Using the transcriptome to annotate the genome

ASAP1, a gene at 8q24, is associated with prostate cancer metastasis

Metastatic prostate cancer is a terminal disease, and the development of reliable prognostic tools and more effective therapy is critically important for improved disease survival and management. This study was aimed at identifying genes that are differentially expressed in metastatic and nonmetastatic prostate cancer cells and, as such, could be critical in the development of metastasis. Long-SAGE analysis was used to compare a transplantable human metastatic prostate cancer subline, PCa1-met, with a nonmetastatic counterpart, PCa2. Both sublines were developed from a patient's prostate cancer specimen via subrenal capsule grafting and subsequent orthotopic implantation into SCID mice. Among various differentially expressed genes identified, ASAP1, an 8q24 gene encoding an ADP-ribosylation factor GTPase-activating protein not previously associated with prostate cancer, was up-regulated in the metastatic subline as confirmed by quantitative real-time PCR. Immunohistochemistry of xenograft sections showed that cytoplasmic ASAP1 protein staining was absent or weak in benign tissue, significantly stronger in nonmetastatic PCa2 tissue, and strongest in PCa1-met tissue. In clinical specimens, ASAP1 protein staining was elevated in 80% of primary prostate cancers and substantially higher in metastatic lesions compared with benign prostate tissue. Moreover, additional ASAP1 gene copies were detected in 58% of the primary prostate cancer specimens. Small interfering RNA-induced reduction of ASAP1 protein expression markedly suppressed in vitro PC-3 cell migration (approximately 50%) and Matrigel invasion (approximately 67%). This study suggests that the ASAP1 gene plays a role in prostate cancer metastasis and may represent a therapeutic target and/or biomarker for metastatic disease.

Human neural crest cells display molecular and phenotypic hallmarks of stem cells

The fields of both developmental and stem cell biology explore how functionally distinct cell types arise from a self-renewing founder population. Multipotent, proliferative human neural crest cells (hNCC) develop toward the end of the first month of pregnancy. It is assumed that most differentiate after migrating throughout the organism, although in animal models neural crest stem cells reportedly persist in postnatal tissues. Molecular pathways leading over time from an invasive mesenchyme to differentiated progeny such as the dorsal root ganglion, the maxillary bone or the adrenal medulla are altered in many congenital diseases. To identify additional components of such pathways, we derived and maintained self-renewing hNCC lines from pharyngulas. We show that, unlike their animal counterparts, hNCC are able to self-renew ex vivo under feeder-free conditions. While cross species comparisons showed extensive overlap between human, mouse and avian NCC transcriptomes, some molecular cascades are only active in the human cells, correlating with phenotypic differences. Furthermore, we found that the global hNCC molecular profile is highly similar to that of pluripotent embryonic stem cells when compared with other stem cell populations or hNCC derivatives. The pluripotency markers NANOG, POU5F1 and SOX2 are also expressed by hNCC, and a small subset of transcripts can unambiguously identify hNCC among other cell types. The hNCC molecular profile is thus both unique and globally characteristic of uncommitted stem cells.

Transcriptome analysis of the normal human mammary cell commitment and differentiation process

Mature mammary epithelial cells are generated from undifferentiated precursors through a hierarchical process, but the molecular mechanisms involved, particularly in the human mammary gland, are poorly understood. To address this issue, we isolated highly purified subpopulations of primitive bipotent and committed luminal progenitor cells as well as mature luminal and myoepithelial cells from normal human mammary tissue and compared their transcriptomes obtained using three different methods. Elements unique to each subset of mammary cells were identified, and changes that accompany their differentiation in vivo were shown to be recapitulated in vitro. These include a stage-specific change in NOTCH pathway gene expression during the commitment of bipotent progenitors to the luminal lineage. Functional studies further showed NOTCH3 signaling to be critical for this differentiation event to occur in vitro. Taken together, these findings provide an initial foundation for future delineation of mechanisms that perturb primitive human mammary cell growth and differentiation.

Clustering-based approaches to SAGE data mining

Serial analysis of gene expression (SAGE) is one of the most powerful tools for global gene expression profiling. It has led to several biological discoveries and biomedical applications, such as the prediction of new gene functions and the identification of biomarkers in human cancer research. Clustering techniques have become fundamental approaches in these applications. This paper reviews relevant clustering techniques specifically designed for this type of data. It places an emphasis on current limitations and opportunities in this area for supporting biologically-meaningful data mining and visualisation.

Identification of a set of genes showing regionally enriched expression in the mouse brain

Background

The Pleiades Promoter Project aims to improve gene therapy by designing human mini-promoters (< 4 kb) that drive gene expression in specific brain regions or cell-types of therapeutic interest. Our goal was to first identify genes displaying regionally enriched expression in the mouse brain so that promoters designed from orthologous human genes can then be tested to drive reporter expression in a similar pattern in the mouse brain.

Results

We have utilized LongSAGE to identify regionally enriched transcripts in the adult mouse brain. As supplemental strategies, we also performed a meta-analysis of published literature and inspected the Allen Brain Atlas in situ hybridization data. From a set of approximately 30,000 mouse genes, 237 were identified as showing specific or enriched expression in 30 target regions of the mouse brain. GO term over-representation among these genes revealed co-involvement in various aspects of central nervous system development and physiology.

Conclusion

Using a multi-faceted expression validation approach, we have identified mouse genes whose human orthologs are good candidates for design of mini-promoters. These mouse genes represent molecular markers in several discrete brain regions/cell-types, which could potentially provide a mechanistic explanation of unique functions performed by each region. This set of markers may also serve as a resource for further studies of gene regulatory elements influencing brain expression.

[Gene expression profiling in drug addicted brain]

Drug addiction, a chronic brain disease caused by interaction of in vitro drug toxification and in vivo gene susceptibility, has been widely studied but its underlining mechanism is so far been elucidated. A major goal in this field is to identify drug-induced molecular changes and their effects on brain function. By the advance of high throughput technologies in genomics and transcriptomics, the whole gene expression profile in addicted brain could be obtained and proved to be a very powerful tool to unclose the molecular mechanism underlying the addiction biology context. Here, we summarized the progress of serial analysis of gene expression (SAGE) and microarray, as well as their application in drug addiction.

