Existing Gene Models Research Articles

Comparative Life Cycle Transcriptomics Revises Leishmania mexicana Genome Annotation and Links a Chromosome Duplication with Parasitism of Vertebrates.

Leishmania spp. are protozoan parasites that have two principal life cycle stages: the motile promastigote forms that live in the alimentary tract of the sandfly and the amastigote forms, which are adapted to survive and replicate in the harsh conditions of the phagolysosome of mammalian macrophages. Here, we used Illumina sequencing of poly-A selected RNA to characterise and compare the transcriptomes of L. mexicana promastigotes, axenic amastigotes and intracellular amastigotes. These data allowed the production of the first transcriptome evidence-based annotation of gene models for this species, including genome-wide mapping of trans-splice sites and poly-A addition sites. The revised genome annotation encompassed 9,169 protein-coding genes including 936 novel genes as well as modifications to previously existing gene models. Comparative analysis of gene expression across promastigote and amastigote forms revealed that 3,832 genes are differentially expressed between promastigotes and intracellular amastigotes. A large proportion of genes that were downregulated during differentiation to amastigotes were associated with the function of the motile flagellum. In contrast, those genes that were upregulated included cell surface proteins, transporters, peptidases and many uncharacterized genes, including 293 of the 936 novel genes. Genome-wide distribution analysis of the differentially expressed genes revealed that the tetraploid chromosome 30 is highly enriched for genes that were upregulated in amastigotes, providing the first evidence of a link between this whole chromosome duplication event and adaptation to the vertebrate host in this group. Peptide evidence for 42 proteins encoded by novel transcripts supports the idea of an as yet uncharacterised set of small proteins in Leishmania spp. with possible implications for host-pathogen interactions.

A robust blood gene expression-based prognostic model for castration-resistant prostate cancer

BackgroundCastration-resistant prostate cancer (CRPC) is associated with wide variations in survival. Recent studies of whole blood mRNA expression-based biomarkers strongly predicted survival but the genes used in these biomarker models were non-overlapping and their relationship was unknown. We developed a biomarker model for CRPC that is robust, but also captures underlying biological processes that drive prostate cancer lethality.MethodsUsing three independent cohorts of CRPC patients, we developed an integrative genomic approach for understanding the biological processes underlying genes associated with cancer progression, constructed a novel four-gene model that captured these changes, and compared the performance of the new model with existing gene models and other clinical parameters.ResultsOur analysis revealed striking patterns of myeloid- and lymphoid-specific distribution of genes that were differentially expressed in whole blood mRNA profiles: up-regulated genes in patients with worse survival were overexpressed in myeloid cells, whereas down-regulated genes were noted in lymphocytes. A resulting novel four-gene model showed significant prognostic power independent of known clinical predictors in two independent datasets totaling 90 patients with CRPC, and was superior to the two existing gene models.ConclusionsWhole blood mRNA profiling provides clinically relevant information in patients with CRPC. Integrative genomic analysis revealed patterns of differential mRNA expression with changes in gene expression in immune cell components which robustly predicted the survival of CRPC patients. The next step would be validation in a cohort of suitable size to quantify the prognostic improvement by the gene score upon the standard set of clinical parameters.Electronic supplementary materialThe online version of this article (doi:10.1186/s12916-015-0442-0) contains supplementary material, which is available to authorized users.

Pistillody mutant reveals key insights into stamen and pistil development in wheat (Triticum aestivum L.).

