Mitsunori Kayano

Functional Cluster Analysis via Orthonormalized Gaussian Basis Expansions and Its Application

We propose functional cluster analysis (FCA) for multidimensional functional data sets, utilizing orthonormalized Gaussian basis functions. An essential point in FCA is the use of orthonormal bases that yield the identity matrix for the integral of the product of any two bases. We construct orthonormalized Gaussian basis functions using Cholesky decomposition and derive a property of Cholesky decomposition with respect to Gram-Schmidt orthonormalization. The advantages of the functional clustering are that it can be applied to the data observed at different time points for each subject, and the functional structure behind the data can be captured by removing the measurement errors. Numerical experiments are conducted to investigate the effectiveness of the proposed method, as compared to conventional discrete cluster analysis. The proposed method is applied to three-dimensional (3D) protein structural data that determine the 3D arrangement of amino acids in individual protein.

Efficiently finding genome-wide three-way gene interactions from transcript-and genotype-data

Motivation: We address the issue of finding a three-way gene interaction, i.e. two interacting genes in expression under the genotypes of another gene, given a dataset in which expressions and genotypes are measured at once for each individual. This issue can be a general, switching mechanism in expression of two genes, being controlled by categories of another gene, and finding this type of interaction can be a key to elucidating complex biological systems. The most suitable method for this issue is likelihood ratio test using logistic regressions, which we call interaction test, but a serious problem of this test is computational intractability at a genome-wide level. Results: We developed a fast method for this issue which improves the speed of interaction test by around 10 times for any size of datasets, keeping highly interacting genes with an accuracy of ∼85%. We applied our method to ∼3 × 108 three-way combinations generated from a dataset on human brain samples and detected three-way gene interactions with small P-values. To check the reliability of our results, we first conducted permutations by which we can show that the obtained P-values are significantly smaller than those obtained from permuted null examples. We then used GEO (Gene Expression Omnibus) to generate gene expression datasets with binary classes to confirm the detected three-way interactions by using these datasets and interaction tests. The result showed us some datasets with significantly small P-values, strongly supporting the reliability of the detected three-way interactions. Availability: Software is available from http://www.bic.kyoto-u.ac.jp/pathway/kayano/bioinfo_three-way.html Contact: kayano@kuicr.kyoto-u.ac.jp Supplementary information: Supplementary data are available at Bioinformatics online.

Plasma microRNA biomarker detection for mild cognitive impairment using differential correlation analysis

BackgroundMild cognitive impairment (MCI) is an intermediate state between normal aging and dementia including Alzheimer’s disease. Early detection of dementia, and MCI, is a crucial issue in terms of secondary prevention. Blood biomarker detection is a possible way for early detection of MCI. Although disease biomarkers are detected by, in general, using single molecular analysis such as t-test, another possible approach is based on interaction between molecules.ResultsDifferential correlation analysis, which detects difference on correlation of two variables in case/control study, was carried out to plasma microRNA (miRNA) expression profiles of 30 age- and race-matched controls and 23 Japanese MCI patients. The 20 pairs of miRNAs, which consist of 20 miRNAs, were selected as MCI markers. Two pairs of miRNAs (hsa-miR-191 and hsa-miR-101, and hsa-miR-103 and hsa-miR-222) out of 20 attained the highest area under the curve (AUC) value of 0.962 for MCI detection. Other two miRNA pairs that include hsa-miR-191 and hsa-miR-125b also attained high AUC value of ≥ 0.95. Pathway analysis was performed to the MCI markers for further understanding of biological implications. As a result, collapsed correlation on hsa-miR-191 and emerged correlation on hsa-miR-125b might have key role in MCI and dementia progression.ConclusionDifferential correlation analysis, a bioinformatics tool to elucidate complicated and interdependent biological systems behind diseases, detects effective MCI markers that cannot be found by single molecule analysis such as t-test.

ROS-DET: robust detector of switching mechanisms in gene expression.

