Gene Selection

Background: Microarray technology is a powerful methodology for identifying differentially expressed genes. However, when thousands of genes in a microarray data set are evaluated simultaneously by fold changes and significance tests, the probability of detecting false positives rises sharply. In this first microarray study of brachial plexus injury, we applied and compared the performance of two recently proposed algorithms for tackling this multiple testing problem, Significance Analysis of Microarrays (SAM) and Westfall and Young step down adjusted p values, as well as t-statistics and Welch statistics, in specifying differential gene expression under different biological states. Results: Using SAM based on t statistics, we identified 73 significant genes, which fall into different functional categories, such as cytokines / neurotrophin, myelin function and signal transduction. Interestingly, all but one gene were down-regulated in the patients. Using Welch statistics in conjunction with SAM, we identified an additional set of up-regulated genes, several of which are engaged in transcription and translation regulation. In contrast, the Westfall and Young algorithm identified only one gene using a conventional significance level of 0.05. In coping with multiple testing problems, Family-wise type I error rate (FWER) and false discovery rate (FDR) are different expressions of Type I error rates. The Westfall and Young algorithm controls FWER. In the context of this microarray study, it is, seemingly, too conservative. In contrast, SAM, by controlling FDR, provides a promising alternative. In this instance, genes selected by SAM were shown to be biologically meaningful.

Background: An important use of data obtained from microarray measurements is the classification of tumor types with respect to genes that are either up or down regulated in specific cancer types. A number of algorithms have been proposed to obtain such classifications. These algorithms usually require parameter optimization to obtain accurate results depending on the type of data. Additionally, it is highly critical to find an optimal set of markers among those up or down regulated genes that can be clinically utilized to build assays for the diagnosis or to follow progression of specific cancer types. In this paper, we employ a mixed integer programming based classification algorithm named hyper-box enclosure method (HBE) for the classification of some cancer types with a minimal set of predictor genes. This optimization based method which is a user friendly and efficient classifier may allow the clinicians to diagnose and follow progression of certain cancer types. Methodology/Principal Findings: We apply HBE algorithm to some well known data sets such as leukemia, prostate cancer, diffuse large B-cell lymphoma (DLBCL), small round blue cell tumors (SRBCT) to find some predictor genes that can be utilized for diagnosis and prognosis in a robust manner with a high accuracy. Our approach does not require any modification or parameter optimization for each data set. Additionally, information gain attribute evaluator, relief attribute evaluator and correlation-based feature selection methods are employed for the gene selection. The results are compared with those from other studies and biological roles of selected genes in corresponding cancer type are described. Conclusions/Significance: The performance of our algorithm overall was better than the other algorithms reported in the literature and classifiers found in WEKA data-mining package. Since it does not require a parameter optimization and it performs consistently very high prediction rate on different type of data sets, HBE method is an effective and consistent tool for cancer type prediction with a small number of gene markers.

Variation in cellular activity in a tissue induces changes in RNA concentration, which affects the validity of gene mRNA abundance analyzed by reverse transcription quantitative PCR (RT-qPCR). A common way of accounting for such variation consists of the use of reference genes for normalization. Programs such as geNorm may be used to select suitable reference genes, although a large set of genes that are not coregulated must be analyzed to obtain accurate results. The objective of this study was to propose an alternative experimental and analytical protocol to assess the invariance of reference genes in porcine mammary tissue using mammary RNA and DNA concentrations as correction factors. Mammary glands were biopsied from 4 sows on d 110 of gestation (prepartum), on d 5 (early) and 17 (peak) of lactation, and on d 5 after weaning (postweaning). Relative expression of 7 potential reference genes, API5, MRPL39, VAPB, ACTB, GAPDH, RPS23, and MTG1, and one candidate gene, SLC7A1, was quantified by RT-qPCR using a relative standard curve approach. Variation in gene expression levels, measured as cycles to threshold at each stage of mammary physiological activity, was tested using a linear mixed model fitting RNA and DNA concentrations as covariates. Results were compared with those obtained with geNorm analysis, and genes selected by each method were used to normalize SLC7A1. Quantified relative mRNA abundance of GAPDH and MRPL39 remained unchanged across stages of mammary physiological activity after accounting for changes in tissue RNA and DNA concentration. In contrast, geNorm analysis selected MTG1, MRPL39, and VAPB as the best reference genes. However, when target gene SL-C7A1 was normalized with genes selected either based on our proposed protocol or by geNorm, fold changes in mRNA abundance did not differ. In conclusion, the proposed analytical protocol assesses expression invari-ance of potential reference genes by accounting for variation in tissue RNA and DNA concentrations and thus represents an alternative method to select suitable reference genes for RT-qPCR analysis.

