Detection of four isomers of the human cytomegalovirus genome using nanopore long-read sequencing.

Human cytomegalovirus has a linear DNA genome with a total length of approximately 235 kb. This large genome is divided into two domains, "Long" and "Short". There are four isomers of the cytomegalovirus genome with different orientations of each domain. To confirm the presence of four types of isomers, it is necessary to identify the sequence of the junction between the domains. However, due to the presence of repeat sequences, it is difficult to determine the junction sequences by next-generation sequencing analysis. To solve this problem, long-read sequencing was performed using the Oxford Nanopore sequencer and the junctions were successfully identified in four isomers in strain Merin and ATCC-2011-3. Nanopore sequencing also revealed the presence of multiple copies of the "a" sequence (a-seq) in the junctions, indicating the diversity of the junction sequences. These results strongly suggest that long-read sequencing using the nanopore sequencer would be beneficial for identifying the complex structure of the cytomegalovirus genome.

Identification and clinical significance of somatic oncogenic mutations in epithelial ovarian cancer.

OBJECTIVE: Epithelial ovarian cancer (EOC) is a heterogeneous disease with diverse clinicopathological features and behaviors, and its heterogeneity may be concerned with the accumulation of multiple somatic oncogenic mutations. The major goals of this study are to systematically perform the comprehensive mutational profiling in EOC patients, and investigate the associations between somatic mutations and clinicopathological characteristics. METHODS: A total of 80 surgical specimens were obtained from EOC patients who had previously undergone primary debulking surgery, and genomic DNAs were extracted from fresh-frozen tissues. We investigated mutational status in hot spot regions of 50 cancer-related genes by targeted next-generation sequencing using an Ion AmpliSeq Cancer Hotspot Panel v2 Kit. RESULTS: Validated mutations were detected in 66 of the 80 tumors (82.5%). The five most frequently mutated genes were TP53 (43.8%), PIK3CA (27.5%), KRAS (23.8%), PTEN (10%) and CTNNB1 (10%). PTEN and CTNNB1 mutations were associated with younger age. PIK3CA1, KRAS and CTNNB1 mutations were observed in early-stage, whereas TP53 mutations were more common in advanced stage. Significant associations were observed between TP53 mutation and serous carcinoma, and between KRAS mutation and mucinous carcinoma. Both PIK3CA mutation and CTNNB1 mutation were also significantly associated with endometrioid and clear cell carcinoma. The patients with PIK3CA and KRAS mutations were significantly associated with favorable progression free survival (PFS). In particular, PIK3CA mutations had more significant associations with favorable PFS than PIK3CA wild-type in the endometrioid subtype (P = 0.012). Patients with mutations only in TP53 were significantly associated with worse PFS. CONCLUSION: EOCs were heterogeneous at the genomic level and harbored somatic oncogenic mutations. Our molecular profiling may have the potential for becoming a novel stratification within histological subtypes of EOC. Further studies are needed to define molecular classification for improved clinical outcomes and treatment of EOC patients in future.

Tumor ß-catenin expression is associated with immune evasion in non-small cell lung cancer with high tumor mutation burden.

ß-catenin expression by tumor cells suppressed dendritic cell recruitment to the tumor microenvironment in a melanoma model, resulting in fewer tumor-infiltrating lymphocytes. Immunohistochemistry was used in the present study to examine the association between the expression of ß-catenin and tumor infiltrating lymphocytes and CD11c+ cells in 122 patients with non-small cell lung cancer (NSCLC), who underwent radical surgery. ß-catenin was positive in 24% of NSCLC tumors compared with 59% of squamous cell carcinomas and 11% of adenocarcinomas. There was no significant association between the expression of ß-catenin and the frequency of CD8+ cell infiltration into tumor tissues, including the stroma. Conversely, the infiltration of CD8+ cells into tumor nests was significantly lower in ß-catenin-positive cases compared with that in negative ß-catenin cases. Similarly, CD11c+ cell infiltration was significantly lower in the ß-catenin-positive group. The ß-catenin-positive group had shorter overall survival and recurrence-free survival times compared with that in the negative group. Furthermore, ß-catenin-positive NSCLC had a high tumor mutation burden, but tended to have a low expression of programmed death-ligand 1. In conclusion, the expression of ß-catenin in NSCLC was negatively associated with CD11c+ cells and cytotoxic T cell infiltration at the tumor site and had a tendency towards a poor prognosis.

Clinical relevance of oncogenic driver mutations identified in endometrial carcinoma.