BackgroundThe pistillody mutant wheat (Triticum aestivum L.) plant HTS-1 exhibits homeotic transformation of stamens into pistils or pistil-like structures. Unlike common wheat varieties, HTS-1 produces three to six pistils per floret, potentially increasing the yield. Thus, HTS-1 is highly valuable in the study of floral development in wheat. In this study, we conducted RNA sequencing of the transcriptomes of the pistillody stamen (PS) and the pistil (P) from HTS-1 plants, and the stamen (S) from the non-pistillody control variety Chinese Spring TP to gain insights into pistil and stamen development in wheat.ResultsApproximately 40 Gb of processed reads were obtained from PS, P, and S. De novo assembly yielded 121,210 putative unigenes, with a mean length of 695 bp. Among these high-quality unigenes, 59,199 (48.84%) had at least one significant match with an existing gene model. A total of 23, 263, and 553 differentially expressed genes were identified in PS vs. P, PS vs. S, and P vs. S, respectively, with differences in expression greater than five-fold. Among the differentially expressed genes, 206 were highly correlated with stamen and pistil development. These genes include WM27B, DL, YAB1, YABBY4, WM 5, CER 1, and WBLH1, which have been implicated in flower development. The expression patterns of 25 differentially expressed genes were confirmed through quantitative real-time reverse transcription PCR.ConclusionsAnalysis of this transcriptome resource enabled us to characterize gene expression profiles, examine differential gene expression, and produce a candidate gene list related to wheat stamen and pistil development. This work is significant for the development of genomic resources for wheat, and provides important insights into the molecular mechanisms of wheat stamen and pistil development.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-015-1453-0) contains supplementary material, which is available to authorized users.

Flower development and sex specification in wild grapevine.

BackgroundWild plants of Vitis closely related to the cultivated grapevine (V. v. vinifera) are believed to have been first domesticated 10,000 years BC around the Caspian Sea. V. v. vinifera is hermaphrodite whereas V. v. sylvestris is a dioecious species. Male flowers show a reduced pistil without style or stigma and female flowers present reflexed stamens with infertile pollen. V. vinifera produce perfect flowers with all functional structures. The mechanism for flower sex determination and specification in grapevine is still unknown.ResultsTo understand which genes are involved during the establishment of male, female and complete flowers, we analysed and compared the transcription profiles of four developmental stages of the three genders. We showed that sex determination is a late event during flower development and that the expression of genes from the ABCDE model is not directly correlated with the establishment of sexual dimorphism. We propose a temporal comprehensive model in which two mutations in two linked genes could be players in sex determination and indirectly establish the Vitis domestication process. Additionally, we also found clusters of genes differentially expressed between genders and between developmental stages that suggest a role involved in sex differentiation. Also, the detection of differentially transcribed regions that extended existing gene models (intergenic regions) between sexes suggests that they may account for some of the variation between the subspecies.ConclusionsThere is no evidence of differences of expression levels in genes from the ABCDE model that could explain the shift from hermaphroditism to dioecy. We propose that sex specification occurs after floral organ identity has been established and therefore, sex determination genes might be having an effect downstream of the ABCDE model genes.For the first time a full transcriptomic analysis was performed in different flower developmental stages in the same individual. Our experimental approach enabled us to create a comprehensive catalogue of transcribed genes across developmental stages and genders that will contribute for future work in sex determination in seed plants.Electronic supplementary materialThe online version of this article (doi:10.1186/1471-2164-15-1095) contains supplementary material, which is available to authorized users.

RNA-Seq reveals leaf cuticular wax-related genes in Welsh onion.

The waxy cuticle plays a very important role in plant resistance to various biotic and abiotic stresses and is an important characteristic of Welsh onions. Two different types of biangan Welsh onions (BG) were selected for this study: BG, a wild-type covered by wax, which forms a continuous lipid membrane on its epidermal cells, and GLBG, a glossy mutant of BG whose epidermal cells are not covered by wax. To elucidate the waxy cuticle-related gene expression changes, we used RNA-Seq to compare these two Welsh onion varieties with distinct differences in cuticular wax. The de novo assembly yielded 42,881 putative unigenes, 25.41% of which are longer than 1,000 bp. Among the high-quality unique sequences, 22,289 (52.0%) had at least one significant match to an existing gene model. A total of 798 genes, representing 1.86% of the total putative unigenes, were differentially expressed between these two Welsh onion varieties. The expression patterns of four important unigenes that are related to waxy cuticle biosynthesis were confirmed by RT-qPCR and COG class annotation, which demonstrated that these genes play an important role in defense mechanisms and lipid transport and metabolism. To our knowledge, this study is the first exploration of the Welsh onion waxy cuticle. These results may help to reveal the molecular mechanisms underlying the waxy cuticle and will be useful for waxy gene cloning, genetics and breeding as well as phylogenetic and evolutionary studies of the Welsh onion.

