PMC:540072 JSONTXT 3 Projects

ECgene: genome annotation for alternative splicing Abstract ECgene provides annotation for gene structure, function and expression, taking alternative splicing events into consideration. The gene-modeling algorithm combines the genome-based expressed sequence tag (EST) clustering and graph-theoretic transcript assembly procedures. The website provides several viewers and applications that have many unique features useful for the analysis of the transcript structure and gene expression. The summary viewer shows the gene summary and the essence of other annotation programs. The genome browser and the transcript viewer are available for comparing the gene structure of splice variants. Changes in the functional domains by alternative splicing can be seen at a glance in the transcript viewer. We also provide two unique ways of analyzing gene expression. The SAGE tags deduced from the assembled transcripts are used to delineate quantitative expression patterns from SAGE libraries available publically. Furthermore, the cDNA libraries of EST sequences in each cluster are used to infer qualitative expression patterns. It should be noted that the ECgene website provides annotation for the whole transcriptome, not just the alternatively spliced genes. Currently, ECgene supports the human, mouse and rat genomes. The ECgene suite of tools and programs is available at http://genome.ewha.ac.kr/ECgene/. INTRODUCTION Alternative splicing (AS) is a major mechanism of increasing transcriptome diversity in eucaryotic genomes (1–3). Recent studies on AS estimated that 40–70% of human genes show evidence of encoding transcripts that are alternatively spliced (4–6). Numerous databases on AS, mostly concentrating on specific aspects of gene structure (7–14), have been published so far. Databases on AS are generated either by data mining the experimental databases such as GenBank, Swiss-Prot/TrEMBL and Medline, or by comparing sequence alignments. The former includes ASDB (14), Xpro (12) and AEdb in ASD (7). Computational approaches are even more diverse. Many attempts compared the expressed sequence tag (EST) alignments with mRNA or known gene sequences (10,12,13,15,16). However, examining genomic alignment of EST and mRNA sequences has many advantages once the genome map is available. This includes AltExtron (17), TAP (5), ASAP (8) and altGraphX (18). Notably, the European Bioinformatics Institute (EBI) launched the Alternative Splicing Database (ASD) consortium to annotate AS events (7). Currently, three major databases are available in ASD—AltSplice and AltExtron from computational pipeline and AEdb which is manually generated from the literature. AceView at the NCBI is another website that provides ample information on alternatively spliced genes (D. Thierry-Mieg, J. Thierry-Mieg, M. Potdevin and M. Sienkiewicz, unpublished data; http://www.ncbi.nih.gov/IEB/Research/Acembly/). Recently, we developed a gene prediction program, ECgene (Gene prediction by EST Clustering), taking alternative splicing into consideration. The algorithm combines the genome-based EST clustering and graph-theoretic transcript assembly procedures in a coherent fashion. A brief summary of the algorithm was described with the ASmodeler work, the web server version (19). Further details of the algorithm will be published elsewhere (N. Kim, S. Shin and S. Lee, manuscript submitted). In this paper, we describe the ECgene database and website with several application programs. Besides the gene prediction algorithm, our viewers and applications have many unique features useful for analyzing the transcript structure, gene function and expression pattern. ECgene DATABASE The ECgene algorithm has been applied to the human, mouse and rat genomes to produce transcriptomes that include alternative splicing events. ECgene transcriptome must be one of the most complete versions even though it contains many unreal transcripts. We classify the transcript models into three groups according to reliability. ECgene Part A represents transcripts of almost RefSeq quality with a clone evidence covering all exons of the transcript. Part B includes transcripts of slightly lower quality (but still highly probable). Their transcript structure requires concatenation of minimum two clones to recover all exons in the transcript. No single clone is available to be full-length for transcript model. All other transcripts belong to Part C and are of low reliability. Their gene structure may be questionable since it requires concatenation of more than two clones. However, individual event of alternative splicing implied in the transcripts should be real unless the genomic alignment of mRNA/EST sequences is wrong. They have a fair chance to be real transcripts with more sequence data. Statistics on the number of genes and transcripts is available at the website. WEB INTERFACE AND APPLICATIONS The ECgene website contains several independent but closely related applications—the summary viewer, genome browser, alignment viewer, ECfunction and ECexpression. The summary viewer shows the overall picture of the gene. Each application page provides more detailed information and diverse options to explore a specific aspect of the gene. Summary viewer The summary viewer shows the gene summary and the essence of other annotation programs. It supports querying by gene symbol, accession number of mRNA/EST and the ECgene ID. Figure 1 shows a part of output from the ECgene summary page. The output