PURPOSE: Endometrial carcinoma (EC) is a clinically heterogeneous disease characterized by a number of different histological subtypes, and its heterogeneity may be involved in the accumulation of multiple genetic alterations. The aim of this work was to investigate the comprehensive mutational profile of EC tumors, and examine the associations between somatic mutations and clinicopathological features or survival in EC patients. METHODS: A total of 100 surgical tumors were obtained from EC patients who had previously undergone surgery. Genomic DNA samples extracted from fresh-frozen tissues were analyzed using the Ion AmpliSeq Cancer Hotspot Panel v2 Kit, covering 50 tumor-related genes. RESULTS: Validated mutations were detected in 91 of the 100 tumors (91%) and identified in eight of the most frequently mutated genes, namely PTEN (57%), PIK3CA (51%), TP53 (30%), KRAS (23%), CTNNB1 (21%), FBFR2 (13%), FBXW7(10%) and RB1 (9%). PTEN mutations were found to associated with young age (< 60), early-stage, endometrioid histology, non-recurrence and better overall survival (OS). CTNNB1 mutations were associated with young age, endometrioid histology and better OS. On the other hands, TP53 mutations were associated with late-stage, non-endometrioid histology, high-grade, recurrence and worse OS. FBWX7 mutations were associated with late-stage, vascular invasion and lymph node metastasis. FGFR2 mutations correlated with deep (≥ 1/2) myometrial invasion. CONCLUSION: Our comprehensive mutational profile will be useful for understanding and evaluating the molecular characteristics of EC tumors, and may lead to the establishment of novel treatment strategies that improve the survival of patients with EC in the future.

Clinical implication of oncogenic somatic mutations in early-stage cervical cancer with radical hysterectomy.

It is well known that tumour initiation and progression are primarily an accumulation of genetic mutations. The mutation status of a tumour may predict prognosis and enable better selection of targeted therapies. In the current study, we analysed a total of 55 surgical tumours from stage IB-IIB cervical cancer (CC) patients who had undergone radical hysterectomy including pelvic lymphadenectomy, using a cancer panel covering 50 highly mutated tumorigenesis-related genes. In 35 patients (63.6%), a total 52 mutations were detected (58.3% in squamous cell carcinoma, 73.7% in adenocarcinoma), mostly in PIK3CA (34.5%) and KRAS and TP53 (9.1%). Being mutation-positive was significantly correlated with pelvic lymph node (PLN) metastasis (P = 0.035) and tended to have a worse overall survival (P = 0.076). In particular, in the patients with squamous cell carcinoma, there was a significant association between being mutation-positive and relapse-free survival (P = 0.041). The patients with PLN metastasis had a significantly worse overall survival than those without (P = 0.006). These results indicate that somatic mutation status is a predictive biomarker for PLN metastasis in early-stage CC, and is consequently related to poor prognosis. Therefore, comprehensive genetic mutations, rather than a single genetic mutation, should be examined widely in order to identify novel genetic indicators with clinical usefulness.

Non-pathogenic Escherichia coli acquires virulence by mutating a growth-essential LPS transporter.

The molecular mechanisms that allow pathogenic bacteria to infect animals have been intensively studied. On the other hand, the molecular mechanisms by which bacteria acquire virulence functions are not fully understood. In the present study, we experimentally evaluated the evolution of a non-pathogenic strain of Escherichia coli in a silkworm infection model and obtained pathogenic mutant strains. As one cause of the high virulence properties of E. coli mutants, we identified amino acid substitutions in LptD (G580S) and LptE (T95I) constituting the lipopolysaccharide (LPS) transporter, which translocates LPS from the inner to the outer membrane and is essential for E. coli growth. The growth of the LptD and LptE mutants obtained in this study was indistinguishable from that of the parent strain. The LptD and LptE mutants exhibited increased secretion of outer membrane vesicles containing LPS and resistance against various antibiotics, antimicrobial peptides, and host complement. In vivo cross-linking studies revealed that the conformation of the LptD-LptE complex was altered in the LptD and LptE mutants. Furthermore, several clinical isolates of E. coli carried amino acid substitutions of LptD and LptE that conferred resistance against antimicrobial substances. This study demonstrated an experimental evolution of bacterial virulence properties in an animal infection model and identified functional alterations of the growth-essential LPS transporter that led to high bacterial virulence by conferring resistance against antimicrobial substances. These findings suggest that non-pathogenic bacteria can gain virulence traits by changing the functions of essential genes, and provide new insight to bacterial evolution in a host environment.

