Guoli Ji

Genome-wide landscape of polyadenylation in Arabidopsis provides evidence for extensive alternative polyadenylation

Alternative polyadenylation (APA) has been shown to play an important role in gene expression regulation in animals and plants. However, the extent of sense and antisense APA at the genome level is not known. We developed a deep-sequencing protocol that queries the junctions of 3′UTR and poly(A) tails and confidently maps the poly(A) tags to the annotated genome. The results of this mapping show that 70% of Arabidopsis genes use more than one poly(A) site, excluding microheterogeneity. Analysis of the poly(A) tags reveal extensive APA in introns and coding sequences, results of which can significantly alter transcript sequences and their encoding proteins. Although the interplay of intron splicing and polyadenylation potentially defines poly(A) site uses in introns, the polyadenylation signals leading to the use of CDS protein-coding region poly(A) sites are distinct from the rest of the genome. Interestingly, a large number of poly(A) sites correspond to putative antisense transcripts that overlap with the promoter of the associated sense transcript, a mode previously demonstrated to regulate sense gene expression. Our results suggest that APA plays a far greater role in gene expression in plants than previously expected.

Genome level analysis of rice mRNA 3′-end processing signals and alternative polyadenylation

The position of a poly(A) site of eukaryotic mRNA is determined by sequence signals in pre-mRNA and a group of polyadenylation factors. To reveal rice poly(A) signals at a genome level, we constructed a dataset of 55 742 authenticated poly(A) sites and characterized the poly(A) signals. This resulted in identifying the typical tripartite cis-elements, including FUE, NUE and CE, as previously observed in Arabidopsis. The average size of the 3′-UTR was 289 nucleotides. When mapped to the genome, however, 15% of these poly(A) sites were found to be located in the currently annotated intergenic regions. Moreover, an extensive alternative polyadenylation profile was evident where 50% of the genes analyzed had more than one unique poly(A) site (excluding microheterogeneity sites), and 13% had four or more poly(A) sites. About 4% of the analyzed genes possessed alternative poly(A) sites at their introns, 5′-UTRs, or protein coding regions. The authenticity of these alternative poly(A) sites was partially confirmed using MPSS data. Analysis of nucleotide profile and signal patterns indicated that there may be a different set of poly(A) signals for those poly(A) sites found in the coding regions. Based on the features of rice poly(A) signals, an updated algorithm termed PASS-Rice was designed to predict poly(A) sites.

Using partial least squares and support vector machines for bankruptcy prediction

The evaluation of corporate financial distress has attracted significant global attention as a result of the increasing number of worldwide corporate failures. There is an immediate and compelling need for more effective financial distress prediction models. This paper presents a novel method to predict bankruptcy. The proposed method combines the partial least squares (PLS) based feature selection with support vector machine (SVM) for information fusion. PLS can successfully identify the complex nonlinearity and correlations among the financial indicators. The experimental results demonstrate its superior predictive ability. On the one hand, the proposed model can select the most relevant financial indicators to predict bankruptcy and at the same time identify the role of each variable in the prediction process. On the other hand, the proposed models high levels of prediction accuracy can translate into benefits to financial organizations through such activities as credit approval, and loan portfolio and security management.

Genome-Wide Control of Polyadenylation Site Choice by CPSF30 in Arabidopsis

This work shows that poly(A) site choice is affected in 45% or more of all genes in an Arabidopsis thaliana mutant that lacks a core polyadenylation factor subunit and that a novel poly(A) signal exists that can function in the absence of the affected protein. These results provide new insight into mechanisms of alternative polyadenylation in plants. The Arabidopsis thaliana ortholog of the 30-kD subunit of the mammalian Cleavage and Polyadenylation Specificity Factor (CPSF30) has been implicated in the responses of plants to oxidative stress, suggesting a role for alternative polyadenylation. To better understand this, poly(A) site choice was studied in a mutant (oxt6) deficient in CPSF30 expression using a genome-scale approach. The results indicate that poly(A) site choice in a large majority of Arabidopsis genes is altered in the oxt6 mutant. A number of poly(A) sites were identified that are seen only in the wild type or oxt6 mutant. Interestingly, putative polyadenylation signals associated with sites that are seen only in the oxt6 mutant are decidedly different from the canonical plant polyadenylation signal, lacking the characteristic A-rich near-upstream element (where AAUAAA can be found); this suggests that CPSF30 functions in the handling of the near-upstream element. The sets of genes that possess sites seen only in the wild type or mutant were enriched for those involved in stress and defense responses, a result consistent with the properties of the oxt6 mutant. Taken together, these studies provide new insights into the mechanisms and consequences of CPSF30-mediated alternative polyadenylation.

Determinants of public T cell responses

Historically, sharing T cell receptors (TCRs) between individuals has been speculated to be impossible, considering the dramatic discrepancy between the potential enormity of the TCR repertoire and the limited number of T cells generated in each individual. However, public T cell response, in which multiple individuals share identical TCRs in responding to a same antigenic epitope, has been extensively observed in a variety of immune responses across many species. Public T cell responses enable individuals within a population to generate similar antigen-specific TCRs against certain ubiquitous pathogens, leading to favorable biological outcomes. However, the relatively concentrated feature of TCR repertoire may limit T cell response in a population to some other pathogens. It could be a great benefit for human health if public T cell responses can be manipulated. Therefore, the mechanistic insight of public TCR generation is important to know. Recently, high-throughput DNA sequencing has revolutionized the study of immune receptor repertoires, which allows a much better understanding of the factors that determine the overlap of TCR repertoire among individuals. Here, we summarize the current knowledge on public T-cell response and discuss future challenges in this field.

