Earticle

Recent advancement in microarray technologies and large high throughput data generated has made it very challenging to decipher and draw a feasible biological conclusion from current microarray experiments. The difficulty arises when the number of samples available for analysis is smaller than the huge numbers of genes that need to be considered. Currently, pathway analysis is a preferable tool in extracting and understanding the biological information obtained from high throughput experiments. It is essential to analyze microarray experiments along with their biological information to represent the underlying structure of the biological network. Currently, there are numerous software developed for pathway analysis available with the same goal of mining the information from the microarray experiments with biological relevance over the extensive amounts of data. This paper discusses the comparisons between pathway analysis software in terms of their performance, advantages and limitations as well as the available pathway databases in terms of their data availability and organization. The aim of this review is to provide a better understanding of the capabilities of these software and helps to select the tools most suited for a particular purpose.

The estimation of probability density function (pdf) by the nonparametric kernel methods requires a reliable estimate of the bandwidth. There have been several studies on how to efficiently estimate this parameter. In this work, we propose a new optimization method of the smoothing parameter of the variable kernel estimator (VKE) based on the statistical properties of the probability distributions of random variables. In this setting, we show how to use the maximum entropy principle for estimating the smoothing parameter. The optimized estimator is after used in building the Bayesian classifier. In the same setting, the estimated probability density function is called optimal in the sense of having a minimum error rate of classifying data. Finally, a practical implementation with the aid of a dataset of DNA microarray is used to illustrate the behavior of the optimization technique.

The Gram-positive soil bacterium, S. coelicolor A3(2) produces various antibiotics. There are various approaches to improve antibiotics production and we focused transcriptional regulators. By controlling not only pathway-specific transcriptional regulators but also novel regulators, we can control antibiotics production. To find novel regulators, we used DACA method developed in our lab, and the method to analyse microarray data. It is possible to get the raw array data from the NCBI’s GEO. We obtained 11 Streptomyces GEO datasets, which are the mRNA expression data from various conditions. After combining every dataset, we performed normalization. And we compared every normalized profile with interesting antibiotics cluster profiles, and in this case, we used RED cluster and ACT cluster. By using Pearson product-moment correlation, we selected 119 genes using RED cluster and 85 genes using ACT cluster. Our hypothesis is that genes which show similar expression profiles with antibiotics cluster have possibility to affect antibiotics production. In conclusion, we found 15 regulator genes, some are already reported and the others are not yet reported, but studied in our lab.

Aiming at the characteristics of high dimension and small samples in microarray data, this paper proposes a selective ensemble method to classify microarray data. Firstly, kruskal-wallis test is used to filter irrelevant genes with classification task and to obtain a set of genes, and then a reduced training set is produced from original training set according to gene subset obtained. Secondly, multiple gene subsets are generated by using neighborhood rough set model with different radius and used to construct training subsets on above reduced training set. Thirdly, every constructed training subset is used to train a classifier by using SVM algorithm, and then multiple classifiers are produced as base classifiers. Finally, a set of base classifiers are selected by using teaching-learning-based optimization and build an ensemble classifier by weighted voting. Five benchmarks tumor microarray datasets are applied to evaluate performance of our proposed method. Experimental results indicate our proposed method is very effective and efficient for classifying microarray data, and it improves not only classification accuracy, but also decrease memory costs and computation times.

To promote molecular biology studies, public repositories for microarray data need to be constructed; the minimum contents for analysis of microarray experiment have been defined and standardized. Public repositories have been constructed by some researches which follow the standards such as MIAME-compliant data and MAGE-OM/ML. However, enough consideration has not been taken into the design of storage structure for the hierarchy of microarray data. In this paper, we propose alternative mapping strategy to mine the structural similarity and an advanced mapping rule from the algorithm. Object-relational mapping technique is used for extracting advanced storage design schema for microarray data and structural similarity of elements is evaluated for efficient storage construction. The mapping strategy reduced the number of relational tables remarkably. The strategy will contribute to design of the storage structure of microarray data and performance enhancement of a public repository.

Public repositories for microarray data have been constructed by some researches which follow the standards such as MIAME-compliant data and MAGE-OM/ML. However, enough consideration has not been taken into the design of storage structure for the hierarchy of microarray data. In this paper, we propose alternative mapping strategy to mine the structural similarity and an advanced mapping rule from the algorithm. Object-relational mapping technique is used for extracting advanced storage design schema for microarray data and structural similarity of elements is evaluated for efficient storage construction. The mapping strategy reduced the number of relational tables remarkably.