In the field of medical science diseases diagnosis by Tissue microarrays is one of the active areas of research .There are various gene selection techniques in the literature. Gene selection provides genes subsets that are capable to describe in which category those gene are (active, hyperactive or silent).Various application areas like combinatorial chemistry, text mining, multivariate imaging, or bioinformatics are using huge data sets. The problem has been addressed of selection of a small subset of genes from broad patterns of gene expression data, recorded on DNA micro-arrays for cancer classification. Usually till now survey paper discuss various conventional & evolutionary methods of gene selection like filters, wrappers methods.

Microarray technologies are rapidly becoming available for new species including teleost fishes. We constructed a rainbow trout cDNA microarray targeted at the identification of genes which are differentially expressed in response to environmental stressors. This platform included clones from normalized and subtracted libraries and genes selected through functional annotation. Present study focused on time-course comparisons of stress responses in the brain and kidney and the identification of a set of genes which are diagnostic for stress response. Fish were stressed with handling and samples were collected 1, 3 and 5 days after the first exposure. Gene expression profiles were analysed in terms of Gene Ontology categories. Stress affected different functional groups of genes in the tissues studied. Mitochondria, extracellular matrix and endopeptidases (especially collagenases) were the major targets in kidney. Stress response in brain was characterized with dramatic temporal alter...

Kinesin super family protein 2A (KIF2A) is an ATP-dependent microtubule destabilizer, that belongs to the kinesin-13 family. It is highly expressed in juvenile brains, but its postnatal function has not been determined due to mortality in KIF2A-deficient mice. In addition, KIF2A is a causal gene in human Malformation of Cortical Development (MCD) with epilepsy, but the contribution of KIF2A to its pathogenesis is not yet understood due to the small number of human patients. In this study, we first analyzed the prenatal function of KIF2A using kif2a-KO mice. These mice were pale, breathed irregularly, and exhibited frequent twitching in addition to malformations resembling those in kif2a-mutated human patients. To determine the post-migration function of KIF2A, we generated newly tamoxifen-inducible conditional knockout mice. Despite successful neuronal migration, all offspring displayed hyper-activity, weight loss, severe cortical/hippocampal-focus epilepsy, and died. Interestingly, KIF2A was highly expressed in excitatory axons, in cortical neurites in layers II/III/V and in dentate mossy fibers (MF). In the cortex, the loss of KIF2A generated aberrant axon sprouting in those layers. In the hippocampus, the loss of KIF2A resulted in the development of hippocampal sclerosis, MF sprouting, and recurrent excitatory circuits resembling but different from TLE. These phenotypes developed without excitation. cKO granule cells exhibited failed axon/dendrite determination, and developed multiple short axons throughout the entire molecular layer. These results suggest that KIF2A is crucial postnatally for establishing accurate neuronal circuits, in addition to its role in prenatal cell survival, migration, and axon elongation.

Cancer has been suggested to result from interactions between genetic and environmental factors, and certain subgroups in the general population may be at increased risk because of their relatively higher susceptibility to environmental carcinogens. The current study, part of a large biomonitoring study conducted in Flanders from 2002 to 2006 (The Flanders Environment and Health Survey), aims to determine these susceptible subpopulations based on multiple genotypic differences between individuals. A random selection of 429 adolescents and 361 adults was genotyped for 36 polymorphisms in 23 genes selected because of their known role in carcinogen metabolism, DNA repair, and oxidative stress. In both age groups, relationships between endogenous exposure to organochloride substances (polychlorinated biphenyl, hexachlorobenzene, dichlorodiphenyl dichloroethane), metals (cadmium, lead), and urinary metabolites (1-hydroxypyrene, trans-trans muconic acid) versus genotoxic effects (Comet as...

Summary A serial analysis of gene expression (SAGE) library is a collection of thousands of small DNA “tags,” each of which represents a distinct messenger RNA (mRNA) transcript. Existing methods have been proposed for analyzing single library data (i.e., one library per group) or one tag at a time. The practice of lumping all libraries together (in a multi‐library setting) to form a “mega” library for each group is obviously unsatisfactory, but nonetheless performed frequently due to the lack of alternative methods. Because the tag counts within each library are interrelated as they are drawn from a multinomial distribution, analyzing thousands of tags one at a time is undoubtedly inadequate. Not only does such a practice ignore the dependency, but it also faces the multiple testing adjustment issue. This article is an attempt to address both of these issues so that all tags from multi‐library groups can be analyzed together. The methods proposed also gear toward multi‐group data. ...