Comparison of the expression profiles of promastigotes and axenic amastigotes in Leishmania donovani using serial analysis of gene expression

In this study, we examined the transcriptome of Leishmania donovani promastigotes and axenic amastigotes to identify differentially regulated mRNAs utilizing the serial analysis of gene expression (SAGE). The axenic culture of amastigotes was initiated from stationary phase promastigotes. Transformation from promastigote to amastigote occurred when cultures in Medium 199 (pH 5.5), supplemented with 20% (v/v) fetal bovine serum, were transferred from 26 degrees C to 37 degrees C. A total of 20,299 and 20,132 tags were generated from promastigote and amastigote libraries, respectively. The containing unique genes identified in these two SAGE libraries were 8,615 and 7,835, respectively. Characteristics of the expressed genes' frequency distribution were remarkably similar in both libraries: the most abundant tags (frequency>or=20), whose levels were equal to or >1.3% of the identified tags, constituted >23% of the total sequenced tags. Correspondingly, 75.72% or 71.65% of the genes accounted for those tags present at low abundance (frequency=1) contributed only 32.13% or 27.89% of the total tags. A total of 968 genes (11.2% of the total genes in promastigotes and 12.4% in amastigotes) were recorded to have statistically different transcript levels between promastigotes and axenic amastigotes. Of the 968 distinct total genes, there are 326 promastigote-enriched transcripts and 642 amastigote-enriched mRNAs.

StemBase: a resource for the analysis of stem cell gene expression data

StemBase is a database of gene expression data obtained from stem cells and derivatives mainly from mouse and human using DNA microarrays and Serial Analysis of Gene Expression. Here, we describe this database and indicate ways to use it for the study the expression of particular genes in stem cells or to search for genes with particular expression profiles in stem cells, which could be associated to stem cell function or used as stem cell markers.

Serial analysis of the vascular endothelial transcriptome under static and shear stress conditions

We have utilized serial analysis of gene expression (SAGE) to analyze the response of human coronary artery endothelial cells (HCAECs) to laminar shear stress (LSS). Primary cultures of HCAECs were exposed to 15 dyn/cm(2) LSS for 24 h in a parallel plate flow chamber and compared with identical same passage cells cultured under static conditions. The expression levels of a number of functional categories of genes were reduced by shear stress including those encoding proteins involved in cell proliferation (CDC10, CDC20, CDC23, CCND1, CCNB1), angiogenesis (ANGPTL4, CTGF, CYR61, ENG, EPAS1, EGFR, LGALS3, PGK1, and SPARC), extracellular matrix and cell-matrix adhesion (EFEMP1, LOXL2, P4HB, FBN1, FN1, ITGA5, ITGAE, ITGAV, ILK, LAMR1) and ATP synthesis (ATP5G3, ATP5J2, ATP5L, ATP5D). We also observed an increase in the LSS-responsive expression of genes encoding stress response proteins, including HMOX1, which is significant since HMOX1 may have anti-inflammatory and vasodilatory vascular effects. The autosomal dominant polycystic kidney disease (ADPKD) genes PKD1 and PKD2 were also elevated by LSS. ADPKD is associated with vascular malfunction, including the impairment of vasoreactive processes. To our knowledge, this is the first SAGE-based analysis of the shear stress-responsive endothelial cell transcriptome. These immortal data provide a resource for further analyses of the molecular mechanisms underlying the biological response to LSS and contribute to the expanding collection of publicly available SAGE data.

A genomic analysis of RNA polymerase II modification and chromatin architecture related to 3' end RNA polyadenylation

Genomic analyses have been applied extensively to analyze the process of transcription initiation in mammalian cells, but less to transcript 3' end formation and transcription termination. We used a novel approach to prepare 3' end fragments from polyadenylated RNA, and mapped the position of the poly(A) addition site using oligonucleotide arrays tiling 1% of the human genome. This approach revealed more 3' ends than had been annotated. The distribution of these ends relative to RNA polymerase II (PolII) and di- and trimethylated lysine 4 and lysine 36 of histone H3 was compared. A substantial fraction of unannotated 3' ends of RNA are intronic and antisense to the embedding gene. Poly(A) ends of annotated messages lie on average 2 kb upstream of the end of PolII binding (termination). Near the termination sites, and in some internal sites, unphosphorylated and C-terminal domain (CTD) serine 2 phosphorylated PolII (POLR2A) accumulate, suggesting pausing of the polymerase and perhaps dephosphorylation prior to release. Lysine 36 trimethylation occurs across transcribed genes, sometimes alternating with stretches of DNA in which lysine 36 dimethylation is more prominent. Lysine 36 methylation decreases at or near the site of polyadenylation, sometimes disappearing before disappearance of phosphorylated RNA PolII or release of PolII from DNA. Our results suggest that transcription termination loss of histone 3 lysine 36 methylation and later release of RNA polymerase. The latter is often associated with polymerase pausing. Overall, our study reveals extensive sites of poly(A) addition and provides insights into the events that occur during 3' end formation.

Genome-wide analysis of protein-DNA interactions

The human genome is predominantly composed of nonprotein-coding sequences whose function remains largely undefined. A significant portion of the noncoding DNA is believed to serve as transcriptional regulatory elements that control gene expression in specific cell types at appropriate developmental stages. Identifying these regulatory sequences and determining the mechanisms by which they act present a great challenge in the postgenomic era. Previous investigations using genetic, molecular, and biochemical approaches have uncovered a large number of proteins involved in regulating transcription. Knowledge of the genomic locations of DNA binding for these proteins in the nucleus should define the identity and nature of the transcriptional regulatory sequences and reveal the gene regulatory networks in cells. Chromatin immunoprecipitation (ChIP) is a common method for detecting interactions between a protein and a DNA sequence in vivo. In recent years, this method has been combined with DNA microarrays and other high-throughput technologies to enable genome-wide identification of DNA-binding sites for various nuclear proteins. Here, we review recent advances in ChIP-based methods for genome-wide detection of protein-DNA interactions, and discuss their significance in enhancing our knowledge of the gene regulatory networks and epigenetic mechanisms in cells.

Detailed characterization of the mouse embryonic stem cell transcriptome reveals novel genes and intergenic splicing associated with pluripotency

Background

Transcriptional control of embryonic stem (ES) cell pluripotency has been a subject of intense study. Transcriptional regulators including Oct4 (Oct3/4 index), Sox2 and Nanog are fundamental for maintaining the undifferentiated state. However, the ES cell transcriptome is not limited to their targets, and exhibits considerable complexity when assayed with microarray, MPSS, cDNA/EST sequencing, and SAGE technologies. To identify novel genes associated with pluripotency, we globally searched for ES transcripts not corresponding to known genes, validated their sequences, determined their expression profiles, and employed RNAi to test their function.

