Krzysztof Simek

Gene Expression Profile of Papillary Thyroid Cancer: Sources of Variability and Diagnostic Implications

The study looked for an optimal set of genes differentiating between papillary thyroid cancer (PTC) and normal thyroid tissue and assessed the sources of variability in gene expression profiles. The analysis was done by oligonucleotide microarrays (GeneChip HG-U133A) in 50 tissue samples taken intraoperatively from 33 patients (23 PTC patients and 10 patients with other thyroid disease). In the initial group of 16 PTC and 16 normal samples, we assessed the sources of variability in the gene expression profile by singular value decomposition which specified three major patterns of variability. The first and the most distinct mode grouped transcripts differentiating between tumor and normal tissues. Two consecutive modes contained a large proportion of immunity-related genes. To generate a multigene classifier for tumor-normal difference, we used support vector machines-based technique (recursive feature replacement). It included the following 19 genes: DPP4, GJB3, ST14, SERPINA1, LRP4, MET, EVA1, SPUVE, LGALS3, HBB, MKRN2, MRC2, IGSF1, KIAA0830, RXRG, P4HA2, CDH3, IL13RA1, and MTMR4, and correctly discriminated 17 of 18 additional PTC/normal thyroid samples and all 16 samples published in a previous microarray study. Selected novel genes (LRP4, EVA1, TMPRSS4, QPCT, and SLC34A2) were confirmed by Q-PCR. Our results prove that the gene expression signal of PTC is easily detectable even when cancer cells do not prevail over tumor stroma. We indicate and separate the confounding variability related to the immune response. Finally, we propose a potent molecular classifier able to discriminate between PTC and nonmalignant thyroid in more than 90% of investigated samples.

Using SVD and SVM methods for selection, classification, clustering and modeling of DNA microarray data

Abstract DNA microarray technology is the latest and the most advanced tool for parallel measuring of the activity and interactions of thousands of genes. This modern technology promises new insight into mechanisms of living systems, for example only two high-density oligonucleotide microarrays are sufficient to inspect the whole human genome. However, it provides unprecedented amount of data that require application of advanced computational methods. The appropriate choice of data analysis technique depends both on data and on goals of an experiment. In this paper we focus on two promising methods: singular value decomposition and support vector machines. We discuss the possibility of application of these methods for different purposes; particularly for clustering, classification, feature selection and modeling of dynamics of gene expression. We use for testing presented approaches existing data sets, which are widely available via Internet, and one new tumor/normal thyroid microarray data set.

Advanced Technologies for Intelligent Systems of National Border Security

One of the worlds leading problems in the field of national security is protection of borders and borderlands. This book addresses multiple issues on advanced innovative methods of multi-level control of both ground (UGVs) and aerial drones (UAVs). Those objects combined with innovative algorithms become autonomous objects capable of patrolling chosen borderland areas by themselves and automatically inform the operator of the system about potential place of detection of a specific incident. This is achieved by using sophisticated methods of generation of non-collision trajectory for those types of objects and enabling automatic integration of both ground and aerial unmanned vehicles. The topics included in this book also cover presentation of complete information and communication technology (ICT) systems capable of control, observation and detection of various types of incidents and threats. This book is a valuable source of information for constructors and developers of such solutions for uniformed services. Scientists and researchers involved in computer vision, image processing, data fusion, control algorithms or IC can find many valuable suggestions and solutions. Multiple challenges for such systems are also presented.

Gene Expression Profiling in Hereditary, BRCA1-linked Breast Cancer: Preliminary Report

Global analysis of gene expression by DNA microarrays is nowadays a widely used tool, especially relevant for cancer research. It helps the understanding of complex biology of cancer tissue, allows identification of novel molecular markers, reveals previously unknown molecular subtypes of cancer that differ by clinical features like drug susceptibility or general prognosis. Our aim was to compare gene expression profiles in breast cancer that develop against a background of inherited predisposing mutations versus sporadic breast cancer. In this preliminary study we analysed seven hereditary, BRCA1 mutation-linked breast cancer tissues and seven sporadic cases that were carefully matched by histopathology and ER status. Additionally, we analysed 6 samples of normal breast tissue. We found that while the difference in gene expression profiles between tumour tissue and normal breast can be easily recognized by unsupervised algorithms, the difference between those two types of tumours is more discrete. However, by supervised methods of data analysis, we were able to select a set of genes that may differentiate between hereditary and sporadic tumours. The most significant difference concerns genes that code for proteins engaged in regulation of transcription, cellular metabolism, signalling, proliferation and cell death. Microarray results for chosen genes (TOB1, SEPHS2) were validated by real-time RT-PCR.