Microarray experiments generate large datasets with expression values for thousands of genes, but not more than a few dozens of samples. A challenging task with these data is to reveal groups of genes which act together and whose collective expression is strongly associated with an outcome variable of interest. To find these groups, we suggest the use of supervised algorithms: these are procedures which use external information about the response variable for grouping the genes. We present Pelora, an algorithm based on penalized logistic r egression analysis, that combines gene selection, gene grouping and sample classification in a supervised, simultaneous way. With an empirical study on six different microarray datasets, we show that Pelora identifies gene groups whose expression centroids have very good predictive potential and yield results that can keep up with state-of-the-art classification methods based on single genes. Thus, our gene groups can be beneficial in medical diagnostics and prognostics, but they may also provide more biological insights into gene function and regulation.

Although X chromosome inactivation in female mammals evolved to balance the expression of X chromosome and autosomal genes in the two sexes, female embryos pass through developmental stages in which both X chromosomes are active in somatic cells. Bovine blastocysts show higher expression of many X genes in XX than XY embryos, suggesting that X inactivation is not complete. Here we reanalyzed bovine blastocyst microarray expression data from a network perspective with a focus on interactions between X chromosome and autosomal genes. Whereas male to female ratios of expression of autosomal genes were distributed around a mean of 1, X chromosome genes were clearly shifted towards higher expression in females. We generated gene coexpression networks and identified a major module of genes with correlated gene expression that includes female-biased X genes and sexually dimorphic autosomal genes for which the sexual dimorphism is likely driven by the X genes. In this module, expression of X chromosome genes correlates with autosome genes, more than the expression of autosomal genes with each other. Our study identifies correlated patterns of autosomal and X-linked genes that are likely influenced by the sexual imbalance of X gene expression when X inactivation is inefficient.

Breast cancer is the second leading cause of female mortality globally. Effective diagnostic tools, such as biosensors that utilize reliable biomarkers, are essential for early detection, particularly in low-income countries. This study introduces a novel bioinformatics pipeline that uses machine learning algorithms (MLAs) to identify genetic biomarkers for classifying breast cancer into non-malignant, non-triple-negative, and triple-negative categories. Five Gene Selection Approaches (GSAs) were employed: LASSO (Least Absolute Shrinkage and Selection Operator), Membrane LASSO, Surfaceome LASSO, Network Analysis, and Feature Importance Score (FIS). We implemented three factorial designs to assess the impact of MLAs and GSAs on classification performance (F1 Macro and Accuracy) in both cell lines and patient samples. Using Recursive Feature Elimination (RFE) and Genetic Algorithms (GAs) in the first four GSAs, we reduced the gene count to eight per GSA while maintaining an F1 Macro ≥80 %. Consequently, 95.5 % of our treatments with these gene sets achieved an F1 Macro or Accuracy ranging from 70.3 % to 97.2 %. We analyzed 37 genes for their predictive power in terms of five-year survival and relapse-free survival and compared them with genes from four commercial panels. Notably, thirteen genes (MFSD2A, TMEM74, SFRP1, UBXN10, CACNA1H, ERBB2, SIDT1, TMEM129, MME, FLRT2, CA12, ESR1, and TBC1D9) showed significant predictive capabilities for up to five years of survival. TBC1D9, UBXN10, SFRP1, and MME were significant for relapse-free survival after five years. The FOXC1, MLPH, FOXA1, ESR1, ERBB2, and SFRP1 genes also matched those described in commercial panels. The influence of MLA on F1 Macro and Accuracy was not statistically significant. Altogether, the genetic biomarkers identified in this study hold potential for use in biosensors aimed at breast cancer diagnosis and treatment.