A switching mechanism in gene expression, where two genes are positively correlated in one condition and negatively correlated in the other condition, is a key to elucidating complex biological systems. There already exist methods for detecting switching mechanisms from microarrays. However, current approaches have problems under three real cases: outliers, expression values with a very small range and a small number of examples. ROS-DET overcomes these three problems, keeping the computational complexity of current approaches. We demonstrated that ROS-DET outperformed existing methods, under that all these three situations are considered. Furthermore, for each of the top 10 pairs ranked by ROS-DET, we attempted to identify a pathway, i.e. consecutive biological phenomena, being related with the corresponding two genes by checking the biological literature. In 8 out of the 10 pairs, we found two parallel pathways, one of the two genes being in each of the two pathways and two pathways coming to (or starting with) the same gene. This indicates that two parallel pathways would be cooperatively used under one experimental condition, corresponding to the positive correlation, and the two pathways might be alternatively used under the other condition, corresponding to the negative correlation. ROS-DET is available from http://www.bic.kyoto-u.ac.jp/pathway/kayano/ros-det.htm.

Detecting differentially coexpressed genes from labeled expression data: a brief review

We review methods for capturing differential coexpression, which can be divided into two cases by the size of gene sets: 1) two paired genes and 2) multiple genes. In the first case, two genes are positively and negatively correlated with each other under one and the other conditions, respectively. In the second case, multiple genes are coexpressed and randomly expressed under one and the other conditions, respectively. We summarize a variety of methods for the first and second cases into four and three approaches, respectively. We describe each of these approaches in detail technically, being followed by thorough comparative experiments with both synthetic and real data sets. Our experimental results imply high possibility of improving the efficiency of the current methods, particularly in the case of multiple genes, because of low performance achieved by the best methods which are relatively simple intuitive ones.

Construction of a virtual Mycobacterium tuberculosis consensus genome and its application to data from a next generation sequencer

BackgroundAlthough Mycobacterium tuberculosis isolates are consisted of several different lineages and the epidemiology analyses are usually assessed relative to a particular reference genome, M. tuberculosis H37Rv, which might introduce some biased results. Those analyses are essentially based genome sequence information of M. tuberculosis and could be performed in sillico in theory, with whole genome sequence (WGS) data available in the databases and obtained by next generation sequencers (NGSs). As an approach to establish higher resolution methods for such analyses, whole genome sequences of the M. tuberculosis complexes (MTBCs) strains available on databases were aligned to construct virtual reference genome sequences called the consensus sequence (CS), and evaluated its feasibility in in sillico epidemiological analyses.ResultsThe consensus sequence (CS) was successfully constructed and utilized to perform phylogenetic analysis, evaluation of read mapping efficacy, which is crucial for detecting single nucleotide polymorphisms (SNPs), and various MTBC typing methods virtually including spoligotyping, VNTR, Long sequence polymorphism and Beijing typing. SNPs detected based on CS, in comparison with H37Rv, were utilized in concatemer-based phylogenetic analysis to determine their reliability relative to a phylogenetic tree based on whole genome alignment as the gold standard. Statistical comparison of phylogenic trees based on CS with that of H37Rv indicated the former showed always better results that that of later. SNP detection and concatenation with CS was advantageous because the frequency of crucial SNPs distinguishing among strain lineages was higher than those of H37Rv. The number of SNPs detected was lower with the consensus than with the H37Rv sequence, resulting in a significant reduction in computational time. Performance of each virtual typing was satisfactory and accorded with those published when those are available.ConclusionsThese results indicated that virtual CS constructed from genome sequence data is an ideal approach as a reference for MTBC studies.

Multi-omics Approach for Estimating Metabolic Networks Using Low-Order Partial Correlations

Two typical purposes of metabolome analysis are to estimate metabolic pathways and to understand the regulatory systems underlying the metabolism. A powerful source of information for these analyses is a set of multi-omics data for RNA, proteins, and metabolites. However, integrated methods that analyze multi-omics data simultaneously and unravel the systems behind metabolisms have not been well established. We developed a statistical method based on low-order partial correlations with a robust correlation coefficient for estimating metabolic networks from metabolome, proteome, and transcriptome data. Our method is defined by the maximum of low-order, particularly first-order, partial correlations (MF-PCor) in order to assign a correct edge with the highest correlation and to detect the factors that strongly affect the correlation coefficient. First, through numerical experiments with real and synthetic data, we showed that the use of protein and transcript data of enzymes improved the accuracy of the estimated metabolic networks in MF-PCor. In these experiments, the effectiveness of the proposed method was also demonstrated by comparison with a correlation network (Cor) and a Gaussian graphical model (GGM). Our theoretical investigation confirmed that the performance of MF-PCor could be superior to that of the competing methods. In addition, in the real data analysis, we investigated the role of metabolites, enzymes, and enzyme genes that were identified as important factors in the network established by MF-PCor. We then found that some of them corresponded to specific reactions between metabolites mediated by catalytic enzymes that were difficult to be identified by analysis based on metabolite data alone.