Results

Gene Identification Signature (GIS) analysis, a SAGE derivative distinguished by paired 5' and 3' transcript end tags, identified 153 candidate novel transcriptional units (TUs) distinct from known genes in a mouse E14 ES mRNA library. We focused on 16 TUs free of artefacts and mapping discrepancies, five of which were validated by RTPCR product sequencing. Two of the TUs were revealed by annotation to represent novel protein-coding genes: a PRY-domain cluster member and a KRAB-domain zinc finger. The other three TUs represented intergenic splicing events involving adjacent, functionally unrelated protein-coding genes transcribed in the same orientation, with one event potentially encoding a fusion protein containing domains from both component genes (Clk2 and Scamp3). Expression profiling using embryonic samples and adult tissue panels confirmed that three of the TUs were unique to or most highly expressed in ES cells. Expression levels of all five TUs dropped dramatically during three distinct chemically induced differentiation treatments of ES cells in culture. However, siRNA knockdowns of the TUs did not alter mRNA levels of pluripotency or differentiation markers, and did not affect cell morphology.

Conclusion

Transcriptome libraries retain considerable potential for novel gene discovery despite massive recent cDNA and EST sequencing efforts; cDNA and EST evidence for these ES cell TUs had been limited or absent. RTPCR and full-length sequencing remain essential in resolving the bottleneck between numerous candidate novel transcripts inferred from high-throughput sequencing and the small fraction that can be validated. RNAi results indicate that, despite their strong association with pluripotency, these five transcriptomic novelties may not be required for maintaining it.

Interrogating global gene expression in rat neuronal cultures using SAGE

The normal function of the mammalian brain is regulated by complex networks of interactions between cells and molecules, which are to a considerable extent dependent on mechanisms of transcriptional regulation. Disruption of such interactions by neurotoxic stimuli may lead to severe forms of dementia and to other neuropsychiatric disorders. Therefore, critical insight into mechanisms of neuronal dysfunction may be obtained by examining global patterns of gene expression in mammalian models of neurotoxicity. In this regard, the combined use of rat neuronal cultures and serial analysis of gene expression (SAGE) can be viewed as a general platform to enable the search for molecular targets involved in neurotoxic processes. Here, we discuss potential advantages of this approach, highlighting the need for generation of robust SAGE libraries from rat neuronal cultures. The availability and current limitations of bioinformatics tools for SAGE data derived from rat samples is also discussed.

5'- and 3'-RACE from LongSAGE tags

Serial analysis of gene expression (SAGE) studies often yield numerous tags that cannot be mapped to known gene sequences. Intriguingly, these may represent unknown genes, unknown parts of genes, or transcript variants. In order to elucidate the origin of these tags, 3'- and 5'-rapid amplification of complementary DNA ends (RACE) reactions can be performed using primers identical or complementary to SAGE tags. This way, transcript fragments, or indeed the entire uncharacterized transcript, can be cloned and sequenced.

Low-cost-medium throughput Sanger dideoxy sequencing

Serial analysis of gene expression (SAGE) requires the sequencing of DNA. The principal cost of SAGE is largely determined by the cost of sequencing. Therefore, it is important to have access to a robust and affordable sequencing system. Here, we describe such a system based on the sequencing of amplified inserts of concatemer-containing clones.

Genome-wide allele-specific expression analysis using Massively Parallel Signature Sequencing (MPSS) reveals cis- and trans-effects on gene expression in maize hybrid meristem tissue

Allelic differences in expression are important genetic factors contributing to quantitative trait variation in various organisms. However, the extent of genome-wide allele-specific expression by different modes of gene regulation has not been well characterized in plants. In this study we developed a new methodology for allele-specific expression analysis by applying Massively Parallel Signature Sequencing (MPSS), an open ended and sequencing based mRNA profiling technology. This methodology enabled a genome-wide evaluation of cis- and trans-effects on allelic expression in six meristem stages of the maize hybrid. Summarization of data from nearly 400 pairs of MPSS allelic signature tags showed that 60% of the genes in the hybrid meristems exhibited differential allelic expression. Because both alleles are subjected to the same trans-acting factors in the hybrid, the data suggest the abundance of cis-regulatory differences in the genome. Comparing the same allele expressed in the hybrid versus its inbred parents showed that 40% of the genes were differentially expressed, suggesting different trans-acting effects present in different genotypes. Such trans-acting effects may result in gene expression in the hybrid different from allelic additive expression. With this approach we quantified gene expression in the hybrid relative to its inbred parents at the allele-specific level. As compared to measuring total transcript levels, this study provides a new level of understanding of different modes of gene regulation in the hybrid and the molecular basis of heterosis.

Paired-end diTagging for transcriptome and genome analysis

The Paired-End diTagging (PET) procedure enables one to obtain sequence information from both termini of any contiguous DNA fragment. This is achieved by a series of enzymatic manipulations that introduce MmeI sites directly flanking each DNA insert during the construction of a plasmid library. Subsequent MmeI digestion and self-ligation results in the production of covalently-linked paired-end ditags (PETs) that can be extracted and then concatenated for efficient sequencing. By mapping the PET sequences to assembled genomes, the original DNA fragments from which the PETs were derived can be precisely localized. This unit details two applications of PET technology. In GIS-PET, ditagging of mRNA converted to full-length cDNA enables whole-transcriptome analysis, including novel gene identification, gene prediction validation, and gene expression studies. In ChIP-PET, ditagging of chromatin immunoprecipitation-enriched genomic DNA fragments enables the global mapping of transcription factor binding sites. A recent innovation (Multiplex Sequencing of Paired-End ditags; MS-PET) enables PETs to be sequenced using high-throughput 454 sequencing, greatly increasing the amount of data that can be collected in each run.

LongSAGE analysis of skeletal muscle at three prenatal stages in Tongcheng and Landrace pigs

BACKGROUND

Obese and lean pig breeds show obvious differences in muscle growth; however, the molecular mechanism underlying phenotype variation remains unknown. Prenatal muscle development programs postnatal performance. Here, we describe a genome-wide analysis of differences in prenatal skeletal muscle between Tongcheng (a typical indigenous Chinese breed) and Landrace (a leaner Western breed) pigs.