Tumor mutation burden and immunological, genomic, and clinicopathological factors as biomarkers for checkpoint inhibitor treatment of patients with non-small-cell lung cancer.

Cancer treatment using immune checkpoint inhibitors is widely used, although biomarkers predictive of response are not well established. However, both the expressions of programmed cell death ligand 1 (PD-L1) and the tumor mutation burden (TMB) hold promise as such biomarkers for immune checkpoint inhibitors; however, its characteristics and clinical and immunological impacts have not been fully analyzed. We, therefore, evaluated the clinical and immunological parameters related to TMB to identify potential new biomarkers. We enrolled 92 patients with non-small-cell lung cancer who underwent surgery at Fukushima Medical University Hospital from 2013 to 2016. TMB of individual tumors was calculated by whole-exome sequencing analysis. Major cancer-related gene mutations were evaluated using panel sequencing. Expression of PD-L1 and abundance of tumor-infiltrating lymphocytes were evaluated by immunohistochemistry using surgical samples. The median TMB value was 60. TMB was significantly higher in men, current or former smokers, and in patients with squamous cell carcinoma, tumor size ≥ 2.8 cm, wild-type EGFR, TP53 gene mutation-positive status, and cyclin-dependent kinase-inhibitor gene 2A mutation-positive status. According to multivariate analysis, TMB was significantly associated with EGFR gene mutation-negative status (p = 0.0111) and TP53 gene mutation-positive status (p = 0.0425). If TMB is identified as a robust biomarker for immune checkpoint inhibitor administration, analysis of TP53 and EGFR mutations may provide a relatively rapid and easy proxy for predicting TMB.

Prognostic Impact of Tumor Mutation Burden in Patients With Completely Resected Non-Small Cell Lung Cancer: Brief Report.

INTRODUCTION: Tumor mutation burden (TMB) is thought to be associated with the amount of neoantigen in the tumor and to have an important role in predicting the effect of immune checkpoint inhibitors. However, the relevance of TMB to prognosis is not yet fully understood. In this study, we investigated the clinical significance of TMB in patients with NSCLC and examined the relationship between TMB and prognosis. METHODS: We calculated TMB within individual tumors by whole-exome sequencing analysis using next-generation sequencing. We included that there were 90 patients with NSCLC who underwent surgery in the Hospital of Fukushima Medical University from 2013 to 2016. No patients received chemotherapy or immunotherapy before surgery. We assessed the correlation between TMB and prognosis. RESULTS: TMB greater than 62 was associated with worse overall survival (OS) of patients with NSCLC (hazard ratio [HR] = 6.633, p = 0.0003). Multivariate analysis showed poor prognosis with high TMB (HR = 12.31, p = 0.019). In patients with stage I NSCLC, higher TMB was associated with worse prognosis for both OS (HR = 7.582, p = 0.0018) and disease-free survival (HR = 6.07, p = 0.0072). CONCLUSIONS: High TMB in NSCLC is a poor prognostic factor. If high TMB is a predictor of the efficacy of immune checkpoint inhibitors, postoperative adjuvant therapy with immune checkpoint inhibitors may contribute to improvement of recurrence and OS.

[Exploring the Possibility of New Biomarkers of Immune Checkpoint Inhibitors for Non-Small Cell Lung Cancer (NSCLC)].

Mutation burden in a tumor, presumably involving neo-antigens in the tumor tissue, is also thought to be one of the better predictors for the efficacy of immune checkpoint inhibitors. However, it is difficult to analyze the mutation burden routinely in the clinic. Here, we describe more convenient factors that can be used as surrogate markers of mutation burden. Ninety-four patients with NSCLC who underwent resection in our institution were recruited for this study. Mutation burden and major gene alterations were analyzed by using next generation sequencing. Several immunological parameters were also assessed using immunohistochemistry. Statistical analysis was performed on mutation burden, major gene alternations, immunohistochemistry, and clinical parameters. The median mutation load was 54 mutations(range, 10-363 mutations). Squamous cell carcinoma, EGFRmutation -negativity, and TP53 alteration-positivity were closely connected with higher mutation burden. Multiple regression analysis showed that mutation burden in the tumor could be associated with EGFRmutation and TP53 alteration status.

Identification of a long non-coding RNA as a novel biomarker and potential therapeutic target for metastatic prostate cancer.

