Bioinformatics Advance Access originally published online on December 7, 2004
Bioinformatics 2005 21(8):1389-1392; doi:10.1093/bioinformatics/bti205
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
LongSAGE analysis revealed the presence of a large number of novel antisense genes in the mouse genome
1Institute of Developmental Genetics, GSF-National Research Center for Environment and Health Ingolstädter Landstrasse 1, D-85764 Neuherberg, Germany
2Institute of Pathology, GSF-National Research Center for Environment and Health Ingolstädter Landstrasse 1, D-85764 Neuherberg, Germany
*To whom correspondence should be addressed.
 |
Abstract |
|---|
 TOP Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Motivation: Despite the increasing notions of the functional importance of antisense transcripts in gene regulation, the genome-wide overview on the ontology of antisense genes has not been obtained. Therefore, we tried to find novel antisense genes genome-wide by using our LongSAGE dataset of 202 015 tags (consisting of 41 718 unique tags), experimentally generated from mouse embryonic tail libraries.
Results: We identified 1260 potential antisense genes, of which 1001 are not annotated in EnsEMBL, thereby being regarded as novel. Interestingly their sense counterparts were co-expressed in the majority of the cases.
Conclusions: The use of LongSAGE transcriptome data is extremely powerful in the identification of thus-far unknown antisense transcripts, even in the case of well-characterized organisms like the mouse.
Contact: imai{at}gsf.de
|
INTRODUCTION |
|---|
|
TOPAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Control of gene expression by naturally occurring antisense transcripts has been discovered in prokaryotes more than 20 years ago (reviewed by Simons, 1988). In recent years, global gene expression surveys by microarrays (Yamada et al., 2003), as well as computational analyses based on full-length cDNA (Kiyosawa et al., 2003) or expressed sequence tag (EST) sequences (Shendure and Church, 2002) suggested that the genomes of higher vertebrates also contain a substantial number of genes that harbor oppositely oriented overlapping transcripts. However, the number of well-characterized antisense genes is still small (Vanhee-Brossollet and Vaquero, 1998). Therefore, the genome-wide overview on the ontology of antisense genes (i.e. how many genes are associated with their antisense genes) remains largely unknown. Furthermore, it cannot be ruled out that at least some of the identified sense–antisense gene pairs in the above-cited papers are artifacts, owing to experimental errors intrinsic to the corresponding approach. For example, hybridization-based approaches are sensitive to cross-hybridization, and the alignment of transcript sequences to the genome can be erroneous, owing to the wrong annotation of the transcript orientation. Furthermore, around half of the antisense transcripts are single-exon genes, and therefore the correct orientation cannot be confirmed by the analysis of canonical splice sites (Kiyosawa et al., 2003). Thus, most approaches are only qualitative, and do not evaluate the quantitative aspect of sense/antisense transcript pairs in a certain tissue.
Serial Analysis of Gene Expression (SAGE) (Velculescu et al., 1995) is a promising method to evaluate antisense transcripts in a genome-wide scale. For a SAGE analysis, a priori knowledge of transcript sequences is not required, which is an intrinsic and principle advantage of SAGE over microarray approaches. Furthermore, the SAGE method is far more effective than EST sequencing (Sun et al., 2004) and supplies quantitative gene expression data as digital counts. In the present study, we analyzed a dataset of 202 015 LongSAGE tags generated from mouse embryonic tails and examined it in comparison with public cDNA and EST datasets to determine and evaluate antisense genes in the mouse genome.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Tissue dissection
Tails of 268 stage-matched E10.5 mouse embryos of C57BL/6N origin (Charles River) were microdissected by four sagittal cuts into four tissue parts, including the tail bud, the caudal two-third or the rostral one-third of the presomitic mesoderm and two pairs of most recently formed somites. The collected tissues were homogenized in the Binding/Lysis buffer (Dynal) and immediately stored at –80°C until use.
LongSAGE library construction
Poly(A)-positive RNA was isolated using the mRNA DIRECT kit (Dynal), bound to oligo(dT)25 magnetic beads, and immediately proceeded to LongSAGE library construction, according to the standard protocol (Saha et al., 2002), with our-own modifications (Wahl et al., 2004). For each LongSAGE library construction, distinct linker/primer combinations were used in order to avoid cross-contaminations of LongSAGE tags between the libraries (see Supplementary Table). Sequencing was performed with big-dye terminators on an ABI 3100 DNA Analyzer (Applied Biosystems).
LongSAGE data acquisition
By using Phred (Ewing et al., 1998), extracted LongSAGE tags were considered only when all bases had a Phred score of 10 or higher. LongSAGE tag sequences were further assessed by SAGEScreen (Akmaev and Wang, 2004). The complete LongSAGE tag dataset used in this study is deposited at the Gene Expression Omnibus database at NCBI with the accession number GSM26978
[NCBI GEO]
.