Gene set differential analysis of time course expression profiles via sparse estimation in functional logistic model with application to time-dependent biomarker detection

High-throughput time course expression profiles have been available in the last decade due to developments in measurement techniques and devices. Functional data analysis, which treats smoothed curves instead of originally observed discrete data, is effective for the time course expression profiles in terms of dimension reduction, robustness, and applicability to data measured at small and irregularly spaced time points. However, the statistical method of differential analysis for time course expression profiles has not been well established. We propose a functional logistic model based on elastic net regularization (F-Logistic) in order to identify the genes with dynamic alterations in case/control study. We employ a mixed model as a smoothing method to obtain functional data; then F-Logistic is applied to time course profiles measured at small and irregularly spaced time points. We evaluate the performance of F-Logistic in comparison with another functional data approach, i.e. functional ANOVA test (F-ANOVA), by applying the methods to real and synthetic time course data sets. The real data sets consist of the time course gene expression profiles for long-term effects of recombinant interferon β on disease progression in multiple sclerosis. F-Logistic distinguishes dynamic alterations, which cannot be found by competitive approaches such as F-ANOVA, in case/control study based on time course expression profiles. F-Logistic is effective for time-dependent biomarker detection, diagnosis, and therapy.

Experimental investigation of bone mineral density in Thoroughbreds using quantitative computed tomography

ABSTRACT Bone mineral density (BMD) is one of the indications of the strength and health. BMD measured by quantitative computed tomography (QCT) was compared with that measured by dual energy X-ray absorptiometry (DXA) and radiographic bone aluminum equivalence (RBAE). Limbs were removed from horses that had been euthanized for reasons not associated with this study. Sixteen limbs (left and right metacarpals and metatarsals) from 4 horses were used to compare BMD as measured by QCT with those measured by DXA and RBAE. There was a strong correlation between BMD values measured by QCT and those measured by DXA (R2=0.85); correlation was also observed between values obtained by QCT and those obtained by RBAE (R2=0.61). To investigate changes in BMD with age, 37 right metacarpal bones, including 7 from horses euthanized because of fracture were examined by QCT. The BMD value of samples from horses dramatically increased until 2 years of age and then plateaued, a pattern similar to the growth curve. The BMD values of bone samples from horses euthanized because of fracture were within the population range, and samples of morbid fracture were not included. The relationship between BMD and age provides a reference for further quantitative studies of bone development and remodeling. Quantitative measurement of BMD using QCT may have great potential for the evaluation of bone biology for breeding and rearing management.

Short communication: Development of the first follicular wave dominant follicle on the ovary ipsilateral to the corpus luteum is associated with decreased conception rate in dairy cattle

In this study, we examined the effect of the locations of the first-wave dominant follicle (DF) and corpus luteum (CL) on fertility. In total, 350 artificial insemination (AI) procedures were conducted (lactating dairy cows: n=238, dairy heifers: n=112). Ovulation was confirmed 24 h after AI. The locations of the first-wave DF and CL were examined 5 to 9d after AI using rectal palpation or transrectal ultrasonography. Lactating dairy cows and dairy heifers were divided into 2 groups: (1) the ipsilateral group (IG), in which the DF was ipsilateral to the CL; and (2) the contralateral group (CG), in which the DF was contralateral to the CL. Pregnancy was diagnosed using transrectal ultrasonography 40d after AI. Conception rates were 54.0% in all cattle: 48.9% in lactating dairy cows, and 58.9% in dairy heifers. The incidence of the first-wave DF location did not differ between IG and CG (all cattle: 184 vs. 166; lactating cows: 129 vs. 109; heifers: 55 vs. 57 for IG vs. CG). Conception rates were lower in IG than in CG (all cattle: 40.2 vs. 69.3%; lactating dairy cows: 38.0 vs. 67.0%; dairy heifers: 45.5 vs. 73.7%, for IG vs. CG). Conception rate was not affected by season or live weight in heifers and lactating cows. In addition, days in milk at AI, milk production, body condition score, and parity did not affect conception in lactating cows. In summary, development of the first-wave DF in the ovary ipsilateral to the CL was associated with reduced conception rates in both lactating cows and heifers.