Metastatic prostate cancer (PCa) is still an incurable disease. Long non-coding RNAs (lncRNAs) may be an overlooked source of cancer biomarkers and therapeutic targets. We therefore performed RNA sequencing on paired metastatic/non-metastatic PCa xenografts derived from clinical specimens. The most highly up-regulated transcript was LOC728606, a lncRNA now designated PCAT18. PCAT18 is specifically expressed in the prostate compared to 11 other normal tissues (p<0.05) and up-regulated in PCa compared to 15 other neoplasms (p<0.001). Cancer-specific up-regulation of PCAT18 was confirmed on an independent dataset of PCa and benign prostatic hyperplasia samples (p<0.001). PCAT18 was detectable in plasma samples and increased incrementally from healthy individuals to those with localized and metastatic PCa (p<0.01). We identified a PCAT18-associated expression signature (PES), which is highly PCa-specific and activated in metastatic vs. primary PCa samples (p<1E-4, odds ratio>2). The PES was significantly associated with androgen receptor (AR) signalling. Accordingly, AR activation dramatically up-regulated PCAT18 expression in vitro and in vivo. PCAT18 silencing significantly (p<0.001) inhibited PCa cell proliferation and triggered caspase 3/7 activation, with no effect on non-neoplastic cells. PCAT18 silencing also inhibited PCa cell migration (p<0.01) and invasion (p<0.01). These results position PCAT18 as a potential therapeutic target and biomarker for metastatic PCa.

Identification and characterization of novel genotoxic stress-inducible nuclear long noncoding RNAs in mammalian cells.

Whole transcriptome analyses have revealed a large number of novel transcripts including long and short noncoding RNAs (ncRNAs). Currently, there is great interest in characterizing the functions of the different classes of ncRNAs and their relevance to cellular processes. In particular, nuclear long ncRNAs may be involved in controlling various aspects of biological regulation, such as stress responses. By a combination of bioinformatic and experimental approaches, we identified 25 novel nuclear long ncRNAs from 6,088,565 full-length human cDNA sequences. Some nuclear long ncRNAs were conserved among vertebrates, whereas others were found only among primates. Expression profiling of the nuclear long ncRNAs in human tissues revealed that most were expressed ubiquitously. A subset of the identified nuclear long ncRNAs was induced by the genotoxic agents mitomycin C or doxorubicin, in HeLa Tet-off cells. There were no commonly altered nuclear long ncRNAs between mitomycin C- and doxorubicin-treated cells. These results suggest that distinct sets of nuclear long ncRNAs play roles in cellular defense mechanisms against specific genotoxic agents, and that particular long ncRNAs have the potential to be surrogate indicators of a specific cell stress.

Genome-wide determination of RNA stability reveals hundreds of short-lived noncoding transcripts in mammals.

Mammalian genomes produce huge numbers of noncoding RNAs (ncRNAs). However, the functions of most ncRNAs are unclear, and novel techniques that can distinguish functional ncRNAs are needed. Studies of mRNAs have revealed that the half-life of each mRNA is closely related to its physiological function, raising the possibility that the RNA stability of an ncRNA reflects its function. In this study, we first determined the half-lives of 11,052 mRNAs and 1418 ncRNAs in HeLa Tet-off (TO) cells by developing a novel genome-wide method, which we named 5'-bromo-uridine immunoprecipitation chase-deep sequencing analysis (BRIC-seq). This method involved pulse-labeling endogenous RNAs with 5'-bromo-uridine and measuring the ongoing decrease in RNA levels over time using multifaceted deep sequencing. By analyzing the relationship between RNA half-lives and functional categories, we found that RNAs with a long half-life (t(1/2) ≥ 4 h) contained a significant proportion of ncRNAs, as well as mRNAs involved in housekeeping functions, whereas RNAs with a short half-life (t(1/2) < 4 h) included known regulatory ncRNAs and regulatory mRNAs. The stabilities of a significant set of short-lived ncRNAs are regulated by external stimuli, such as retinoic acid treatment. In particular, we identified and characterized several novel long ncRNAs involved in cell proliferation from the group of short-lived ncRNAs. We designated this novel class of ncRNAs with a short half-life as Short-Lived noncoding Transcripts (SLiTs). We propose that the strategy of monitoring RNA half-life will provide a powerful tool for investigating hitherto functionally uncharacterized regulatory RNAs.

HGPD: Human Gene and Protein Database, 2012 update.