DNA microarray technology can be used to measure expression levels for thousands of genes in a single experiment across different samples. Within a gene expression matrix there are usually several particular Macroscopic Phenotypes of samples related to some diseases or drug effects such as diseased samples, normal samples or drug treated samples. The goal of sample based clustering is to find the phenotype structure or substructure of the samples. Currently most of research work focuses on the supervised analysis, relatively less attention has been paid to unsupervised approaches in sample based analysis which is important when domain knowledge is incomplete or hard to obtain. The standard k-means algorithm is effective in producing clusters for many practical applications. But the computational complexity of the original k-means algorithm is very high in high dimensional data and the accuracy of the clustering result depends on the initial centroid. In this paper, we present a new framework for unsupervised sample based clustering using informative genes for microarray data. We proposed a method to find initial centroid for k-means and we have used similarity measure to find the informative genes. The goal of our clustering approach is to perform better cluster discovery on sample with informative gene.

In this work, we focus on nonparametric kernel methods for estimating the probability density function (pdf). The convergence of a kernel estimator depends crucially on the choice of the smoothing parameter. We present in this paper, a new method for optimizing the bandwidth of an estimator of the probability density function: the adaptive kernel estimator. This optimized estimator is used to construct the Bayes classifier. In this sense, we have proposed a new approach to optimize the pdf based on the statistical properties of the probability distributions of random variables. We adopt the maximum entropy principle (MEP) in order to determine the optimal value of the smoothing parameter used in the estimator. In the proposed criterion, the estimated probability density function is called optimal in the sense of having a minimum error rate of classifying data. Finally, we illustrate the robustness of our optimization process of the kernel estimation methods by using a set of DNA microarray data showing that our approach effectively improves the performance of the classification process.

Recently, there has been greater attention to the use of classifier systems in medical diagnosis. Medical diagnostic tools provide automated procedures for objective decisions by making use of quantitative measures and machine learning techniques. These tools are effective and helpful for medical experts to diagnose diseases. One of such diseases is breast cancer which is the second largest cause of cancer deaths among women. To build an intelligent tool, it is very important to have an effective set of features. Two types of feature sets have been commonly implemented for the purpose of breast cancer diagnosis: image shape-based features and microarray gene expression data. Both types of feature sets have been widely implemented; however, there has been no work that directly compared the classification performance of these two feature sets. In this paper, we intensively review related works that used both types of feature sets and we also review the implemented machine learning algorithms. Moreover, we run extensive experiments to compare the classification performance of the aforementioned feature sets. Our results show that the image shape-based features are more discriminative for breast cancer classification when tested with ten-fold cross validation. To check the robustness of the best performing feature set, we further examine it with five-fold cross validation and with a variety of generative classification algorithms.

난소암은 조기진단이 어려워 대부분 진행된 병기(III기 또는 IV기)에서 발견되며, 이로 인해 부인과 암 중 가장 높은 사망률을 보인다. 병기에 따라 치료 전략과 예후가 크게 달라지므로, 정확한 병기 예측은 임상적으로 매우 중요하 다. 본 연구에서는 난소암의 병기 예측에 활용할 수 있는 유전자 또는 유전자 군을 탐색하기 위해, Affymetrix GPL570 플랫폼 기반의 여섯 개 마이크로어레이 데이터셋을 통합하여 총 995개의 샘플을 확보하였다. 통합 마이크로어레이 데이 터에 랜덤 포레스트(Random Forest) 모델을 적용하고, 피처 중요도에 기반하여 100개의 유전자를 선별하였다. 이들 유전자의 발현 패턴을 활용한 이진 분류(초기 병기 I–II vs. 진행 병기 III–IV)에서 총괄 정확도 98.0%를 달성하였다. 본 연구에서 도출된 유전자군은 난소암의 병기 예측을 기반으로 한 정밀 의료 전략 수립에 유용하게 활용될 수 있을 것으로 기대된다.

Ovarian cancer is difficult to diagnose early and is typically detected at advanced stages (Stage III or IV), resulting in the highest mortality rate among gynecological cancers. As treatment strategies and prognosis vary significantly according to stage, accurate stage prediction is clinically crucial. In this study, we integrated six microarray datasets based on the Affymetrix GPL570 platform, securing a total of 995 samples to explore genes or gene sets that could be utilized for ovarian cancer stage prediction. By applying a Random Forest model to the integrated data and selecting 100 genes based on feature importance, we achieved an overall accuracy of 98.0% in binary classification (early stages I-II vs. advanced stages III-IV) using the expression patterns of these genes. The gene set identified in this study could be effectively utilized for establishing precision medicine strategies based on ovarian cancer stage prediction.