page shows a brief summary of the gene and the cluster, a graphic illustration of the transcript structure, the summary table of splice variants, functional analysis from the InterPro (20) and Swiss-Prot databases (21), and gene expression analysis using SAGE and cDNA libraries. Links to other well-known databases, predicted mRNA and protein sequences, and sequence members in the cluster are shown at the bottom of the page. Graphic images are from ECfunction and ECexpression, which are described in later sections. Links to various viewers are given next to the gene structure in Figure 1. We provide our own genome browser and the transcript/motif/domain viewer whose details are described in the following sections. The alignment viewer (picture not shown) can be used to see the alignment and tissue source of ESTs (organ and cancer). Viewers via these links use the default options, and users should access the application page for various options. Clicking on the variant in the picture or in the table opens the mRNA page. This page provides more detailed information on the selected transcript in a similar interface. Domain/motif information is added in the picture. ECgene genome browser The transcript structure and cluster members can be seen via the ECgene genome browser that adds ECgene models as custom tracks in the UCSC genome browser. Figure 2 shows the transcript structure of the BRCA2 gene using the ECgene genome browser available at http://genome.ewha.ac.kr/ECgene/gbr/. The GUI design is almost identical to the UCSC genome browser as shown in Figure 2. The most useful feature would be the option of showing the EST alignment that adds each transcript model and member sequences as a separate custom track. The title line includes a brief summary of the transcript model and clones. Clicking on the transcript or the title line expands the picture to show the alignment in the pack-view mode. We also provide the option of hiding unspliced alignments since many of them are incomplete or are artifacts. Furthermore, one can see the result of UniGene clustering for comparison. It is just a collection of alignments without transcript models. We find that all 56 sequences in the UniGene cluster for the BRCA2 gene appear in this region. ECfunction—transcript/domain/motif viewer ECfunction available at http://genome.ewha.ac.kr/ECgene/ECfunction/ collects information related to gene function. It includes the InterPro domains, transmembrane helices, signal peptides, BLAST hits and GO annotations. ECfunction has a unique transcript viewer as shown in Figure 3. It shows the transcript structure in the mRNA coordinates where introns are depicted as short lines of fixed length. Overlapping exons are grouped together to show all splice variants on a single ruler scale given above the picture. The nucleotide ranges on the exons indicate exons from splice site variation so that one can see the difference of splice variants at a glance. It should be noted that merging overlapping exons to generate the master coordinate might create unrealistically long exons in case the introns are retained in one of the transcripts. To remedy this problem, we provide the option of removing intron retention events. Alternatively, the user may choose the specific isoforms to draw in the transcript viewer. Being on the mRNA coordinate, the transcript viewer can be used to show the functional domains and motifs. This enables the user to recognize immediately any loss of functional domains due to alternative splicing. ECexpression ECexpression available at http://genome.ewha.ac.kr/ECgene/ECexpression/ analyzes the gene expression pattern using ∼260 SAGE and ∼8600 cDNA libraries available publically. Our SAGE analysis is much different from the NCBI's SAGEmap (22) in that SAGE tags are extracted from the assembled transcript models, not from EST sequences in the cluster. The clustering is also different from the UniGene even though both are genome-based at this point. We explain here the output pictures briefly. Figure 4 shows the normalized tag frequencies (i.e. tags per million count) for 28 organs with public SAGE data. In cases the splice variants have different SAGE tags, SAGE tags can be used to find variant-specific gene expression. We divide the normal and cancer libraries whose tag frequencies are depicted in the upper and lower half of the picture, respectively. This layout makes it particularly easy to find any tissue-specific or cancer-specific isoforms that have tremendous value for clinical application. Gene expression from cDNA libraries is quite similar except that the graph represents the number of EST sequences from the specified tissue in the cluster. We manually classified ∼8600 human cDNA libraries in terms of tissue (organ), pathology and developmental stage. Gene expression predicted by using EST members is mostly qualitative since many cDNA libraries are normalized or subtracted to find genes with low expression level. However, the coverage of cDNA libraries is much more extensive as can be seen in the number of libraries available publically. FUTURE DIRECTIONS ECgene is an ongoing project. It aims to be one of the reference sites for genome annotation. A number of new features and applications are currently under development. Gene profiler, ortholog finder and ChimerDB are among the applications to be completed in the near future. We plan to expand and optimize the system resources to shorten the response time. Other model organisms will be added too.