A bioinformatics approach to reanalyze the genome annotation of kinetoplastid protozoan parasite Leishmania donovani

Leishmania donovani is a kinetoplastid protozoan parasite which causes the fatal disease visceral leishmaniasis in humans. Genome sequencing of L. donovani revealed information about the arrangement of genes and genome architecture. After curation of the genome sequence, many genes in L. donovani were assigned as truncated or “partial” genes by the genome sequencing group. In the present study, we have carried out an extensive analysis and attempted to improve the gene models of these partial genes. Our analysis resulted in the identification of 308 partial genes in L. donovani, which were further categorized as C-terminal extensions, joining of genes, tandemly repeated paralogs and wrong chromosomal assignments. We have analyzed each of these genes from these categories and have improved the annotation of existing gene models in L. donovani. Some of these corrections have been confirmed by mass spectrometry derived peptide data from our previous comparative proteogenomics study in L. donovani.

Transcript Assembly and Quantification by RNA-Seq Reveals Differentially Expressed Genes between Soft-Endocarp and Hard-Endocarp Hawthorns

Hawthorn (Crataegus spp.) is an important pome with a long history as a fruit, an ornamental, and a source of medicine. Fruits of hawthorn are marked by hard stony endocarps, but a hawthorn germplasm with soft and thin endocarp was found in Liaoning province of China. To elucidate the molecular mechanism underlying the soft endocarp of hawthorn, we conducted a de novo assembly of the fruit transcriptome of Crataegus pinnatifida and compared gene expression profiles between the soft-endocarp and the hard-endocarp hawthorn varieties. De novo assembly yielded 52,673 putative unigenes, 20.4% of which are longer than 1,000 bp. Among the high-quality unique sequences, 35,979 (68.3%) had at least one significant match to an existing gene model. A total of 1,218 genes, represented 2.31% total putative unigenes, were differentially expressed between the soft-endocarp hawthorn and the hard-endocarp hawthorn. Among these differentially expressed genes, a number of lignin biosynthetic pathway genes were down-regulated while almost all the flavonoid biosynthetic pathway genes were strongly up-regulated, concomitant with the formation of soft endocarp. In addition, we have identified some MYB and NAC transcription factors that could potentially control lignin and flavonoid biosynthesis. The altered expression levels of the genes encoding lignin biosynthetic enzymes, MYB and NAC transcription factors were confirmed by quantitative RT-PCR. This is the first transcriptome analysis of Crataegus genus. The high quality ESTs generated in this study will aid future gene cloning from hawthorn. Our study provides important insights into the molecular mechanisms underlying soft endocarp formation in hawthorn.

Moving from unsequenced to sequenced genome: Reanalysis of the proteome of Leishmania donovani

The kinetoplastid protozoan parasite, Leishmania donovani, is the causative agent of kala azar or visceral leishmaniasis. Kala azar is a severe form of leishmaniasis that is fatal in the majority of untreated cases. Studies on proteomic analysis of L. donovani thus far have been carried out using homology-based identification based on related Leishmania species (L. infantum, L. major and L. braziliensis) whose genomes have been sequenced. Recently, the genome of L. donovani was fully sequenced and the data became publicly available. We took advantage of the availability of its genomic sequence to carry out a more accurate proteogenomic analysis of L. donovani proteome using our previously generated dataset. This resulted in identification of 17,504 unique peptides upon database-dependent search against the annotated proteins in L. donovani. These peptides were assigned to 3999 unique proteins in L. donovani. 2296 proteins were identified in both the life stages of L. donovani, while 613 and 1090 proteins were identified only from amastigote and promastigote stages, respectively. The proteomic data was also searched against six-frame translated L. donovani genome, which led to 255 genome search-specific peptides (GSSPs) resulting in identification of 20 novel genes and correction of 40 existing gene models in L. donovani. Leishmania donovani genome sequencing was recently completed, which permitted us to use a proteogenomic approach to map its proteome and to carry out annotation of it genome. This resulted in mapping of 50% (3999 proteins) of L. donovani proteome. Our study identified 20 novel genes previously not predicted from the L. donovani genome in addition to correcting annotations of 40 existing gene models. The identified proteins may help in better understanding of stage-specific protein expression profiles in L. donovani and to identify novel stage-specific drug targets in L. donovani which could be used in the treatment of leishmaniasis. This article is part of a Special Issue entitled: Trends in Microbial Proteomics.