Tag-to-gene mapping
Assignments of obtained LongSAGE tags to corresponding genes were carried out, using information from public databases such as the UniGene, Mouse Genome Database (MGD) and EnsEMBL, and described in detail in the accompanying paper (Wahl et al., 2005).
Identification of antisense transcripts
All cDNA and EST sequences overlapping with a LongSAGE tag were retrieved from the EnsEMBL databases: i.e. mus_musculus_core_19_30, mus_musculus_estgene_19_30 and mus_musculus_est_19_30, which were derived from the NCBI mouse genome assembly (build 30). Sequences were processed further only when canonical splice donor and/or acceptor sites could be identified immediately flanking its HSP (transcript sequence against the genome), or when the whole sequence is aligned to the genomic sequence to take single-exon genes into account. Next, all transcripts overlapping with a LongSAGE tag were compared against EnsEMBL and EST genes. In the case where a LongSAGE tag was found on the opposite strand within a 10 kb region of a EnsEMBL or EST gene, the LongSAGE tag/transcript sequence pair was considered to represent a antisense gene, according to the following criteria: (1) the transcript sequence(s) also overlapped with an exon of the EnsEMBL gene (following splicing rules), or (2) they shared a minimum percentage identity of 95% over at least 150 bp. Antisense LongSAGE tags assigned to the same sense EnsEMBL gene were considered as being derived from the same antisense gene.
Quantification of gene expression for sense and antisense tags
The expression level of each of the sense or antisense genes listed in Table 3 was determined as the sum of counts of all alternative LongSAGE tags derived from the single gene.
|
|
RESULTS |
|---|
|
TOPAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
LongSAGE library construction
C57BL/6 mice were used in this study, so that LongSAGE tag sequences could be best compared to mouse genome sequences that were mostly of C57BL/6 origin. The tail of E10.5 mouse embryos was microdissected into four different parts, and mRNA was extracted and used for generating a total of eight LongSAGE libraries: two independent libraries from each of the four tail parts. From each library, excluding linker tags and duplicate ditags, around 25 000 tags were sequenced and a total of 202 015 tags were collected, corresponding to 41 714 unique tags (Table 1). Although a very small amount of starting RNA material required a slightly increased number of PCR cycles for ditag amplification (exactly 36 cycles for each library), the frequency of duplicate ditags (1–4%) was low. Furthermore, GC-content (Margulies et al., 2001b) as well as the incidence of linker tags (0.5–4%) is comparable to other good SAGE libraries (Margulies et al., 2001a), and a comparison of the pairs of LongSAGE libraries generated from the same tissue part showed a high reproducibility, pointing to the good quality and consistency of the LongSAGE libraries.
|
Identification of sense–antisense gene pairs
To minimize false-positive cases, we determined only those antisense transcripts, which were supported by LongSAGE tag as well as cDNA and/or EST sequences. Therefore, ESTs overlapping with the genomic position of a LongSAGE tag were determined and considered to be corresponding to the same transcript as the LongSAGE tag, only when they could be aligned to the genome on the same strand following splicing rules (Fig. 1). The LongSAGE tag/transcript pair was considered to be in an antisense orientation to a EnsEMBL gene, if the LongSAGE tag was located either within an exon of the sense EnsEMBL gene (Fig. 1A), or the associated transcript overlapped outside the LongSAGE tag with the EnsEMBL gene (Fig. 1B). As summarized in Table 2, among a total of 18 205 transcript-verified LongSAGE tags, a total of 1468 were in an antisense orientation to an annotated EnsEMBL gene, suggesting that these antisense LongSAGE tags represented potential antisense genes.
|
|
Evaluation of sense—antisense gene pairs
We further analyzed the observed antisense LongSAGE tags supported by transcript sequences more in detail. As depicted in Table 2, these 1468 antisense tags correspond to 1260 genes, by considering that different LongSAGE tags antisense to the same sense gene were derived from the same gene. For 78% of the antisense genes, at least one transcript for the sense counterpart was detected in our LongSAGE dataset. In general, the sense genes were expressed at a higher level (average tag count: 32.1) than the antisense genes (average tag count: 4.8). The 20 most abundant antisense transcripts not included in the EnsEMBL gene set are listed in Table 3. Furthermore, a complete list of 1260 sense-antisense gene pairs as an online Supplementary material in a tab-delimited text format is available for downloading from the OUP server.