The Human Gene and Protein Database (HGPD; http://www.HGPD.jp/) is a unique database that stores information on a set of human Gateway entry clones in addition to protein expression and protein synthesis data. The HGPD was launched in November 2008, and 33,275 human Gateway entry clones have been constructed from the open reading frames (ORFs) of full-length cDNA, thus representing the largest collection in the world. Recently, research objectives have focused on the development of new medicines and the establishment of novel diagnostic methods and medical treatments. And, studies using proteins and protein information, which are closely related to gene function, have been undertaken. For this update, we constructed an additional 9974 human Gateway entry clones, giving a total of 43,249. This set of human Gateway entry clones was named the Human Proteome Expression Resource, known as the 'HuPEX'. In addition, we also classified the clones into 10 groups according to protein function. Moreover, in vivo cellular localization data of proteins for 32,651 human Gateway entry clones were included for retrieval from the HGPD. In 'Information Overview', which presents the search results, the ORF region of each cDNA is now displayed allowing the Gateway entry clones to be searched more easily.

Statistical analysis of features associated with protein expression/solubility in an in vivo Escherichia coli expression system and a wheat germ cell-free expression system.

Recombinant protein technology is an important tool in many industrial and pharmacological applications. Although the success rate of obtaining soluble proteins is relatively low, knowledge of protein expression/solubility under 'standard' conditions may increase the efficiency and reduce the cost of proteomics studies. In this study, we conducted a genome-scale experiment to assess the overexpression and the solubility of human full-length cDNA in an in vivo Escherichia coli expression system and a wheat germ cell-free expression system. We evaluated the influences of sequence and structural features on protein expression/solubility in each system and estimated a minimal set of features associated with them. A comparison of the feature sets related to protein expression/solubility in the in vivo Escherichia coli expression system revealed that the structural information was strongly associated with protein expression, rather than protein solubility. Moreover, a significant difference was found in the number of features associated with protein solubility in the two expression systems.

Comparative analysis of human SRC-family kinase substrate specificity in vitro.

Src family kinases (SFKs) are the earliest known family of tyrosine kinases and are widely thought to play essential roles in cellular signal transduction. Although numerous functional analyses have been performed, no study has analyzed the specificity of all SFKs on an equal platform. To gain a better understanding of SFK phosphorylation, we designed a high-throughput in vitro kinase assay on the subproteome scale using surface plasmon resonance. We reacted each of the 11 human SFKs with 519 substrate proteins, and significant phosphorylation was detected in 33.6% (1921) of the total 5709 kinase-substrate combinations. A large number of novel phosphorylations were included among them. Many substrates were shown to be phosphorylated by multiple SFKs, which might reflect functional complementarity of SFKs. Clustering analysis of phosphorylation results grouped substrates into 10 categories, while the similarity of SFK catalytic specificity exhibited no significant correlation with that of amino acid sequences. In silico predictions of SRC-specific phosphorylation sites were not consistent with experimental results, implying some unknown SRC recognition modes. In an attempt to find biologically meaningful novel substrates, phosphorylation data were integrated with annotation data. The extensive in vitro data obtained in this study would provide valuable clues for further understanding SFK-mediated signal transduction.

Alternative splicing of genes during neuronal differentiation of NT2 pluripotential human embryonal carcinoma cells.

We analyzed the mRNA diversity of genes after inducing neuronal differentiation in human NT2 teratocarcinoma cells using all-trans retinoic acid (RA). DNA microarray analyses of cells treated with RA identified 358 RA-responsive genes. mRNA diversity analysis revealed that 274 genes produced multiple protein-coding transcripts by alternative splicing. Among these 274 genes, we chose 26 genes that showed AS in their C-terminus and 12 transcription factor genes for further analysis. By using transcript-specific primers, we performed quantitative real-time PCR analysis to examine the expression profiles of all the protein-coding transcripts. Consequently, we identified genes which showed different RA-induced changes in the expression of their protein-coding transcripts.

Identification and functional analyses of 11,769 full-length human cDNAs focused on alternative splicing.