Kinesin super family protein 2A (KIF2A) is an ATP-dependent microtubule destabilizer, that belongs to the kinesin-13 family. It is highly expressed in juvenile brains, but its postnatal function has not been determined due to mortality in KIF2A-deficient mice. In addition, KIF2A is a causal gene in human Malformation of Cortical Development (MCD) with epilepsy, but the contribution of KIF2A to its pathogenesis is not yet understood due to the small number of human patients. In this study, we first analyzed the prenatal function of KIF2A using kif2a-KO mice. These mice were pale, breathed irregularly, and exhibited frequent twitching in addition to malformations resembling those in kif2a-mutated human patients. To determine the post-migration function of KIF2A, we generated newly tamoxifen-inducible conditional knockout mice. Despite successful neuronal migration, all offspring displayed hyper-activity, weight loss, severe cortical/hippocampal-focus epilepsy, and died. Interestingly, KIF2A was highly expressed in excitatory axons, in cortical neurites in layers II/III/V and in dentate mossy fibers (MF). In the cortex, the loss of KIF2A generated aberrant axon sprouting in those layers. In the hippocampus, the loss of KIF2A resulted in the development of hippocampal sclerosis, MF sprouting, and recurrent excitatory circuits resembling but different from TLE. These phenotypes developed without excitation. cKO granule cells exhibited failed axon/dendrite determination, and developed multiple short axons throughout the entire molecular layer. These results suggest that KIF2A is crucial postnatally for establishing accurate neuronal circuits, in addition to its role in prenatal cell survival, migration, and axon elongation.

Napier grass (Pennisetum purpureum Schumach.) is a highly productive C4 tropical forage grass that has been targeted as a potential bioenergy crop. To further increase the efficiency of bioethanol production by molecular breeding, a reliable protocol for genetically transforming napier grass is essential. In this study, we report the creation of transgenic napier grass plants derived from embryogenic callus cultures of shoot apices. Embryogenic callus was initiated in three accessions of napier grass and a napier grass×pearl millet hybrid using Murashige and Skoog (MS) medium supplemented with 2.0 mg L -1 2,4-dichlorophenoxyacetic acid (2,4-D), 0.5 mg L -1 6-benzylaminopurine (BAP) and 50 µM copper sulfate (CuSO 4 ). Of the accessions tested, a dwarf type with late-heading (DL line) had the best response for embryogenic callus formation. Highly regenerative calli that formed dense polyembryogenic clusters were selected as target tissues for transformation. A plasmid vector, pAHC25, containing an herbicide-resistance gene (bar) and the β-glucuronidase (GUS) reporter gene was used in particle bombardment experiments. Target tissues treated with 0.6 M osmoticum were bombarded, and transgenic plants were selected under 5.0 mg L -1 bialaphos selection. Although a total of 1400 target tissues yielded nine GUS-positive bialaphos-resistant calli, only one transgenic line that was derived from target tissue with the shortest culture term produced four transgenic plants. Thus, the length of time that the target tissue is in callus culture was one of the most important factors for acquiring transgenic plants in napier grass. This is the first report of successfully producing transgenic napier grass plants.

Cancer can be considered as one of the leading causes of death widely. One of the most effective tools to be able to handle cancer diagnosis, prognosis, and treatment is by using expression profiling technique which is based on microarray gene. For each data point (sample), gene data expression usually receives tens of thousands of genes. As a result, this data is large-scale, high-dimensional, and highly redundant. The classification of gene expression profiles is considered to be a (NP)-Hard problem. Feature (gene) selection is one of the most effective methods to handle this problem. A hybrid cancer classification approach is presented in this paper, and several machine learning techniques were used in the hybrid model: Pearson’s correlation coefficient as a correlation-based feature selector and reducer, a Decision Tree classifier that is easy to interpret and does not require a parameter, and Grid Search CV (cross-validation) to optimize the maximum depth hyperparameter. Seven ...

In microarray studies, the number of samples is relatively small compared to the number of genes per sample. An important aspect of microarray studies is the prediction of patient survival based on their gene expression profile. This naturally calls for the use of a dimension reduction procedure together with the survival prediction model. In this study, a new method based on combining wavelet approximation coefficients and Cox regression was presented. The proposed method was compared with supervised principal component and supervised partial least squares methods. The different fitted Cox models based on supervised wavelet approximation coefficients, the top number of supervised principal components, and partial least squares components were applied to the data. The results showed that the prediction performance of the Cox model based on supervised wavelet feature extraction was superior to the supervised principal components and partial least squares components. The results sugge...

Gene expression profiling has been widely used to study molecular signatures of many diseases and to develop molecular diagnostics for disease prediction. Gene selection, as an important step for improved diagnostics, screens tens of thousands of genes and identifies a small subset that discriminates between disease types. A two-step gene selection method is proposed to identify informative gene subsets for accurate classification of multiclass phenotypes. In the first step, individually discriminatory genes (IDGs) are identified by using one-dimensional weighted Fisher criterion (wFC). In the second step, jointly discriminatory genes (JDGs) are selected by sequential search methods, based on their joint class separability measured by multidimensional weighted Fisher criterion (wFC). The performance of the selected gene subsets for multiclass prediction is evaluated by artificial neural networks (ANNs) and/or support vector machines (SVMs). By applying the proposed IDG/JDG approach to two microarray studies, that is, small round blue cell tumors (SRBCTs) and muscular dystrophies (MDs), we successfully identified a much smaller yet efficient set of JDGs for diagnosing SRBCTs and MDs with high prediction accuracies (96.9% for SRBCTs and 92.3% for MDs, resp.). These experimental results demonstrated that the two-step gene selection method is able to identify a subset of highly discriminative genes for improved multiclass prediction.

