. 2009, 6: S22-S32. 10.1038/nmeth.1371.PubMed CentralView ArticlePubMed
- Ouyang ZQ, Zhou Q, Wong WH: ChIP-Seq of transcription factors predicts absolute and differential gene expression in embryonic stem cells. Proc Natl Acad Sci USA. 2009, 106: 21521-21526. 10.1073/pnas.0904863106.PubMed CentralView ArticlePubMed
- Sharma CM, Vogel J: Experimental approaches for the discovery anal characterization of regulatory small RNA. Curr Opin Microbiol. 2009, 12: 536-546. 10.1016/j.mib.2009.07.006.View ArticlePubMed
- Boyle AP, Song LY, Lee BK, London D, Keefe D, Birney E, Iyer VR, Crawford GE, Furey TS: High-resolution genome-wide in vivo footprinting of diverse transcription factors in human cells. Genome Res. 2011, 21: 456-464. 10.1101/gr.112656.110.PubMed CentralView ArticlePubMed
- Khalil AM, Guttman M, Huarte M, Garber M, Raj A, Morales DR, Thomas K, Presser A, Bernstein BE, van Oudenaarden A: Many human large intergenic noncoding RNAs associate with chromatin-modifying complexes and affect gene expression. Proc Natl Acad Sci USA. 2009, 106: 11667-11672. 10.1073/pnas.0904715106.PubMed CentralView ArticlePubMed
- Gupta RA, Shah N, Wang KC, Kim J, Horlings HM, Wong DJ, Tsai MC, Hung T, Argani P, Rinn JL: Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature. 2010, 464: 1071-U1148. 10.1038/nature08975.PubMed CentralView ArticlePubMed
- Tsai MC, Manor O, Wan Y, Mosammaparast N, Wang JK, Lan F, Shi Y, Segal E, Chang HY: Long noncoding RNA as modular scaffold of histone modification complexes. Science. 2010, 329: 689-693. 10.1126/science.1192002.PubMed CentralView ArticlePubMed
- Guttman M, Donaghey J, Carey BW, Garber M, Grenier JK, Munson G, Young G, Lucas AB, Ach R, Bruhn L: lincRNAs act in the circuitry controlling pluripotency and differentiation. Nature. 2011, 477: 295-U260. 10.1038/nature10398.PubMed CentralView ArticlePubMed
- Pauli A, Rinn JL, Schier AF: Non-coding RNAs as regulators of embryogenesis. Nat Rev Genet. 2011, 12: 136-149.PubMed CentralView ArticlePubMed
- Nagano T, Fraser P: No-nonsense functions for long noncoding RNAs. Cell. 2011, 145: 178-181. 10.1016/j.cell.2011.03.014.View ArticlePubMed
- Li H, Lovci MT, Kwon YS, Rosenfeld MG, Fu XD, Yeo GW: Determination of tag density required for digital transcriptome analysis: application to an androgen-sensitive prostate cancer model. Proc Natl Acad Sci USA. 2008, 105: 20179-20184. 10.1073/pnas.0807121105.PubMed CentralView ArticlePubMed
- Blencowe BJ, Ahmad S, Lee LJ: Current-generation high-throughput sequencing: deepening insights into mammalian transcriptomes. Genes Dev. 2009, 23: 1379-1386. 10.1101/gad.1788009.View ArticlePubMed
- Tarazona S, Garcia-Alcalde F, Dopazo J, Ferrer A, Conesa A: Differential expression in RNA-seq: a matter of depth. Genome Res. 2011, 21: 2213-2223. 10.1101/gr.124321.111.PubMed CentralView ArticlePubMed
- Haas BJ, Chin M, Nusbaum C, Birren BW, Livny J: How deep is deep enough for RNA-Seq profiling of bacterial transcriptomes?. Bmc Genomics. 2012, 13: 734-10.1186/1471-2164-13-734.PubMed CentralView ArticlePubMed
- Tang F, Barbacioru C, Bao S, Lee C, Nordman E, Wang X, Lao K, Surani MA: Tracing the derivation of embryonic stem cells from the inner cell mass by single-cell RNA-Seq analysis. Cell Stem Cell. 2010, 6: 468-478. 10.1016/j.stem.2010.03.015.PubMed CentralView ArticlePubMed
- Tang F, Barbacioru C, Nordman E, Li B, Xu N, Bashkirov VI, Lao K, Surani MA: RNA-Seq analysis to capture the transcriptome landscape of a single cell. Nat Protoc. 2010, 5: 516-535.View ArticlePubMed
- Westermann AJ, Gorski SA, Vogel J: Dual RNA-seq of pathogen and host. Nat Rev Microbiol. 2012, 10: 618-630. 10.1038/nrmicro2852.View ArticlePubMed
- Iourov IY, Vorsanova SG, Yurov YB: Single cell genomics of the brain: focus on neuronal diversity and neuropsychiatric diseases. Curr Genomics. 2012, 13: 477-488. 10.2174/138920212802510439.PubMed CentralView ArticlePubMed
- Tischler J, Surani MA: Investigating transcriptional states at single-cell-resolution. Curr Opin Biotechnol. 2013, 24: 69-78. 10.1016/j.copbio.2012.09.013.View ArticlePubMed
- Crawford JE, Guelbeogo WM, Sanou A, Traore A, Vernick KD, Sagnon N, Lazzaro BP: De novo transcriptome sequencing in anopheles funestus using illumina RNA-Seq technology. Plos One. 2010, 5 (12): e14202-10.1371/journal.pone.0014202.PubMed CentralView ArticlePubMed