본 논문에서는 파킨슨병의 분자병리적 기작 중 하나인 미토콘드리아 기능 장애로 인한 신경세포 사멸과정을 유전자 제어 네트워크 관점에서 분석하고자 하였다. 유전자 상호 간의 미치는 영향을 추정하기 위해 상태변수와 관측변 수로 구성된 상태공간 모델을 이용하였으며, 상태공간 모델의 파라미터는 파킨슨병 실험모델로부터 얻어진 시계열 마이 크로어레이 데이터를 통해 계산하였다. 유전자 제어 네트워크는 상태공간 모델의 파라미터를 통해 계산된 유전자-유전 자 상호작용 행렬로부터 구축되었으며, 지속적인 발현 유지를 가능하게 하는 양의 self-loop 구조를 보이는 15개 유전 자가 관측되었으며, 이 가운데 8개는 허브 유전자였다. 이들 유전자는 세포 사멸과정에서 지속적인 발현을 유지하면서 중요한 역활을 감당할 것으로 판단되며 미토콘드리아 기능 장애로 발병되는 파킨슨병에 대한 치료전략 개발에 실마리를 제공할 수 있을 것으로 기대된다.

In this paper, we aimed to analyze the neuronal cell death process resulting from mitochondrial dysfunction, which is one of the molecular pathological mechanisms of Parkinson's disease from the perspective of a gene regulatory network. To estimate the influence between genes, a state space model consisting of state variables and observation variables was utilized, and the parameters of the state-space model were calculated through time-series microarray data obtained from a Parkinson's disease experimental model. The gene regulatory network was constructed from the gene-gene interaction matrix calculated through the parameters of the state-space model. Fifteen genes showing a positive self-loop structure that enables continuous gene expression were observed, among which eight were hub genes. These genes are judged to play a critical role while maintaining continuous expression during the cell death process, and it is expected that they can provide insights for developing therapeutic strategies for Parkinson's disease caused by mitochondrial dysfunction.

A normal mixture model with which dependence between classes is incorporated is proposed in order to detect differentially expressed genes. Gene clustering approaches suffer from the high dimensional column of microarray expression data matrix which leads to the over-fit problem. Various methods are proposed to solve the problem. In this paper, use of simple averaging data within each class is proposed to overcome the various problems due to high dimensionality when the normal mixture model is fitted. Some experiments through simulated data set and real data set show its availability in actuality.

Due to the polygenic nature of cancer, it is believed that breast cancer is caused by the perturbation of multiple genes and their complex interactions, which contribute to the wide aspects of disease phenotypes. A systems biology approach for the identification of subnetworks of interconnected genes as functional modules is required to understand the complex nature of diseases such as breast cancer. In this study, we apply a 3-step strategy for the interpretation of microarray data, focusing on identifying significantly perturbed metabolic pathways rather than analyzing a large amount of overexpressed and underexpressed individual genes. The selected pathways are considered to be dysregulated functional modules that putatively contribute to the progression of disease. The subnetwork of protein-protein interactions for these dysregulated pathways are constructed for further detailed analysis. We evaluated the method by analyzing microarray datasets of breast cancer tissues; i.e., normal and invasive breast cancer tissues. Using the strategy of microarray analysis, we selected several significantly perturbed pathways that are implicated in the regulation of progression of breast cancers, including the extracellular matrix-receptor interaction pathway and the focal adhesion pathway. Moreover, these selected pathways include several known breast cancer-related genes. It is concluded from this study that the present strategy is capable of selecting interesting perturbed pathways that putatively play a role in the progression of breast cancer and provides an improved interpretability of networks of protein-protein interactions.