RESULTS

We generated transcriptome profiles of skeletal muscle from Tongcheng and Landrace pigs at 33, 65 and 90 days post coitus (dpc), using long serial analysis of gene expression (LongSAGE). We sequenced 317,115 LongSAGE tags and identified 1,400 and 1,201 differentially expressed transcripts during myogenesis in Tongcheng and Landrace pigs, respectively. From these, the Gene Ontology processes and expression patterns of these differentially expressed genes were constructed. Most of the genes showed different expression patterns in the two breeds. We also identified 532, 653 and 459 transcripts at 33, 65 and 90 dpc, respectively, that were differentially expressed between the two breeds. Growth factors, anti-apoptotic factors and genes involved in the regulation of protein synthesis were up-regulated in Landrace pigs. Finally, 12 differentially expressed genes were validated by quantitative PCR.

CONCLUSION

Our data show that gene expression phenotypes differ significantly between the two breeds. In particular, a slower muscle growth rate and more complicated molecular changes were found in Tongcheng pigs, while genes responsible for increased cellular growth and myoblast survival were up-regulated in Landrace pigs. Our analyses will assist in the identification of candidate genes for meat production traits and elucidation of the development of prenatal skeletal muscle in mammals.

LongSAGE profiling of nine human embryonic stem cell lines

To facilitate discovery of novel human embryonic stem cell (ESC) transcripts, we generated 2.5 million LongSAGE tags from 9 human ESC lines. Analysis of this data revealed that ESCs express proportionately more RNA binding proteins compared with terminally differentiated cells, and identified novel ESC transcripts, at least one of which may represent a marker of the pluripotent state.

aRNA-LongSAGE: SAGE with antisense RNA

In order to generate serial analysis of gene expression (SAGE) libraries from very small samples such as microdissected cells, the starting material must first be amplified via PCR or linear amplification of RNA. In microarray experiments, it has been shown that linear amplification of RNA can be used to generate reliable gene expression profiles and leads to the detection of expression differences that are not seen with nonamplified starting material. As the product of the amplification is amplified antisense RNA (aRNA), linear amplification of RNA cannot be used in combination with the conventional SAGE protocol. The aRNA-LongSAGE protocol described herein is an adaptation of the MicroSAGE protocol to the use of aRNA as starting material.

Correction of technology-related artifacts in serial analysis of gene expression

Serial analysis of gene expression (SAGE) is a powerful technique for measuring global gene expression through sampling of transcript tags. SAGE tag collections or libraries serve as a rich data source for differential gene expression analysis, transcriptome mapping, and gene discovery. Transcriptome mapping and gene discovery are facilitated by extensions of SAGE, e.g., Long SAGE, where the transcript tags are elongated by utilization of a different tagging enzyme. SAGE, as a sequencing-based technique, is prone to errors resulting in artifact SAGE tag sequences and erroneous tag numbers. A methodology to pinpoint and correct tag artifacts is necessary to fully exploit the value of large SAGE libraries. SAGEScreen is a tag sequence correction algorithm. The algorithm is a multistep procedure that addresses error rates and performs ditag and tag processing. The error rate estimates are based on a stochastic model of PCR and sequencing related mutations. The ditag processing step is essential for calculation of unbiased tag numbers, and the tag processing step allows for filtration of tag sequence artifacts and adjustment of tag numbers.

Modeling transcriptome based on transcript-sampling data

Background

Newly-evolved multiplex sequencing technology has been bringing transcriptome sequencing into an unprecedented depth. Millions of transcript tags now can be acquired in a single experiment through parallelization. The significant increase in throughput and reduction in cost required us to address some fundamental questions, such as how many transcript tags do we have to sequence for a given transcriptome? How could we estimate the total number of unique transcripts for different cell types (transcriptome diversity) and the distribution of their copy numbers (transcriptome dynamics)? What is the probability that a transcript with a given expression level to be detected at a certain sampling depth?

Methodology/Principal Findings

We developed a statistical model to evaluate these parameters based on transcriptome-sampling data. Three mixture models were exploited for their potentials to model the sampling frequencies. We demonstrated that relative abundances of all transcripts in a transcriptome follow the generalized inverse Gaussian distribution. The widely known beta and gamma distributions failed to fulfill the singular characteristics of relative abundance distribution, i.e., highly skewed toward zero and with a long tail. An estimator of transcriptome diversity and an analytical form of sampling growth curve were proposed in a coherent framework. Experimental data fitted this model very well and Monte Carlo simulations based on this model replicated sampling experiments in a remarkable precision.

Conclusions

Taking human embryonic stem cell as a prototype, we demonstrated that sequencing tens of thousands of transcript tags in an ordinary EST/SAGE experiment was far from sufficient. In order to fully characterize a human transcriptome, millions of transcript tags had to be sequenced. This model lays a statistical basis for transcriptome-sampling experiments and in essence can be used in all sampling-based data.

Identifying chromosomal targets of DNA-binding proteins by Sequence Tag Analysis of Genomic Enrichment (STAGE)

Sequence Tag Analysis of Genomic Enrichment (STAGE) is a method for experimentally identifying the in vivo chromosomal targets of DNA-binding proteins in any sequenced genome. STAGE generates 21-bp tags derived from DNA isolated by chromatin immunoprecipitation (ChIP; UNIT 21.3). Concatamers of tags are cloned and sequenced to yield a STAGE library. Tags in the library represent DNA fragments that were occupied by the DNA-binding protein, and mapping these tag sequences to the genome identifies the binding loci of the DNA-binding protein in vivo. STAGE can be applied to any sequenced genome to identify targets of DNA-binding proteins without requiring extensive microarray resources.

Quantitative gene expression profiling of mouse brain regions reveals differential transcripts conserved in human and affected in disease models

Using serial analysis of gene expression, we collected quantitative transcriptome data in 11 regions of the adult wild-type mouse brain: the orbital, prelimbic, cingulate, motor, somatosensory, and entorhinal cortices, the caudate-putamen, the nucleus accumbens, the thalamus, the substantia nigra, and the ventral tegmental area. With >1.2 million cDNA tags sequenced, this database is a powerful resource to explore brain functions and disorders. As an illustration, we performed interregional comparisons and found 315 differential transcripts. Most of them are poorly characterized and 20% lack functional annotation. For 78 differential transcripts, we provide independent expression level measurements in mouse brain regions by real-time quantitative RT-PCR. We also show examples where we used in situ hybridization to achieve infrastructural resolution. For 30 transcripts, we next demonstrated that regional enrichment is conserved in the human brain. We then quantified the expression levels of region-enriched transcripts in the R6/2 mouse model of Huntington disease and the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mouse model of Parkinson disease and observed significant alterations in the striatum, cerebral cortex, thalamus and substantia nigra of R6/2 mice and in the striatum of MPTP-treated mice. These results show that the gene expression data provided here for the mouse brain can be used to explore pathophysiological models and disclose transcripts differentially expressed in human brain regions.