Recombinatorial Biases and Convergent Recombination Determine Interindividual TCRβ Sharing in Murine Thymocytes

Overlap of TCR repertoires among individuals provides the molecular basis for public T cell responses. By deep-sequencing the TCRβ repertoires of CD4+CD8+ thymocytes from three individual mice, we observed that a substantial degree of TCRβ overlap, comprising ∼10–15% of all unique amino acid sequences and ∼5–10% of all unique nucleotide sequences across any two individuals, is already present at this early stage of T cell development. The majority of TCRβ sharing between individual thymocyte repertoires could be attributed to the process of convergent recombination, with additional contributions likely arising from recombinatorial biases; the role of selection during intrathymic development was negligible. These results indicate that the process of TCR gene recombination is the major determinant of clonotype sharing between individuals.

PLS-Based Gene Selection and Identification of Tumor-Specific Genes

In view of the characteristics of high-dimensional small sample, strong relevance, and high noise of the identification of tumor-specific genes on microarray, a novel partial least squares (PLS) based gene-selection method, which synthesizes genetic relatedness and is suitable for multicategory classification, is presented. Using the explanation difference of independent variables on dependent variable (class), we define three indicators for global gene selection, which takes into accounts the combined effects of all the genes and the correlation among the genes. Integrated with the linear kernel support vector classifier (SVC), the proposed method is tested by MIT acute myeloid leukemia/acute lymphoblastic leukemia (AML/ALL) and small round blue cell tumors (SRBCT) data sets. A subset of specific genes with small numbers and high identification are obtained. The results indicate that our proposed PLS-based method for tumor-specific genes selection is highly efficient. Compared to the literature, the selected specific genes from both two-category dataset AML/ALL and multicategory dataset SRBCT are credible. Further investigation shows that the proposed gene-selection method is robust. Overall, the proposed method can effectively solve feature-selection problem on high-dimensional small sample. At the same time, it has good performance for multicategory classification as well.

A classification-based prediction model of messenger RNA polyadenylation sites

Messenger RNA polyadenylation is one of the essential processing steps during eukaryotic gene expression. The site of polyadenylation [(poly(A) site] marks the end of a transcript, which is also the end of a gene. A computation program that is able to recognize poly(A) sites would not only prove useful for genome annotation in finding genes ends, but also for predicting alternative poly(A) sites. Features that define the poly(A) sites can now be extracted from the poly(A) site datasets to build such predictive models. Using methods, including K-gram pattern, Z-curve, position-specific scoring matrix and first-order inhomogeneous Markov sub-model, numerous features were generated and placed in an original feature space. To select the most useful features, attribute selection algorithms, such as information gain and entropy, were employed. A training model was then built based on the Bayesian network to determine a subset of the optimal features. Test models corresponding to the training models were built to predict poly(A) sites in Arabidopsis and rice. Thus, a prediction model, termed Poly(A) site classifier, or PAC, was constructed. The uniqueness of the model lies in its structure in that each sub-model can be replaced or expanded, while feature generation, selection and classification are all independent processes. Its modular design makes it easily adaptable to different species or datasets. The algorithms high specificity and sensitivity were demonstrated by testing several datasets and, at the best combinations, they both reached 95%. The software package may be used for genome annotation and optimizing transgene structure.

Expectation–Maximization Approach to Fault Diagnosis With Missing Data

This paper introduces a data-driven approach for fault diagnosis in the presence of incomplete monitor data. The expectation-maximization (EM) algorithm is applied to handle missing data in order to obtain a maximum-likelihood solution for the discrete (or categorical) distribution. Because of the nature of categorical distributions, the maximization step of the EM algorithm is shown in this paper to have an easily calculated analytical solution, making this method computationally simple. An experimental study on a ball-and-tube system is investigated to demonstrate advantages of the proposed approach.

Gene expression responses of threespine stickleback to salinity: implications for salt-sensitive hypertension

Despite recent success with genome-wide association studies (GWAS), identifying hypertension (HTN)-susceptibility loci in the general population remains difficult. Here, we present a novel strategy to address this challenge by studying salinity adaptation in the threespine stickleback, a fish species with diverse salt-handling ecotypes. We acclimated native freshwater (FW) and anadromous saltwater (SW) threespine sticklebacks to fresh, brackish, and sea water for 30 days, and applied RNA sequencing to determine the gene expression in fish kidneys. We identified 1844 salt-responsive genes that were differentially expressed between FW sticklebacks acclimated to different salinities and/or between SW and FW sticklebacks acclimated to full-strength sea water. Significant overlap between stickleback salt-responsive genes and human genes implicated in HTN was detected (P < 10−7, hypergeometric test), suggesting a possible similarity in genetic mechanisms of salt handling between threespine sticklebacks and humans. The overlapping genes included a newly discovered HTN gene—MAP3K15, whose expression in FW stickleback kidneys decreases with salinity. These also included genes located in the GWAS loci such as AGTRAP-PLOD1 and CYP1A1-ULK3, which contain multiple potentially causative genes contributing to HTN susceptibility that need to be prioritized for study. Taken together, we show that stickleback salt-responsive genes provide valuable information facilitating the identification of human HTN genes. Thus, threespine sticklebacks may be used as a model, complementary to existing animal models, in human HTN research.