A Proteogenomic Survey of the Medicago truncatula Genome

Peptide sequencing by computational assignment of tandem mass spectra to a database of putative protein sequences provides an independent approach to confirming or refuting protein predictions based on large-scale DNA and RNA sequencing efforts. This use of mass spectrometrically-derived sequence data for testing and refining predicted gene models has been termed proteogenomics. We report herein the application of proteogenomic methodology to a database of 10.9 million tandem mass spectra collected over a period of two years from proteolytically generated peptides isolated from the model legume Medicago truncatula. These spectra were searched against a database of predicted M. truncatula protein sequences generated from public databases, in silico gene model predictions, and a whole-genome six-frame translation. This search identified 78,647 distinct peptide sequences, and a comparison with the publicly available proteome from the recently published M. truncatula genome supported translation of 9,843 existing gene models and identified 1,568 novel peptides suggesting corrections or additions to the current annotations. Each supporting and novel peptide was independently validated using mRNA-derived deep sequencing coverage and an overall correlation of 93% between the two data types was observed. We have additionally highlighted examples of several aspects of structural annotation for which tandem MS provides unique evidence not easily obtainable through typical DNA or RNA sequencing. Proteogenomic analysis is a valuable and unique source of information for the structural annotation of genomes and should be included in such efforts to ensure that the genome models used by biologists mirror as accurately as possible what is present in the cell.

Gene2DGE: A Perl Package for Gene Model Renewal with Digital Gene Expression Data

For transcriptome analysis, it is critical to precisely define all the transcripts across the whole genome. More and more digital gene expression (DGE) scannings have indicated the presence of huge amount of novel transcripts in addition to the known gene models. However, almost all these studies still depend crucially on existing annotation. Here, we present Gene2DGE, a Perl software package for gene model renewal with DGE data. We applied Gene2DGE to the mouse blastomere transcriptome, and defined 98,532 read-enriched regions (RERs) by read clustering supported by more than four reads for each base pair. Taking advantage of this ab initio method, we refined 2,104 exonic regions (4% of a total of 48,501 annotated transcribed regions) with remarkable extension into un-annotated regions (>50bp). For 5% of uniquely mapped reads falling within intron regions, we identified 13,291 additional possible exons. As a result, we renewed 4,788 gene models, which account for 39% of a total of 12,277 transcribed genes. Furthermore, we identified 12,613 intergenic RERs, suggesting the possible presence of novel genes outside the existing gene models. In this study, therefore, we have developed a suitable tool for renewal of known gene models by ab initio prediction in transcriptome dissection. The Gene2DGE package is freely available at http://bighapmap.big.ac.cn/.