|
DISCUSSION |
|---|
|
TOPAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our analysis has identified 1260 potential antisense genes. The absence of annotated EnsEMBL genes for most of the antisense tags detected in our dataset points out that the EnsEMBL gene annotation pipeline omits most of the genes in antisense orientation to protein-coding genes. Indeed, the number of sense–antisense gene pairs within the EnsEMBL dataset, 217 in human (Shendure and Church, 2002), is dramatically lower than those 2418 pairs observed in mouse full-length cDNA sequences (Kiyosawa et al., 2003). Interestingly, the number of potential antisense genes in the LongSAGE libraries is more than half compared to the numbers found in the Riken Fantom2 set. Since only limited types of tissues were analyzed in this study, it is conceivable that the number of antisense genes represented in the Riken Fantom2 dataset still underestimates the quantity of existing antisense genes. This is in accordance with the notion that only half of known sense/antisense gene pairs were detected in the Riken Fantom2 dataset (Kiyosawa et al., 2003). Furthermore, a recent global gene expression study in Arabidopsis detected that more than one-third of all genes have an expressed antisense counterpart, of which
10% are even co-expressed in the same tissue (Yamada et al., 2003). This indicates that a large fraction of eukaryotic genes have an antisense counterpart. It is also interesting that, in the majority of the cases, expression of an antisense gene is associated with the expression of its sense gene, and that the expression level of sense genes is much higher than that of corresponding antisense genes. These observations are consistent with the notions that antisense genes are expressed often at low levels and that their transcripts are often unstable (Storz, 2002).
The biological function of those antisense transcripts might be of great interest. Recently, a major focus in the community is on micro RNAs (miRNAs) that are generated by the successive processing of RNAs to 21–23 base long RNAs, which inhibit the transcription of its targets (reviewed by Bartel, 2004). Yet computational predictions (Lim et al., 2003a,b) and experimental cloning (Lagos-Quintana et al., 2002) led to the identification of less than 900 miRNAs in all species (miRNA registry) (Griffiths-Jones, 2004), and are therefore outnumbered by the antisense genes mentioned above. However, owing to its short length (22 bases) and since precursors of miRNAs do not have to be bidirectionally located on the opposite strand of the genome, many miRNAs will not be captured with the strategy applied. Because of its different structure the identified antisense transcripts might function by a different mechanism other than that of miRNAs. Studies over the last few years suggest that antisense transcripts function in various different ways, finally regulating or antagonizing its sense counterpart. Cases have been reported in which antisense transcripts affect alternative splicing (Munroe and Lazar, 1991), RNA editing (Kumar and Carmichael, 1997), X-inactivation (Lee and Lu, 1999), translational regulation (Li and Murphy, 2000), imprinting (Sleutels et al., 2002) and transcriptional repression by methylation (Tufarelli et al., 2003). Experimental validations are required in future studies to define the nature and functions of the potential antisense genes identified in this study.
|
Acknowledgments |
|---|
We thank Rudi Balling (GBF) for valuable comments to this work. We also thank Ken Kinzler and Victor Velculescu (Johns Hopkins University) for the SAGE protocol and software, and Jerzy Adamski and Peter Lichtner (Genome Analysis Center, GSF) for providing the sequencing facility. This work was supported by the GSF.
|
Footnotes |
|---|
Present address: Stowers Institute for Medical Research, 1000 E. 50th Street, Kansas City, MO 64110, USA 
Received on September 4, 2004; revised on November 23, 2004; accepted on December 2, 2004
|
REFERENCES |
|---|
|
TOPAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Akmaev, V.R. and Wang, C.J. (2004) Correction of sequence-based artifacts in serial analysis of gene expression. Bioinformatics, 20, 1254–1263
Bartel, D.P. (2004) MicroRNAs: genomics, biogenesis, mechanism, and function. Cell, 116, 281–297[CrossRef][Web of Science][Medline].
Ewing, B., Hillier, L., Wendl, M.C., Green, P. (1998) Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res., 8, 175–185
Griffiths-Jones, S. (2004) The microRNA Registry. Nucleic Acids Res., 32, 109–111.
Kiyosawa, H., Yamanaka, I., Osato, N., Kondo, S., Hayashizaki, Y. (2003) Antisense transcripts with FANTOM2 clone set and their implications for gene regulation. Genome Res., 13, 1324–1334
Kumar, M. and Carmichael, G.G. (1997) Nuclear antisense RNA induces extensive adenosine modifications and nuclear retention of target transcripts. Proc. Natl Acad. Sci., USA, 94, 3542–3547
Lagos-Quintana, M., Rauhut, R., Yalcin, A., Meyer, J., Lendeckel, W., Tuschl, T. (2002) Identification of tissue-specific microRNAs from mouse. Curr. Biol., 12, 735–739[CrossRef][Web of Science][Medline].