- Winbush A, Reed D, Chang PL, Nuzhdin SV, Lyons LC, Arbeitman MN: Identification of gene expression changes associated with long-term memory of courtship rejection in drosophila males. G3-Genes Genomes Genetics. 2012, 2: 1437-1445.PubMed Central
- Bouvaine S, Behmer ST, Lin GG, Faure ML, Grebenok RJ, Douglas AE: The physiology of sterol nutrition in the pea aphid Acyrthosiphon pisum. J Insect Physiol. 2012, 58: 1383-1389. 10.1016/j.jinsphys.2012.07.014.View ArticlePubMed
- Chen X, Hu Y, Zheng HQ, Cao LF, Niu DF, Yu DL, Sun YQ, Hu SN, Hu FL: Transcriptome comparison between honey bee queen- and worker-destined larvae. Insect Biochem Mol Biol. 2012, 42: 665-673. 10.1016/j.ibmb.2012.05.004.View ArticlePubMed
- Teets NM, Peyton JT, Colinet H, Renault D, Kelley JL, Kawarasaki Y, Lee RE, Denlinger DL: Gene expression changes governing extreme dehydration tolerance in an Antarctic insect. Proc Natl Acad Sci USA. 2012, 109: 20744-20749. 10.1073/pnas.1218661109.PubMed CentralView ArticlePubMed
- Dunning LT, Dennis AB, Park D, Sinclair BJ, Newcomb RD, Buckley TR: Identification of cold-responsive genes in a New Zealand alpine stick insect using RNA-Seq. Comp Biochem Physiol D-Genomics Proteomics. 2013, 8: 24-31. 10.1016/j.cbd.2012.10.005.View Article
- Hoffman DR: Hymenoptera venom allergens. Clin Rev Allergy Immunol. 2006, 30 (2): 109-128. 10.1385/CRIAI:30:2:109.View ArticlePubMed
- Peiren N, de Graaf DC, Vanrobaeys F, Danneels EL, Devreese B, Van Beeurnen J, Jacobs FJ: Proteomic analysis of the honey bee worker venom gland focusing on the mechanisms of protection against tissue damage. Toxicon. 2008, 52 (1): 72-83. 10.1016/j.toxicon.2008.05.003.View ArticlePubMed
- Bonizzoni M, Dunn WA, Campbell CL, Olson KE, Marinotti O, James AA: Complex modulation of the Aedes aegypti transcriptome in response to dengue virus infection. Plos One. 2012, 7 (11): e50512-10.1371/journal.pone.0050512.PubMed CentralView ArticlePubMed
- Fontana JR, Crews ST: Transcriptome analysis of drosophila CNS midline cells reveals diverse peptidergic properties and a role for castor in neuronal differentiation. Dev Biol. 2012, 372: 131-142. 10.1016/j.ydbio.2012.09.010.PubMed CentralView ArticlePubMed
- Hughes ME, Grant GR, Paquin C, Qian J, Nitabach MN: Deep sequencing the circadian and diurnal transcriptome of Drosophila brain. Genome Res. 2012, 22: 1266-1281. 10.1101/gr.128876.111.PubMed CentralView ArticlePubMed
- Su Y-L, Li J-M, Li M, Luan J-B, Ye X-D, Wang X-W, Liu S-S: Transcriptomic analysis of the salivary glands of an invasive whitefly. Plos One. 2012, 7 (6): e39303-10.1371/journal.pone.0039303.PubMed CentralView ArticlePubMed
- Huang ZY, Robinson GE: Honey bee colony integration: worker-worker interactions mediate plasticity in endocrine and behavioral development. Proc Natl Acad Sci USA. 1992, 89: 11726-11729. 10.1073/pnas.89.24.11726.PubMed CentralView ArticlePubMed
- Whitfield CW, Ben-Shahar Y, Brillet C, Leoncini I, Crauser D, Le Conte Y, Rodriguez-Zas S, Robinson GE: Genomic dissection of behavioral maturation in the honey bee. Proc Natl Acad Sci USA. 2006, 103 (44): 16068-16075. 10.1073/pnas.0606909103.PubMed CentralView ArticlePubMed
- Crailsheim K: Interadult feeding of jelly in honeybee (Apis mellifera) colonies. J Comp Physiol B-Biochemical Systemic Environ Physiol. 1991, 161: 55-60. 10.1007/BF00258746.View Article
- Wegener J, Huang ZY, Lorenz MW, Bienefeld K: Regulation of hypopharyngeal gland activity and oogenesis in honey bee (Apis mellifera) workers. J Insect Physiol. 2009, 55: 716-725. 10.1016/j.jinsphys.2009.05.003.View ArticlePubMed
- Johnson BR: Division of labor in honeybees: form, function, and proximate mechanisms. Behav Ecol Sociobiol. 2010, 64 (3): 305-316. 10.1007/s00265-009-0874-7.PubMed CentralView ArticlePubMed
- Trapnell C, Pachter L, Salzberg SL: TopHat: discovering splice junctions with RNA-Seq. Bioinformatics. 2009, 25: 1105-1111. 10.1093/bioinformatics/btp120.PubMed CentralView ArticlePubMed
- Martin M: Cutadapt removes adapter sequences from high-throughput sequencing reads. 2011, North America: EMBnet.journal, Available at: http://journal.embnet.org/index.php/embnetjournal/article/view/200
- Anders S, Huber W: Differential expression analysis for sequence count data. Genome Biol. 2010, 11: R106-10.1186/gb-2010-11-10-r106.PubMed CentralView ArticlePubMed
- Robinson MD, McCarthy DJ, Smyth GK: edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010, 26: 139-140. 10.1093/bioinformatics/btp616.PubMed CentralView ArticlePubMed
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.