Good and Poor Response Gene Expression Signatures to Proteasome Inhibitors Using a Mouse Model of Multiple Myeloma

Abstract 1843The proteasome inhibitor bortezomib (Bz) has been used extensively and with much success in the treatment of multiple myeloma (MM) patients; however, patients eventually relapse, many as non-responders to subsequent treatments with Bz making drug resistance a significant problem. Here we utilized cell lines created using a iMycCa/Bcl-xL transgenic mouse model of MM (Cheung, et al. J Clin Invest (2004) 113: 1763) to identify 1) gene expression signatures of Bz response, 2) differences in gene expression between sensitive and resistant cell lines, and 3) cytogenetic abnormalities associated with Bz sensitive and resistant phenotypes.The iMycCa/Bcl-xL transgenic mice develop plasma cell tumors with 100% penetrance and have shown strikingly strong similarities to human MM by extensive gene expression profiling (GEP), spectral karyotyping and histology (Boylan, et al. Cancer Res (2007) 67: 4069). Six cell lines created from these mice were dose escalated with Bz over approximately six months to create Bz resistant (BzR) cell lines with approximately 5–8 fold increase in IC50 to Bz compared to their sensitive counterparts. The BzR characteristics were stable, as lines grown in the absence of drug for as long as 6 months maintained drug resistance upon subsequent challenge. Notably, BzR lines showed cross resistance to other investigational proteasome inhibitors (MLN9708 and carfilzomib) while maintaining sensitivity to other chemotherapeutic agents (dexamethasone and melphalan), suggesting a common mechanism of emerging resistance to proteasome inhibitors.The results of GEP of these mouse tumor cell lines treated with Bz were compared with a recently published human drug trial where GEP was completed prior to and 48 hours after a “test dose” of Bz was administered to patients (Shaughnessy, et al. Blood (2011), ahead of print). In the mouse tumor cell lines, 116 genes were differentially expressed upon in vitro Bz treatment (p=0.001, ≥1.5 fold change). Between the mouse and human drug response data sets was an overlapping common 27-gene signature (p=1×10−25, Fishers exact test) of Bz-induced expression changes that has not previously been described.Time points were collected in these mouse cell line GEP experiments at 0, 2, 8, 16, and 24 hours after Bz treatment. A comparison of the Bz sensitive and derived BzR lines prior to drug treatment revealed a 50 gene signature (p=0.05, ≥2 fold change) that distinguishes three pairs of sensitive and resistant lines. Gene-set enrichment analyses have revealed significant pathways that are differentially regulated in the sensitive and resistant responses. Additional GEP differences were seen when time course expression patterns were examined from Bz sensitive compared to resistant tumor lines. Thus, GEP signatures that distinguish tumor lethality from resistance were identified both prior to Bz treatment, as well as in the early response to Bz. In addition, array comparative genomic hybridization on 4 pairs of mouse Bz sensitive and established BzR lines revealed not only gross differences in copy number between the differentially responding groups of cells but copy number abnormalities that may be unique to the emerging resistance.Taken together, these data indicate that this model is useful for the identification of good and poor Bz response signatures in MM. These signatures are currently being evaluated in human tumor cells from single agent bortezomib phase II and phase III clinical trials. Because the in vitro adapted tumor mouse lines can be genetically manipulated using lentiviral vectors, this model can be used as a preclinical platform to validate existing gene models with respect to Bz response, something that cannot be done using human patients. Subsequent transfer of manipulated lines into syngeneic, immunocompetent recipients can further test Bz response in vivo presenting a significant advantage of this robust mouse MM model system over other in vitro systems. Disclosures:Stessman:Millennium: The Takeda Oncology Company: Research Funding. Mansoor:Millennium: The Takeda Oncology Company: Research Funding. Janz:Millennium: The Takeda Oncology Company: Research Funding. Van Ness:Millennium: The Takeda Oncology Company: Research Funding.