Lee, J.T. and Lu, N. (1999) Targeted mutagenesis of Tsix leads to nonrandom X inactivation. Cell, 99, 47–57[CrossRef][Web of Science][Medline].
Li, A.W. and Murphy, P.R. (2000) Expression of alternatively spliced FGF-2 antisense RNA transcripts in the central nervous system: regulation of FGF-2 mRNA translation. Mol. Cell. Endocrinol., 170, 233–242[Medline].
Lim, L.P., Glasner, M.E., Yekta, S., Burge, C.B., Bartel, D.P. (2003a) Vertebrate microRNA genes. Science, 299, 1540
Lim, L.P., Lau, N.C., Weinstein, E.G., Abdelhakim, A., Yekta, S., Rhoades, M.W., Burge, C.B., Bartel, D.P. (2003b) The microRNAs of Caenorhabditis elegans. Genes Dev., 17, 991–1008
Margulies, E.H., Kardia, S.L., Innis, J.W. (2001a) A comparative molecular analysis of developing mouse forelimbs and hindlimbs using serial analysis of gene expression (SAGE). Genome Res., 11, 1686–1698
Margulies, E.H., Kardia, S.L., Innis, J.W. (2001b) Identification and prevention of a GC content bias in SAGE libraries. Nucleic Acids Res., 29, E60.
Munroe, S.H. and Lazar, M.A. (1991) Inhibition of c-erbA mRNA splicing by a naturally occurring antisense RNA. J. Biol. Chem., 266, 22083–22086
Saha, S., Sparks, A.B., Rago, C., Akmaev, V., Wang, C.J., Vogelstein, B., Kinzler, K.W., Velculescu, V.E. (2002) Using the transcriptome to annotate the genome. Nat. Biotechnol., 20, 508–512[CrossRef][Web of Science][Medline].
Shendure, J. and Church, G.M. (2002) Computational discovery of sense-antisense transcription in the human and mouse genomes. Genome Biol., 3, RESEARCH0044.
Simons, R.W. (1988) Naturally occurring antisense RNA control—a brief review. Gene, 72, 35–44[CrossRef][Web of Science][Medline].
Sleutels, F., Zwart, R., Barlow, D.P. (2002) The non-coding Air RNA is required for silencing autosomal imprinted genes. Nature, 415, 810–813[Medline].
Storz, G. (2002) An expanding universe of noncoding RNAs. Science, 296, 1260–1263
Sun, M., Zhou, G., Lee, S., Chen, J., Shi, R.Z., Wang, S.M. (2004) SAGE is far more sensitive than EST for detecting low-abundance transcripts. BMC Genomics, 5, 1[CrossRef][Medline].
Tufarelli, C., Stanley, J.A., Garrick, D., Sharpe, J.A., Ayyub, H., Wood, W.G., Higgs, D.R. (2003) Transcription of antisense RNA leading to gene silencing and methylation as a novel cause of human genetic disease. Nat. Genet., 34, 157–165[CrossRef][Web of Science][Medline].
Vanhee-Brossollet, C. and Vaquero, C. (1998) Do natural antisense transcripts make sense in eukaryotes? Gene, 211, 1–9[CrossRef][Web of Science][Medline].
Velculescu, V.E., Zhang, L., Vogelstein, B., Kinzler, K.W. (1995) Serial analysis of gene expression. Science, 270, 484–487
Wahl, M.B., Heinzmann, U., Imai, K. (2005) LongSAGE analysis significantly improves genome annotation: identifications of novel genes and alternative transcripts in the mouse. Bioinformatics, 21, 1393–1400
Wahl, M., Shukunami, C., Heinzmann, U., Hamajima, K., Hiraki, Y., Imai, K. (2004) Transcriptome analysis of early chondrogenesis in ATDC5 cells induced by bone morphogenetic protein 4. Genomics, 83, 45–58[CrossRef][Web of Science][Medline].
Yamada, K., Lim, J., Dale, J.M., Chen, H., Shinn, P., Palm, C.J., Southwick, A.M., Wu, H.C., Kim, C., Nguyen, M., et al. (2003) Empirical analysis of transcriptional activity in the Arabidopsis genome. Science, 302, 842–846
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Bianchetti, Y. Wu, E. Guerin, F. Plewniak, and O. Poch SAGETTARIUS: a program to reduce the number of tags mapped to multiple transcripts and to plan SAGE sequencing stages Nucleic Acids Res., September 25, 2007; 35(18): e122 - e122. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Ge, Q. Wu, Y.-C. Jung, J. Chen, and S. M. Wang A large quantity of novel human antisense transcripts detected by LongSAGE Bioinformatics, October 15, 2006; 22(20): 2475 - 2479. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||