Jun Sese

A large-scale full-length cDNA analysis to explore the budding yeast transcriptome

We performed a large-scale cDNA analysis to explore the transcriptome of the budding yeast Saccharomyces cerevisiae. We sequenced two cDNA libraries, one from the cells exponentially growing in a minimal medium and the other from meiotic cells. Both libraries were generated by using a vector-capping method that allows the accurate mapping of transcription start sites (TSSs). Consequently, we identified 11,575 TSSs associated with 3,638 annotated genomic features, including 3,599 ORFs, to suggest that most yeast genes have two or more TSSs. In addition, we identified 45 previously undescribed introns, including those affecting current ORF annotations and those spliced alternatively. Furthermore, the analysis revealed 667 transcription units in the intergenic regions and transcripts derived from antisense strands of 367 known features. We also found that 348 ORFs carry TSSs in their 3′-halves to generate sense transcripts starting from inside the ORFs. These results indicate that the budding yeast transcriptome is considerably more complex than previously thought, and it shares many recently revealed characteristics with the transcriptomes of mammals and other higher eukaryotes. Thus, the genome-wide active transcription that generates novel classes of transcripts appears to be an intrinsic feature of the eukaryotic cells. The budding yeast will serve as a versatile model for the studies on these aspects of transcriptome, and the full-length cDNA clones can function as an invaluable resource in such studies.

Transversing itemset lattices with statistical metric pruning

We study how to efficiently compute significant association rules according to common statistical measures such as a chi-squared value or correlation coefficient. For this purpose, one might consider to use of the Apriori algorithm, but the algorithm needs major conversion, because none of these statistical metrics are anti-monotone, and the use of higher support for reducing the search space cannot guarantee solutions in its the search space. We here present a method of estimating a tight upper bound on the statistical metric associated with any superset of an itemset, as well as the novel use of the resulting information of upper bounds to prune unproductive supersets while traversing itemset lattices. Experimental tests demonstrate the efficiency of this method.

Semi-supervised local Fisher discriminant analysis for dimensionality reduction

When only a small number of labeled samples are available, supervised dimensionality reduction methods tend to perform poorly because of overfitting. In such cases, unlabeled samples could be useful in improving the performance. In this paper, we propose a semi-supervised dimensionality reduction method which preserves the global structure of unlabeled samples in addition to separating labeled samples in different classes from each other. The proposed method, which we call SEmi-supervised Local Fisher discriminant analysis (SELF), has an analytic form of the globally optimal solution and it can be computed based on eigen-decomposition. We show the usefulness of SELF through experiments with benchmark and real-world document classification datasets.

Diverse transcriptional initiation revealed by fine, large‐scale mapping of mRNA start sites

Determination of the mRNA start site is the first step in identifying the promoter region, which is of key importance for transcriptional regulation of gene expression. The ‘oligo‐capping’ method enabled us to introduce a sequence tag to the first base of an mRNA by replacing the cap structure of the mRNA. Using cDNA libraries made from oligo‐capped mRNAs, we could identify the transcriptional start site of an individual mRNA just by sequencing the 5′‐end of the cDNA. The fine mapping of transcriptional start sites was performed for 5880 mRNAs in 276 human genes. Contrary to our expectations, the majority of the genes showed a diverse distribution of transcriptional start sites. They were distributed over 61.7 bp with a standard deviation of 19.5. Our finding may reflect the dynamic nature of transcriptional initiation events of human genes in vivo.

5′-end SAGE for the analysis of transcriptional start sites

Identification of the mRNA start site is essential in establishing the full-length cDNA sequence of a gene and analyzing its promoter region, which regulates gene expression. Here we describe the development of a 5′-end serial analysis of gene expression (5′ SAGE) that can be used to globally identify transcriptional start sites and the frequency of individual mRNAs. Of the 25,684 5′ SAGE tags in the HEK293 human cell library, 19,893 matched to the human genome. Among 15,448 tags in one locus of the genome, 85.8%–96.1% of the 5′ SAGE tags were assigned within −500 to +200 nt of mRNA start sites using the RefSeq, UniGene and DBTSS databases. This technique should facilitate 5′-end transcriptome analysis in a variety of cells and tissues.