Integrated approaches to uncovering transcription regulatory networks in mammalian cells

Integrative systems biology has emerged as an exciting research approach in molecular biology and functional genomics that involves the integration of genomics, proteomics, and metabolomics datasets. These endeavors establish a systematic paradigm by which to interrogate, model, and iteratively refine our knowledge of the regulatory events within a cell. Here we review the latest technologies available to collect high-throughput measurements of a cellular state as well as the most successful methods for the integration and interrogation of these measurements. In particular we will focus on methods available to infer transcription regulatory networks in mammals.

SAGE and LongSAGE

Serial analysis of gene expression (SAGE) is a high-throughput method for global gene expression analysis that allows the quantitative and simultaneous analysis of a large number of transcripts. SAGE is a digital method and its sensitivity depends only on the number of tags sequenced. Furthermore, SAGE is a powerful tool for finding novel genes that are expressed under certain conditions or in certain tissues. SAGE has been widely used in fields as diverse as cancer research and the development and study of microorganisms. The SAGE method is a series of routine molecular biology procedure and can, at least in principle, be carried out in any laboratory. However, the number of consecutive steps is quite large and in practice, SAGE has been difficult to carry out on a routine basis.

Robust-LongSAGE (RL-SAGE): an improved LongSAGE method for high-throughput transcriptome analysis

Serial analysis of gene expression (SAGE) is a powerful technique for large-scale transcriptome analysis in eukaryotes. However, technical difficulties in the SAGE library construction, such as low concatemer cloning efficiency, small concatemer size, and a high level of empty clones, has prohibited its widespread use as a routine technique for expression profiling in many laboratories. We recently improved the LongSAGE library construction method considerably and developed a modified version called Robust-LongSAGE, or RL-SAGE. In RL-SAGE, concatemer cloning efficiency and clone insert size were increased significantly. About 20 PCR reactions are sufficient to make a library with more than 150,000 clones. Using RL-SAGE, we have made 10 libraries of rice, maize, and the rice blast fungus Magnaporthe grisea.

High-resolution, genome-wide mapping of chromatin modifications by GMAT

One major postgenomic challenge is to characterize the epigenomes that control genome functions. The epigenomes are mainly defined by the specific association of nonhistone proteins with chromatin and the covalent modifications of chromatin, including DNA methylation and posttranslational histone modifications. The in vivo protein-binding and chromatin-modification patterns can be revealed by the chromatin immunoprecipitation assay (ChIP). By combining the ChIP assays and the serial analysis of gene expression (SAGE) protocols, we have developed an unbiased and high-resolution genome-wide mapping technique (GMAT) to determine the genome-wide protein-targeting and chromatin-modification patterns. GMAT has been successfully applied to mapping the target sites of the histone acetyltransferase, Gcn5p, in yeast and to the discovery of the histone acetylation islands as an epigenetic mark for functional regulatory elements in the human genome.

DeepSAGE: higher sensitivity and multiplexing of samples using a simpler experimental protocol

Combining serial analysis of gene expression (SAGE) with pyrophosphatase-based ultra-high-throughput DNA sequencing provides increased sensitivity and cost-effective gene expression profiling. The combined techniques obviate the formation and cloning of concatemers and the tedious picking and preparation of sequence templates from bacterial clones that are necessary with SAGE alone. Furthermore, multiplexing of samples or replicates of analysis is included in the experimental design.

Transcript profiling by 3'-untranslated region sequencing resolves expression of gene families

Differences in gene expression underlie central questions in plant biology extending from gene function to evolutionary mechanisms and quantitative traits. However, resolving expression of closely related genes (e.g. alleles and gene family members) is challenging on a genome-wide scale due to extensive sequence similarity and frequently incomplete genome sequence data. We present a new expression-profiling strategy that utilizes long-read, high-throughput sequencing to capture the information-rich 3'-untranslated region (UTR) of messenger RNAs (mRNAs). Resulting sequences resolve gene-specific transcripts independent of a sequenced genome. Analysis of approximately 229,000 3'-anchored sequences from maize (Zea mays) ovaries identified 14,822 unique transcripts represented by at least two sequence reads. Total RNA from ovaries of drought-stressed wild-type and viviparous-1 mutant plants was used to construct a multiplex cDNA library. Each sample was labeled by incorporating one of 16 unique three-base key codes into the 3'-cDNA fragments, and combined samples were sequenced using a GS 20 454 instrument. Transcript abundance was quantified by frequency of sequences identifying each unique mRNA. At least 202 unique transcripts showed highly significant differences in abundance between wild-type and mutant samples. For a subset of mRNAs, quantitative differences were validated by real-time reverse transcription-polymerase chain reaction. The 3'-UTR profile resolved 12 unique cellulose synthase (CesA) transcripts in maize ovaries and identified previously uncharacterized members of a histone H1 gene family. In addition, this method resolved nearly identical paralogs, as illustrated by two auxin-repressed, dormancy-associated (Arda) transcripts, which showed reciprocal mRNA abundance in wild-type and mutant samples. Our results demonstrate the potential of 3'-UTR profiling for resolving gene- and allele-specific transcripts.

Duplicate ditag analysis in LongSAGE

The Long serial analysis of gene expression (SAGE) protocol generates ditags from tags with overlapping overhangs, thereby increasing the probability of duplicate ditag formation in LongSAGE. In this chapter, a tool is presented that facilitates the analysis of duplicate ditags in LongSAGE studies to determine whether they should be included or not.

Digital karyotyping

Detection of copy number variation in the human genome is important for identifying naturally occurring copy number polymorphisms as well as alterations that underlie various human diseases, including cancer. Digital karyotyping uses short sequence tags derived from specific genomic loci to provide a quantitative and high-resolution view of copy number changes on a genome-wide scale. Genomic tags are obtained using a combination of enzymatic digests and isolation of short DNA sequences. Individual tags are linked into ditags, concatenated, cloned and sequenced. Tags are matched to reference genome sequences and digital enumeration of groups of neighboring tags provides quantitative copy number information along each chromosome. Digital karyotyping libraries can be generated in about a week, and library sequencing and data analysis require several additional weeks.