Identifying novel genes in C. elegans using SAGE tags

BackgroundDespite extensive efforts devoted to predicting protein-coding genes in genome sequences, many bona fide genes have not been found and many existing gene models are not accurate in all sequenced eukaryote genomes. This situation is partly explained by the fact that gene prediction programs have been developed based on our incomplete understanding of gene feature information such as splicing and promoter characteristics. Additionally, full-length cDNAs of many genes and their isoforms are hard to obtain due to their low level or rare expression. In order to obtain full-length sequences of all protein-coding genes, alternative approaches are required.ResultsIn this project, we have developed a method of reconstructing full-length cDNA sequences based on short expressed sequence tags which is called sequence tag-based amplification of cDNA ends (STACE). Expressed tags are used as anchors for retrieving full-length transcripts in two rounds of PCR amplification. We have demonstrated the application of STACE in reconstructing full-length cDNA sequences using expressed tags mined in an array of serial analysis of gene expression (SAGE) of C. elegans cDNA libraries. We have successfully applied STACE to recover sequence information for 12 genes, for two of which we found isoforms. STACE was used to successfully recover full-length cDNA sequences for seven of these genes.ConclusionsThe STACE method can be used to effectively reconstruct full-length cDNA sequences of genes that are under-represented in cDNA sequencing projects and have been missed by existing gene prediction methods, but their existence has been suggested by short sequence tags such as SAGE tags.

Rnnotator: an automated de novo transcriptome assembly pipeline from stranded RNA-Seq reads

BackgroundComprehensive annotation and quantification of transcriptomes are outstanding problems in functional genomics. While high throughput mRNA sequencing (RNA-Seq) has emerged as a powerful tool for addressing these problems, its success is dependent upon the availability and quality of reference genome sequences, thus limiting the organisms to which it can be applied.ResultsHere, we describe Rnnotator, an automated software pipeline that generates transcript models by de novo assembly of RNA-Seq data without the need for a reference genome. We have applied the Rnnotator assembly pipeline to two yeast transcriptomes and compared the results to the reference gene catalogs of these organisms. The contigs produced by Rnnotator are highly accurate (95%) and reconstruct full-length genes for the majority of the existing gene models (54.3%). Furthermore, our analyses revealed many novel transcribed regions that are absent from well annotated genomes, suggesting Rnnotator serves as a complementary approach to analysis based on a reference genome for comprehensive transcriptomics.ConclusionsThese results demonstrate that the Rnnotator pipeline is able to reconstruct full-length transcripts in the absence of a complete reference genome.

RNA-Seq Atlas of Glycine max: a guide to the soybean transcriptome.

BackgroundNext generation sequencing is transforming our understanding of transcriptomes. It can determine the expression level of transcripts with a dynamic range of over six orders of magnitude from multiple tissues, developmental stages or conditions. Patterns of gene expression provide insight into functions of genes with unknown annotation.ResultsThe RNA Seq-Atlas presented here provides a record of high-resolution gene expression in a set of fourteen diverse tissues. Hierarchical clustering of transcriptional profiles for these tissues suggests three clades with similar profiles: aerial, underground and seed tissues. We also investigate the relationship between gene structure and gene expression and find a correlation between gene length and expression. Additionally, we find dramatic tissue-specific gene expression of both the most highly-expressed genes and the genes specific to legumes in seed development and nodule tissues. Analysis of the gene expression profiles of over 2,000 genes with preferential gene expression in seed suggests there are more than 177 genes with functional roles that are involved in the economically important seed filling process. Finally, the Seq-atlas also provides a means of evaluating existing gene model annotations for the Glycine max genome.ConclusionsThis RNA-Seq atlas extends the analyses of previous gene expression atlases performed using Affymetrix GeneChip technology and provides an example of new methods to accommodate the increase in transcriptome data obtained from next generation sequencing. Data contained within this RNA-Seq atlas of Glycine max can be explored at http://www.soybase.org/soyseq.

Characterization of microsatellites and gene contents from genome shotgun sequences of mungbean (Vigna radiata (L.) Wilczek).