KDD Cup 2001 report

This paper presents results and lessons from KDD Cup 2001. KDD Cup 2001 focused on mining biological databases. It involved three cutting-edge tasks related to drug design and genomics.

Mutual information estimation reveals global associations between stimuli and biological processes

BackgroundAlthough microarray gene expression analysis has become popular, it remains difficult to interpret the biological changes caused by stimuli or variation of conditions. Clustering of genes and associating each group with biological functions are often used methods. However, such methods only detect partial changes within cell processes. Herein, we propose a method for discovering global changes within a cell by associating observed conditions of gene expression with gene functions.ResultsTo elucidate the association, we introduce a novel feature selection method called Least-Squares Mutual Information (LSMI), which computes mutual information without density estimaion, and therefore LSMI can detect nonlinear associations within a cell. We demonstrate the effectiveness of LSMI through comparison with existing methods. The results of the application to yeast microarray datasets reveal that non-natural stimuli affect various biological processes, whereas others are no significant relation to specific cell processes. Furthermore, we discover that biological processes can be categorized into four types according to the responses of various stimuli: DNA/RNA metabolism, gene expression, protein metabolism, and protein localization.ConclusionWe proposed a novel feature selection method called LSMI, and applied LSMI to mining the association between conditions of yeast and biological processes through microarray datasets. In fact, LSMI allows us to elucidate the global organization of cellular process control.

Semi-supervised local fisher discriminant analysis for dimensionality reduction

When only a small number of labeled samples are available, supervised dimensionality reduction methods tend to perform poorly due to overfitting. In such cases, unlabeled samples could be useful in improving the performance. In this paper, we propose a semi-supervised dimensionality reduction method which preserves the global structure of unlabeled samples in addition to separating labeled samples in different classes from each other. The proposed method has an analytic form of the globally optimal solution and it can be computed based on eigendecompositions. Therefore, the proposed method is computationally reliable and efficient. We show the effectiveness of the proposed method through extensive simulations with benchmark data sets.

Sequencing of the genus Arabidopsis identifies a complex history of nonbifurcating speciation and abundant trans-specific polymorphism

The notion of species as reproductively isolated units related through a bifurcating tree implies that gene trees should generally agree with the species tree and that sister taxa should not share polymorphisms unless they diverged recently and should be equally closely related to outgroups. It is now possible to evaluate this model systematically. We sequenced multiple individuals from 27 described taxa representing the entire Arabidopsis genus. Cluster analysis identified seven groups, corresponding to described species that capture the structure of the genus. However, at the level of gene trees, only the separation of Arabidopsis thaliana from the remaining species was universally supported, and, overall, the amount of shared polymorphism demonstrated that reproductive isolation was considerably more recent than the estimated divergence times. We uncovered multiple cases of past gene flow that contradict a bifurcating species tree. Finally, we showed that the pattern of divergence differs between gene ontologies, suggesting a role for selection.

Genome-wide quantification of homeolog expression ratio revealed nonstochastic gene regulation in synthetic allopolyploid Arabidopsis

Genome duplication with hybridization, or allopolyploidization, occurs commonly in plants, and is considered to be a strong force for generating new species. However, genome-wide quantification of homeolog expression ratios was technically hindered because of the high homology between homeologous gene pairs. To quantify the homeolog expression ratio using RNA-seq obtained from polyploids, a new method named HomeoRoq was developed, in which the genomic origin of sequencing reads was estimated using mismatches between the read and each parental genome. To verify this method, we first assembled the two diploid parental genomes of Arabidopsis halleri subsp. gemmifera and Arabidopsis lyrata subsp. petraea (Arabidopsis petraea subsp. umbrosa), then generated a synthetic allotetraploid, mimicking the natural allopolyploid Arabidopsis kamchatica. The quantified ratios corresponded well to those obtained by Pyrosequencing. We found that the ratios of homeologs before and after cold stress treatment were highly correlated (r = 0.870). This highlights the presence of nonstochastic polyploid gene regulation despite previous research identifying stochastic variation in expression. Moreover, our new statistical test incorporating overdispersion identified 226 homeologs (1.11% of 20 369 expressed homeologs) with significant ratio changes, many of which were related to stress responses. HomeoRoq would contribute to the study of the genes responsible for polyploid-specific environmental responses.