Regulation of a virus-induced lethal disease in tomato revealed by LongSAGE analysis

Infection of Cucumber mosaic virus (CMV) and D satellite RNA (satRNA) in tomato plants induces rapid plant death, which has caused catastrophic crop losses. We conducted long serial analysis of gene expression (LongSAGE) in control and virus-infected plants to identify the genes that may be involved in the development of this lethal tomato disease. The transcriptomes were compared between mock-inoculated plants and plants infected with CMV, CMV/D satRNA, or CMV/Dm satRNA (a nonnecrogenic mutant of D satRNA with three mutated nucleotides). The analysis revealed both general and specific changes in the tomato transcriptome after infection with these viruses. A massive transcriptional difference of approximately 400 genes was found between the transcriptomes of CMV/D and CMV/Dm satRNA-infected plants. Particularly, the Long-SAGE data indicated the activation of ethylene synthesis and signaling by CMV/D satRNA infection. Results from inoculation tests with an ethylene-insensitive mutant and treatments with an ethylene action inhibitor further confirmed the role of ethylene in mediating the epinastic leaf symptoms and the secondary cell death in the stem. Results from Northern blot analysis demonstrated the partial contribution of ethylene in the induced defense responses in CMV/D satRNA-infected plants.

Transcriptome changes associated with instructed learning in the barn owl auditory localization pathway

Owls reared wearing prismatic spectacles learn to make adaptive orienting movements. This instructed learning depends on re-calibration of the midbrain auditory space map, which in turn involves the formation of new synapses. Here we investigated whether these processes are associated with differential gene expression, using longSAGE. Newly fledged owls were reared for 8-36 days with prism or control lenses at which time the extent of learning was quantified by electrophysiological mapping. Transciptome profiles were obtained from the inferior colliculus (IC), the major site of synaptic plasticity, and the optic tectum (OT), which provides an instructive signal that controls the direction and extent of plasticity. Twenty-two differentially expressed sequence tags were identified in IC and 36 in OT, out of more than 35,000 unique tags. Of these, only four were regulated in both structures. These results indicate that regulation of two largely independent gene clusters is associated with synaptic remodeling (in IC) and generation of the instructive signal (in OT). Real-time PCR data confirmed the changes for two transcripts, ubiquitin/polyubiquitin and tyrosine 3-monooxgenase/tryotophan 5-monooxygenase activation protein, theta subunit (YWHAQ; also referred to as 14-3-3 protein). Ubiquitin was downregulated in IC, consistent with a model in which protein degradation pathways act as an inhibitory constraint on synaptogenesis. YWHAQ was up-regulated in OT, indicating a role in the synthesis or delivery of instructive information. In total, our results provide a path towards unraveling molecular cascades that link naturalistic experience with synaptic remodeling and, ultimately, with the expression of learned behavior.

Localization of TFIIB binding regions using serial analysis of chromatin occupancy

Background:

RNA Polymerase II (RNAP II) is recruited to core promoters by the pre-initiation complex (PIC) of general transcription factors. Within the PIC, transcription factor for RNA polymerase IIB (TFIIB) determines the start site of transcription. TFIIB binding has not been localized, genome-wide, in metazoans. Serial analysis of chromatin occupancy (SACO) is an unbiased methodology used to empirically identify transcription factor binding regions. In this report, we use TFIIB and SACO to localize TFIIB binding regions across the rat genome.

Results:

A sample of the TFIIB SACO library was sequenced and 12,968 TFIIB genomic signature tags (GSTs) were assigned to the rat genome. GSTs are 20–22 base pair fragments that are derived from TFIIB bound chromatin. TFIIB localized to both non-protein coding and protein-coding loci. For 21% of the 1783 protein-coding genes in this sample of the SACO library, TFIIB binding mapped near the characterized 5' promoter that is upstream of the transcription start site (TSS). However, internal TFIIB binding positions were identified in 57% of the 1783 protein-coding genes. Internal positions are defined as those within an inclusive region greater than 2.5 kb downstream from the 5' TSS and 2.5 kb upstream from the transcription stop. We demonstrate that both TFIIB and TFIID (an additional component of PICs) bound to internal regions using chromatin immunoprecipitation (ChIP). The 5' cap of transcripts associated with internal TFIIB binding positions were identified using a cap-trapping assay. The 5' TSSs for internal transcripts were confirmed by primer extension. Additionally, an analysis of the functional annotation of mouse 3 (FANTOM3) databases indicates that internally initiated transcripts identified by TFIIB SACO in rat are conserved in mouse.

Conclusion:

Our findings that TFIIB binding is not restricted to the 5' upstream region indicates that the propensity for PIC to contribute to transcript diversity is far greater than previously appreciated.

NAC-1 controls cell growth and survival by repressing transcription of Gadd45GIP1, a candidate tumor suppressor

Cancer mortality and morbidity are primarily related to recurrent tumors, and characterization of recurrence-associated genes should illuminate fundamental properties of tumor progression and provide new therapeutic targets. We have previously identified NAC-1, a member of the BTB/POZ gene family and a transcription repressor, as a gene associated with recurrent ovarian carcinomas after chemotherapy. We further showed that homodimerization of NAC-1 proteins is essential for tumor growth and survival. In this study, we applied serial analysis of gene expression and identified growth arrest and DNA-damage-inducible 45-gamma interacting protein (Gadd45GIP1) as one of the downstream genes negatively regulated by NAC-1. NAC-1 knockdown in both SKOV3 and HeLa cells that expressed abundant endogenous NAC-1 induced Gadd45GIP1 expression transcriptionally; on the other hand, engineered expression of NAC-1 in NAC-1-negative RK3E and HEK293 cells suppressed endogenous Gadd45GIP1 expression. In NAC-1-expressing tumor cells, induction of dominant negative NAC-1 conferred a growth-inhibitory effect that can be partially reversed by Gadd45GIP1 knockdown. Induced Gadd45GIP1 expression resulted in growth arrest in SKOV3 and HeLa cells both in vitro and in vivo. In summary, NAC-1 contributes to tumor growth and survival by at least inhibiting Gadd45GIP1 expression, which has a tumor suppressor effect in cancer cells.

Serial analysis of gene expression

Serial analysis of gene expression (SAGE) is a method used to obtain comprehensive, unbiased and quantitative gene-expression profiles. Its major advantage over arrays is that it does not require a priori knowledge of the genes to be analyzed and reflects absolute mRNA levels. Since the original SAGE protocol was developed in a short-tag (10-bp) format, several modifications have been made to produce longer SAGE tags for more precise gene identification and to decrease the amount of starting material necessary. Several SAGE-like methods have also been developed for the genome-wide analysis of DNA copy-number changes and methylation patterns, chromatin structure and transcription factor targets. In this protocol, we describe the 17-bp longSAGE method for transcriptome profiling optimized for a small amount of starting material. The generation of such libraries can be completed in 7-10 d, whereas sequencing and data analysis require an additional 2-3 wk.