Unsupervised analysis reveals two molecular subgroups of serous ovarian cancer with distinct gene expression profiles and survival

PurposeOvarian cancer is typically diagnosed at late stages, and thus, patients’ prognosis is poor. Improvement in treatment outcomes depends, at least partly, on better understanding of ovarian cancer biology and finding new molecular markers and therapeutic targets.MethodsAn unsupervised method of data analysis, singular value decomposition, was applied to analyze microarray data from 101 ovarian cancer samples; then, selected genes were validated by quantitative PCR.ResultsWe found that the major factor influencing gene expression in ovarian cancer was tumor histological type. The next major source of variability was traced to a set of genes mainly associated with extracellular matrix, cell motility, adhesion, and immunological response. Hierarchical clustering based on the expression of these genes revealed two clusters of ovarian cancers with different molecular profiles and distinct overall survival (OS). Patients with higher expression of these genes had shorter OS than those with lower expression. The two clusters did not derive from high- versus low-grade serous carcinomas and were unrelated to histological (ovarian vs. fallopian) origin. Interestingly, there was considerable overlap between identified prognostic signature and a recently described invasion-associated signature related to stromal desmoplastic reaction. Several genes from this signature were validated by quantitative PCR; two of them—DSPG3 and LOX—were validated both in the initial and independent sets of samples and were significantly associated with OS and disease-free survival.ConclusionsWe distinguished two molecular subgroups of serous ovarian cancers characterized by distinct OS. Among differentially expressed genes, some may potentially be used as prognostic markers. In our opinion, unsupervised methods of microarray data analysis are more effective than supervised methods in identifying intrinsic, biologically sound sources of variability. Moreover, as histological type of the tumor is the greatest source of variability in ovarian cancer and may interfere with analyses of other features, it seems reasonable to use histologically homogeneous groups of tumors in microarray experiments.

Intelligent Robust Control of Fault Tolerant Linear Systems

Abstract One way to model linear systems with fault-prone tolerant structure is to use theory of piecewise deterministic processes. In this paper, we consider the continuous time linear systems with abrupt changes of parameters modelled by discrete-state Markov processes. Under the assumption of the existence of a suitable control law, we give the necessary and sufficient conditions for the stochastic stabilizability of the nominal model of this class of systems using intelligent controllers. Moreover assuming that the parameters of the real system may differ from their nominal values we present sufficient conditions of the robustness for the real systems.

SVD Analysis of Gene Expression Data

The analysis of gene expression profiles of cells and tissues, performed by DNA microarray technology, strongly relies on proper bioinformatical methods of data analysis. Due to the large number of analyzed variables (genes) and the usually low number of cases (arrays) in one experiment, limited by the high cost of the technology, the biological reasoning is difficult without previous analysis, leading to a reduction of the problem dimensionality. A wide variety of methods have been developed; the most useful, from a biological point of view, are methods of supervised gene selection withestimation of false discovery rate. However, supervised gene selection is not always satisfying for the user of microarray technology, as the complexity of biological systems analyzed by microarrays rarely can be explained by one variable. Among unsupervised methods of analysis, hierarchical clustering and principal component analysis (PCA) have gained wide biological application. In our opinion, singular value decomposition (SVD) analysis, which is similar to PCA, has additional advantages that are very essential for the interpretation of the biological data. In this chapter we shall present how to apply SVD to unsupervised analysis of transcriptome data obtained by oligonucleotide microarrays. These results have been derived from several experiments, carried out at the DNA oligonucleotide microarray Laboratory at the Institute of Oncology, Gliwice, and are currently analyzed from a biological point of view.

Class Prediction and Pattern Discovery in Microarray Data - Artificial Intelligence and Algebraic Methods

Abstract—In the paper we present a brief survey of our results in processing of data from DNA microarray experiments obtained in our collaborative research with M.C. Sklodowska Centre of Oncology. Our experience therefore is strictly connected with problems resulting from cancer diagnosis and therapy but many results have more general issue. We focus our attention on three important stages of microarray data processing e.g. class prediction, gene selection and pattern discovery.

Robust Stabilization of Fault Tolerant Decentralized Linear Systems

Abstract In this paper, the continuous time linear systems with abrupt changes of parameters modelled by discrete-state Markov processes are considered. Under the assumption of the existence of a suitable control law, the necessary and sufficient conditions for the stochastic stabilizability of the nominal model of this class of systems using JLQ controllers is given. Moreover assuming that the parameters of the real system may differ from their nominal values and the existance of external disturbances sufficient conditions of the robustness for the real systems are presented. Two different control structures are proposed: centralized and decentalized. The comparision of the results is discussed.