Recently, due to molecular biology and engineering technology, DNA microarray makes people watch thousands of genes and the state of variation from the tissue samples of living body. With DNA Microarray, it is possible to construct a genetic group that has similar expression patterns and grasp the progress and variation of gene. This paper practices Cluster Analysis which purposes the discovery of biological subgroup or class by using gene expression information. Hence, the purpose of this paper is to predict a new class which is unknown, open leukaemia data are used for the experiment, and MCL (Markov CLustering) algorithm is applied as an analysis method. The MCL algorithm is based on probability and graph flow theory. MCL simulates random walks on a graph using Markov matrices to determine the transition probabilities among nodes of the graph. If you look at closely to the method, first, MCL algorithm should be applied after getting the distance by using Euclidean distance, then inflation and diagonal factors which are tuning modulus should be tuned, and finally the threshold using the average of each column should be gotten to distinguish one class from another class. Our method has improved the accuracy through using the threshold, namely the average of each column. Our experimental result shows about 70% of accuracy in average compared to the class that is known before. Also, for the comparison evaluation to other algorithm, the proposed method compared to and analyzed SOM (Self-Organizing Map) clustering algorithm which is divided into neural network and hierarchical clustering. The method shows the better result when compared to hierarchical clustering. In further study, it should be studied whether there will be a similar result when the parameter of inflation gotten from our experiment is applied to other gene expression data. We are also trying to make a systematic method to improve the accuracy by regulating the factors mentioned above.

DNA microarray data analysis is a new technology to investigate the expression levels of thousands of genes simultaneously. Since DNA microarray data structures are various and complicative, the data are generally stored in databases for approaching to and controlling the data effectively. But we have some difficulties to analyze and control the data when the data are stored in the several database management systems or that the data are stored to the file format. The existing analysis tools for DNA microarray data have many difficult problems by complicated instructions, and dependency on data types and operating system. In this paper, we design and implement network-based analysis tool for obtaining to useful information from DNA microarray data. When we use this tool, we can analyze effectively DNA microarray data without special knowledge and education for data types and analytical methods.

A systems biology approach for the identification of perturbed molecular functions is required to understand the complex progressive disease such as breast cancer. In this study, we analyze the microarray data with Gene Ontology terms of molecular functions to select perturbed molecular functional modules in breast cancer tissues based on the definition of Gene ontology Functional Code. The Gene Ontology is three structured vocabularies describing genes and its products in terms of their associated biological processes, cellular components and molecular functions. The Gene Ontology is hierarchically classified as a directed acyclic graph. However, it is difficult to visualize Gene Ontology as a directed tree since a Gene Ontology term may have more than one parent by providing multiple paths from the root. Therefore, we applied the definition of Gene Ontology codes by defining one or more GO code(s) to each GO term to visualize the hierarchical classification of GO terms as a network. The selected molecular functions could be considered as perturbed molecular functional modules that putatively contributes to the progression of disease. We evaluated the method by analyzing microarray dataset of breast cancer tissues; i.e., normal and invasive breast cancer tissues. Based on the integration approach, we selected several interesting perturbed molecular functions that are implicated in the progression of breast cancers. Moreover, these selected molecular functions include several known breast cancer-related genes. It is concluded from this study that the present strategy is capable of selecting perturbed molecular functions that putatively play roles in the progression of diseases and provides an improved interpretability of GO terms based on the definition of Gene Ontology codes.

Since microarray data structures are various and complicative, the data are generally stored in databases for approaching to and controlling the data effectively. But we have some difficulties to analyze and control the data when the data are stored in the several database management systems. The existing analysis tools for DNA microarray data have many difficult problems by complicated instructions, and dependency on data types and operating system, and high cost, etc. In this paper, we design and implement the web-based analysis tool for obtaining to useful information from DNA microarray data. When we use this tool, we can analyze effectively DNA microarray data without special knowledge and education for data types and analytical methods.

Background/Aims: The integration of multiple profiling data and the construction of a transcriptional regulatory network may provide additional insights into the molecular mechanisms of hepatocellular carcinoma (HCC). The present study was conducted to investigate the deregulation of genes and the transcriptional regulatory network in HCC. Methods: An integrated analysis of HCC gene expression datasets was performed in Gene Expression Omnibus. Functional annotation of the differentially expression genes (DEGs) was conducted. Furthermore, transcription factors (TFs) were identified, and a global transcriptional regulatory network was constructed. Results: An integrated analysis of eight eligible gene expression profiles of HCC led to 1,835 DEGs. Consistent with the fact that the cell cycle is closely related to various tumors, the functional annotation revealed that genes involved in the cell cycle were significantly enriched. A transcriptional regulatory network was constructed using the 62 TFs, which consisted of 872 TF-target interactions between 56 TFs and 672 DEGs in the context of HCC. The top 10 TFs covering the most downstream DEGs were ZNF354C, NFATC2, ARID3A, BRCA1, ZNF263, FOXD1, GATA3, FOXO3, FOXL1, and NR4A2. This network will appeal to future investigators focusing on the development of HCC. Conclusions: The transcriptional regulatory network can provide additional information that is valuable in understanding the underlying molecular mechanism in hepatic tumorigenesis.