Deep analysis of cellular transcriptomes - LongSAGE versus classic MPSS

BACKGROUND

Deep transcriptome analysis will underpin a large fraction of post-genomic biology. 'Closed' technologies, such as microarray analysis, only detect the set of transcripts chosen for analysis, whereas 'open' e.g. tag-based technologies are capable of identifying all possible transcripts, including those that were previously uncharacterized. Although new technologies are now emerging, at present the major resources for open-type analysis are the many publicly available SAGE (serial analysis of gene expression) and MPSS (massively parallel signature sequencing) libraries. These technologies have never been compared for their utility in the context of deep transcriptome mining.

RESULTS

We used a single LongSAGE library of 503,431 tags and a "classic" MPSS library of 1,744,173 tags, both prepared from the same T cell-derived RNA sample, to compare the ability of each method to probe, at considerable depth, a human cellular transcriptome. We show that even though LongSAGE is more error-prone than MPSS, our LongSAGE library nevertheless generated 6.3-fold more genome-matching (and therefore likely error-free) tags than the MPSS library. An analysis of a set of 8,132 known genes detectable by both methods, and for which there is no ambiguity about tag matching, shows that MPSS detects only half (54%) the number of transcripts identified by SAGE (3,617 versus 1,955). Analysis of two additional MPSS libraries shows that each library samples a different subset of transcripts, and that in combination the three MPSS libraries (4,274,992 tags in total) still only detect 73% of the genes identified in our test set using SAGE. The fraction of transcripts detected by MPSS is likely to be even lower for uncharacterized transcripts, which tend to be more weakly expressed. The source of the loss of complexity in MPSS libraries compared to SAGE is unclear, but its effects become more severe with each sequencing cycle (i.e. as MPSS tag length increases).

CONCLUSION

We show that MPSS libraries are significantly less complex than much smaller SAGE libraries, revealing a serious bias in the generation of MPSS data unlikely to have been circumvented by later technological improvements. Our results emphasize the need for the rigorous testing of new expression profiling technologies.

SAGETTARIUS: a program to reduce the number of tags mapped to multiple transcripts and to plan SAGE sequencing stages

SAGE (Serial Analysis of Gene Expression) experiments generate short nucleotide sequences called ‘tags’ which are assumed to map unambiguously to their original transcripts (1 tag to 1 transcript mapping). Nevertheless, many tags are generated that do not map to any transcript or map to multiple transcripts. Current bioinformatics resources, such as SAGEmap and TAGmapper, have focused on reducing the number of unmapped tags. Here, we describe SAGETTARIUS, a new high-throughput program that performs successive precise Nla3 and Sau3A tag to transcript mapping, based on specifically designed Virtual Tag (VT) libraries. First, SAGETTARIUS decreases the number of tags mapped to multiple transcripts. Among the various mapping resources compared, SAGETTARIUS performed the best in this respect by decreasing up to 11% the number of multiply mapped tags. Second, SAGETTARIUS allows the establishment of a guideline for SAGE experiment sequencing efforts through efficient mapping of the CRT (Cytoplasmic Ribosomal protein Transcripts)-specific tags. Using all publicly available human and mouse Nla3 SAGE experiments, we show that sequencing 100 000 tags is sufficient to map almost all CRT-specific tags and that four sequencing stages can be identified when carrying out a human or mouse SAGE project. SAGETTARIUS is web interfaced and freely accessible to academic users.

SAGE analysis of cell types involved in tolerance induction

Investigations into the mechanisms of immunological tolerance are currently hindered by a paucity of convenient markers, both for the identification and isolation of tolerant cell types and for monitoring the establishment of tolerance in in vivo models. Although high-affinity autoreactive T cells are deleted in the thymus during the establishment of central tolerance, escaping autoreactive cells require modulation in the periphery. Dendritic cells (DC) and regulatory T cells (Treg) are both implicated in the establishment and maintenance of peripheral tolerance, although specific interactions and mechanisms remain to be established. The serial analysis of gene expression (SAGE) approach to transcript profiling offers potential, not only for new insight into tolerogenic mechanisms, unbiased by current dogma, but also for the identification of novel molecular markers of tolerance. SAGE provides both quantitative and qualitative information on transcripts sampled on the basis of frequency of occurrence in the initial mRNA pool. This information is generated in the form of electronic databases that accumulate as a permanent resource and confer on SAGE the ability to readily compare across wide datasets. This offers particular potential when attempting to correlate gene expression with functional phenotype. By comparing variously generated functionally distinct/related immune populations, such as effector T cells and either natural, CD4+CD25+, or adaptive, Tr1, Tregs and/or immune and tolerance prone DC, it should be possible, using SAGE, to identify both individual genes and also signatures of genes associated with protolerogenic rather than immunogenic phenotypes.

Fungal transcriptomics

We have now entered in the post-genomic era, where we have knowledge of plethora of fungal genomes and cutting edge technology is available to study global mRNA, protein and metabolite profiles. These so-called 'omic' technologies (transcriptomics, proteomics and metabolomics) provide the possibility to characterize plant-pathogen interactions and pathogenesis at molecular level. This article provides an overview of transcriptomics and its applications in fungal plant pathology.

Characterization of pancreatic transcription factor Pdx-1 binding sites using promoter microarray and serial analysis of chromatin occupancy

The homeobox transcription factor Pdx-1 is necessary for pancreas organogenesis and beta cell function, however, most Pdx-1-regulated genes are unknown. To further the understanding of Pdx-1 in beta cell biology, we have characterized its genomic targets in NIT-1 cells, a mouse insulinoma cell line. To identify novel targets, we developed a microarray that includes traditional promoters as well as non-coding conserved elements, micro-RNAs, and elements identified through an unbiased approach termed serial analysis of chromatin occupancy. In total, 583 new Pdx-1 target genes were identified, many of which contribute to energy sensing and insulin release in pancreatic beta cells. By analyzing 31 of the protein-coding Pdx-1 target genes, we show that 29 are expressed in beta cells and, of these, 68% are down- or up-regulated in cells expressing a dominant negative mutant of Pdx-1. We additionally show that many Pdx-1 targets also interact with NeuroD1/BETA2, including the micro-RNA miR-375, a known regulator of insulin secretion.