BackgroundMungbean is an important economical crop in Asia. However, genomic research has lagged behind other crop species due to the lack of polymorphic DNA markers found in this crop. The objective of this work is to develop and characterize microsatellite or simple sequence repeat (SSR) markers from genome shotgun sequencing of mungbean.ResultWe have generated and characterized a total of 470,024 genome shotgun sequences covering 100.5 Mb of the mungbean (Vigna radiata (L.) Wilczek) genome using 454 sequencing technology. We identified 1,493 SSR motifs that could be used as potential molecular markers. Among 192 tested primer pairs in 17 mungbean accessions, 60 loci revealed polymorphism with polymorphic information content (PIC) values ranging from 0.0555 to 0.6907 with an average of 0.2594. Majority of microsatellite markers were transferable in Vigna species, whereas transferability rates were only 22.90% and 24.43% in Phaseolus vulgaris and Glycine max, respectively. We also used 16 SSR loci to evaluate phylogenetic relationship of 35 genotypes of the Asian Vigna group. The genome survey sequences were further analyzed to search for gene content. The evidence suggested 1,542 gene fragments have been sequence tagged, that fell within intersected existing gene models and shared sequence homology with other proteins in the database. Furthermore, potential microRNAs that could regulate developmental stages and environmental responses were discovered from this dataset.ConclusionIn this report, we provided evidence of generating remarkable levels of diverse microsatellite markers and gene content from high throughput genome shotgun sequencing of the mungbean genomic DNA. The markers could be used in germplasm analysis, accessing genetic diversity and linkage mapping of mungbean.

Discovery and revision of Arabidopsis genes by proteogenomics

Gene annotation underpins genome science. Most often protein coding sequence is inferred from the genome based on transcript evidence and computational predictions. While generally correct, gene models suffer from errors in reading frame, exon border definition, and exon identification. To ascertain the error rate of Arabidopsis thaliana gene models, we isolated proteins from a sample of Arabidopsis tissues and determined the amino acid sequences of 144,079 distinct peptides by tandem mass spectrometry. The peptides corresponded to 1 or more of 3 different translations of the genome: a 6-frame translation, an exon splice-graph, and the currently annotated proteome. The majority of the peptides (126,055) resided in existing gene models (12,769 confirmed proteins), comprising 40% of annotated genes. Surprisingly, 18,024 novel peptides were found that do not correspond to annotated genes. Using the gene finding program AUGUSTUS and 5,426 novel peptides that occurred in clusters, we discovered 778 new protein-coding genes and refined the annotation of an additional 695 gene models. The remaining 13,449 novel peptides provide high quality annotation (>99% correct) for thousands of additional genes. Our observation that 18,024 of 144,079 peptides did not match current gene models suggests that 13% of the Arabidopsis proteome was incomplete due to approximately equal numbers of missing and incorrect gene models.

Whole genome shotgun sequencing ofBrassica oleraceaand its application to gene discovery and annotation inArabidopsis

Through comparative studies of the model organism Arabidopsis thaliana and its close relative Brassica oleracea, we have identified conserved regions that represent potentially functional sequences overlooked by previous Arabidopsis genome annotation methods. A total of 454,274 whole genome shotgun sequences covering 283 Mb (0.44 x) of the estimated 650 Mb Brassica genome were searched against the Arabidopsis genome, and conserved Arabidopsis genome sequences (CAGSs) were identified. Of these 229,735 conserved regions, 167,357 fell within or intersected existing gene models, while 60,378 were located in previously unannotated regions. After removal of sequences matching known proteins, CAGSs that were close to one another were chained together as potentially comprising portions of the same functional unit. This resulted in 27,347 chains of which 15,686 were sufficiently distant from existing gene annotations to be considered a novel conserved unit. Of 192 conserved regions examined, 58 were found to be expressed in our cDNA populations. Rapid amplification of cDNA ends (RACE) was used to obtain potentially full-length transcripts from these 58 regions. The resulting sequences led to the creation of 21 gene models at 17 new Arabidopsis loci and the addition of splice variants or updates to another 19 gene structures. In addition, CAGSs overlapping already annotated genes in Arabidopsis can provide guidance for manual improvement of existing gene models. Published genome-wide expression data based on whole genome tiling arrays and massively parallel signature sequencing were overlaid on the Brassica-Arabidopsis conserved sequences, and 1399 regions of intersection were identified. Collectively our results and these data sets suggest that several thousand new Arabidopsis genes remain to be identified and annotated.