We analyzed diversity of mRNA produced as a result of alternative splicing in order to evaluate gene function. First, we predicted the number of human genes transcribed into protein-coding mRNAs by using the sequence information of full-length cDNAs and 5'-ESTs and obtained 23 241 of such human genes. Next, using these genes, we analyzed the mRNA diversity and consequently sequenced and identified 11 769 human full-length cDNAs whose predicted open reading frames were different from other known full-length cDNAs. Especially, 30% of the cDNAs we identified contained variation in the transcription start site (TSS). Our analysis, which particularly focused on multiple variable first exons (FEVs) formed due to the alternative utilization of TSSs, led to the identification of 261 FEVs expressed in the tissue-specific manner. Quantification of the expression profiles of 13 genes by real-time PCR analysis further confirmed the tissue-specific expression of FEVs, e.g. OXR1 had specific TSS in brain and tumor tissues, and so on. Finally, based on the results of our mRNA diversity analysis, we have created the FLJ Human cDNA Database. From our result, it has been understood mechanisms that one gene produces suitable protein-coding transcripts responding to the situation and the environment.

Human Gene and Protein Database (HGPD): a novel database presenting a large quantity of experiment-based results in human proteomics.

Completion of human genome sequencing has greatly accelerated functional genomic research. Full-length cDNA clones are essential experimental tools for functional analysis of human genes. In one of the projects of the New Energy and Industrial Technology Development Organization (NEDO) in Japan, the full-length human cDNA sequencing project (FLJ project), nucleotide sequences of approximately 30 000 human cDNA clones have been analyzed. The Gateway system is a versatile framework to construct a variety of expression clones for various experiments. We have constructed 33 275 human Gateway entry clones from full-length cDNAs, representing to our knowledge the largest collection in the world. Utilizing these clones with a highly efficient cell-free protein synthesis system based on wheat germ extract, we have systematically and comprehensively produced and analyzed human proteins in vitro. Sequence information for both amino acids and nucleotides of open reading frames of cDNAs cloned into Gateway entry clones and in vitro expression data using those clones can be retrieved from the Human Gene and Protein Database (HGPD, http://www.HGPD.jp). HGPD is a unique database that stores the information of a set of human Gateway entry clones and protein expression data and helps the user to search the Gateway entry clones.

Differential responses of normal human coronary artery endothelial cells against multiple cytokines comparatively assessed by gene expression profiles.

Endothelial cells play an important role in terms of biological functions by responding to a variety of stimuli in the blood. However, little is known about the molecular mechanism involved in rendering the variety in the cellular response. To investigate the variety of the cellular responses against exogenous stimuli at the gene expression level, we attempted to describe the cellular responses with comprehensive gene expression profiles, dissect them into multiple response patterns, and characterize the response patterns according to the information accumulated so far on the genes included in the patterns. We comparatively analyzed in parallel the gene expression profiles obtained with DNA microarrays from normal human coronary artery endothelial cells (HCAECs) stimulated with multiple cytokines, interleukin-1beta, tumor necrosis factor-alpha, interferon-beta, interferon-gamma, and oncostatin M, which are profoundly involved in various functional responses of endothelial cells. These analyses revealed that the cellular responses of HCAECs against these cytokines included at least 15 response patterns specific to a single cytokine or common to multiple cytokines. Moreover, we statistically extracted genes contained within the individual response patterns and characterized the response patterns with the genes referring to the previously accumulated findings including the biological process defined by the Gene Ontology Consortium (GO). Out of the 15 response patterns in which at least one gene was successfully extracted through the statistical approach, 11 response patterns were differentially characterized by representing the number of genes contained in individual criteria of the biological process in the GO only. The approach to dissect cellular responses into response patterns and to characterize the pattern at the gene expression level may contribute to the gaining of insight for untangling the diversity of cellular functions.

Large-scale identification and characterization of alternative splicing variants of human gene transcripts using 56,419 completely sequenced and manually annotated full-length cDNAs.

We report the first genome-wide identification and characterization of alternative splicing in human gene transcripts based on analysis of the full-length cDNAs. Applying both manual and computational analyses for 56,419 completely sequenced and precisely annotated full-length cDNAs selected for the H-Invitational human transcriptome annotation meetings, we identified 6877 alternative splicing genes with 18 297 different alternative splicing variants. A total of 37,670 exons were involved in these alternative splicing events. The encoded protein sequences were affected in 6005 of the 6877 genes. Notably, alternative splicing affected protein motifs in 3015 genes, subcellular localizations in 2982 genes and transmembrane domains in 1348 genes. We also identified interesting patterns of alternative splicing, in which two distinct genes seemed to be bridged, nested or having overlapping protein coding sequences (CDSs) of different reading frames (multiple CDS). In these cases, completely unrelated proteins are encoded by a single locus. Genome-wide annotations of alternative splicing, relying on full-length cDNAs, should lay firm groundwork for exploring in detail the diversification of protein function, which is mediated by the fast expanding universe of alternative splicing variants.