Identification and analysis of internal promoters in Caenorhabditis elegans operons

The current Caenorhabditis elegans genomic annotation has many genes organized in operons. Using directionally stitched promoterGFP methodology, we have conducted the largest survey to date on the regulatory regions of annotated C. elegans operons and identified 65, over 25% of those studied, with internal promoters. We have termed these operons "hybrid operons." GFP expression patterns driven from internal promoters differ in tissue specificity from expression of operon promoters, and serial analysis of gene expression data reveals that there is a lack of expression correlation between genes in many hybrid operons. The average length of intergenic regions with putative promoter activity in hybrid operons is larger than previous estimates for operons as a whole. Genes with internal promoters are more commonly involved in gene duplications and have a significantly lower incidence of alternative splicing than genes without internal promoters, although we have observed almost all trans-splicing patterns in these two distinct groups. Finally, internal promoter constructs are able to rescue lethal knockout phenotypes, demonstrating their necessity in gene regulation and survival. Our work suggests that hybrid operons are common in the C. elegans genome and that internal promoters influence not only gene organization and expression but also operon evolution.

Transcriptome annotation using tandem SAGE tags

Analysis of several million expressed gene signatures (tags) revealed an increasing number of different sequences, largely exceeding that of annotated genes in mammalian genomes. Serial analysis of gene expression (SAGE) can reveal new Poly(A) RNAs transcribed from previously unrecognized chromosomal regions. However, conventional SAGE tags are too short to identify unambiguously unique sites in large genomes. Here, we design a novel strategy with tags anchored on two different restrictions sites of cDNAs. New transcripts are then tentatively defined by the two SAGE tags in tandem and by the spanning sequence read on the genome between these tagged sites. Having developed a new algorithm to locate these tag-delimited genomic sequences (TDGS), we first validated its capacity to recognize known genes and its ability to reveal new transcripts with two SAGE libraries built in parallel from a single RNA sample. Our algorithm proves fast enough to experiment this strategy at a large scale. We then collected and processed the complete sets of human SAGE tags to predict yet unknown transcripts. A cross-validation with tiling arrays data shows that 47% of these TDGS overlap transcriptional active regions. Our method provides a new and complementary approach for complex transcriptome annotation.

The molecular signature and cis-regulatory architecture of a C. elegans gustatory neuron

Taste receptor cells constitute a highly specialized cell type that perceives and conveys specific sensory information to the brain. The detailed molecular composition of these cells and the mechanisms that program their fate are, in general, poorly understood. We have generated serial analysis of gene expression (SAGE) libraries from two distinct populations of single, isolated sensory neuron classes, the gustatory neuron class ASE and the thermosensory neuron class AFD, from the nematode Caenorhabditis elegans. By comparing these two libraries, we have identified >1000 genes that define the ASE gustatory neuron class on a molecular level. This set of genes contains determinants of the differentiated state of the ASE neuron, such as a surprisingly complex repertoire of transcription factors (TFs), ion channels, neurotransmitters, and receptors, as well as seven-transmembrane receptor (7TMR)-type putative gustatory receptor genes. Through the in vivo dissection of the cis-regulatory regions of several ASE-expressed genes, we identified a small cis-regulatory motif, the "ASE motif," that is required for the expression of many ASE-expressed genes. We demonstrate that the ASE motif is a binding site for the C2H2 zinc finger TF CHE-1, which is essential for the correct differentiation of the ASE gustatory neuron. Taken together, our results provide a unique view of the molecular landscape of a single neuron type and reveal an important aspect of the regulatory logic for gustatory neuron specification in C. elegans.

Human transcriptome subtraction by using short sequence tags to search for tumor viruses in conjunctival carcinoma

Digital transcript subtraction (DTS) was developed to subtract in silico known human sequences from expression library data sets, leaving candidate nonhuman sequences for further analysis. This approach requires precise discrimination between human and nonhuman cDNA sequences. Database comparisons show high likelihood that small viral sequences can be successfully distinguished from human sequences. DTS analysis of 9,026 20-bp tags from an expression library of BCBL-1 cells infected with Kaposi's sarcoma-associated herpesvirus (KSHV) resolved all but three candidate sequences. Two of these sequences belonged to KSHV transcripts, and the third belonged to an unannotated human expression sequence tag. Overall, 0.24% of transcripts from this cell line were of viral origin. DTS analysis of 241,122 expression tags from three squamous cell conjunctival carcinomas revealed that only 21 sequences did not align with sequences from human databases. All 21 candidates amplify human transcripts and have secondary evidence for being of human origin. This analysis shows that it is unlikely that distinguishable viral transcripts are present in conjunctival carcinomas at 20 transcripts per million or higher, which is the equivalent of approximately 4 transcripts per cell. DTS is a simple screening method to discover novel viral nucleic acids. It provides, for the first time, quantitative evidence against some classes of viral etiology when no viral transcripts are found, thereby reducing the uncertainty involved in new pathogen discovery.

A systematic screen for genes expressed in definitive endoderm by Serial Analysis of Gene Expression (SAGE)

Background

The embryonic definitive endoderm (DE) gives rise to organs of the gastrointestinal and respiratory tract including the liver, pancreas and epithelia of the lung and colon. Understanding how DE progenitor cells generate these tissues is critical to understanding the cause of visceral organ disorders and cancers, and will ultimately lead to novel therapies including tissue and organ regeneration. However, investigation into the molecular mechanisms of DE differentiation has been hindered by the lack of early DE-specific markers.

Results

We describe the identification of novel as well as known genes that are expressed in DE using Serial Analysis of Gene Expression (SAGE). We generated and analyzed three longSAGE libraries from early DE of murine embryos: early whole definitive endoderm (0–6 somite stage), foregut (8–12 somite stage), and hindgut (8–12 somite stage). A list of candidate genes enriched for expression in endoderm was compiled through comparisons within these three endoderm libraries and against 133 mouse longSAGE libraries generated by the Mouse Atlas of Gene Expression Project encompassing multiple embryonic tissues and stages. Using whole mount in situ hybridization, we confirmed that 22/32 (69%) genes showed previously uncharacterized expression in the DE. Importantly, two genes identified, Pyy and 5730521E12Rik, showed exclusive DE expression at early stages of endoderm patterning.

Conclusion

The high efficiency of this endoderm screen indicates that our approach can be successfully used to analyze and validate the vast amount of data obtained by the Mouse Atlas of Gene Expression Project. Importantly, these novel early endoderm-expressing genes will be valuable for further investigation into the molecular mechanisms that regulate endoderm development.