It is crucial to accurately categorize cancers using microarray data. Researchers have employed a variety of computational intelligence approaches to analyze gene expression data. It is believed that the most difficult part of the problem of cancer diagnosis is determining which genes are informative. Therefore, selecting genes to study as a starting point for cancer classification is common practice. We offer a novel approach that combines the Runge Kutta optimizer (RUN) with a support vector machine (SVM) as the classifier to select the significant genes in the detection of cancer tissues. As a means of dealing with the high dimensionality that characterizes microarray datasets, the preprocessing stage of the ReliefF method is implemented. The proposed RUN–SVM approach is tested on binary-class microarray datasets (Breast2 and Prostate) and multi-class microarray datasets in order to assess its efficacy (i.e., Brain Tumor1, Brain Tumor2, Breast3, and Lung Cancer). Based on the exp...

Selection of genes that are differentially expressed and critical to a particular biological process has been a major challenge in post-array analysis. Recent development in bioinformatics has made various data sources available such as mRNA and miRNA expression profiles, biological pathway and gene annotation, etc. Efficient and effective integration of multiple data sources helps enrich our knowledge about the involved samples and genes for selecting genes bearing significant biological relevance. In this work, we studied a novel problem of multi-source gene selection: given multiple heterogeneous data sources (or data sets), select genes from expression profiles by integrating information from various data sources. We investigated how to effectively employ information contained in multiple data sources to extract an intrinsic global geometric pattern and use it in covariance analysis for gene selection. We designed and conducted experiments to systematically compare the proposed approach with representative methods in terms of statistical and biological significance, and showed the efficacy and potential of the proposed approach with promising findings.

This paper presents the development of the N-Spherical Minimalist Machine Learning (MML) classifier, an innovative model within the Minimalist Machine Learning paradigm. Using N-spherical coordinates and concepts from metaheuristics and associative models, this classifier e ectively addresses challenges such as data dimensionality and class imbalance in complex datasets. Performance evaluations using the F measure and balanced accuracy demonstrate its superior e ciency and robustness compared to state-of-the-art classifiers. Statistical validation is conducted using the Friedman and Holm tests. Although currently limited to binary classification, this work highlights the potential of minimalist approaches in machine learning for classification of highly dimensional and imbalanced data. Future extensions aim to include multi-class problems and mechanisms for handling categorical data.

The concept of linear separability of gene expression data sets with respect to two classes has been recently studied in the literature. The problem is to efficiently find all pairs of genes which induce a linear separation of the data. It has been suggested that an underlying molecular mechanism relates together the two genes of a separating pair to the phenotype under study, such as a specific cancer. In this paper we study the Containment Angle (CA) defined on the unit circle for a linearly separating gene-pair (LS-pair) as an alternative to the paired t-test ranking function for gene selection. Using the CA we also show empirically that a given classifier's error is related to the degree of linear separability of a given data set. Finally we propose gene subset selection methods based on the CA ranking function for LS-pairs and a ranking function for linearly separation genes (LS-genes), and which select only among LS-genes and LS-pairs. Our methods give better results in terms of subset sizes and classification accuracy when compared to a well-performing method, on many data sets.

It is known that breast cancer is not just one disease, but rather a collection of many different diseases occurring in one site that can be distinguished based in part on characteristic gene expression signatures. Appropriate diagnosis of the specific subtypes of this disease is critical for ensuring the best possible patient response to therapy. Currently, therapeutic direction is determined based on the expression of characteristic receptors; while cost effective, this method is not robust and is limited to predicting a small number of subtypes reliably. Using the original 5 subtypes of breast cancer we hypothesized that machine learning techniques would offer many benefits for feature selection. Unlike existing gene selection approaches, we propose a tree-based approach that conducts gene selection and builds the classifier simultaneously. We conducted computational experiments to select the minimal number of genes that would reliably predict a given subtype. Our results support that this modified approach to gene selection yields a small subset of genes that can predict subtypes with greater than 95% overall accuracy. In addition to providing a valuable list of targets for diagnostic purposes, the gene ontologies of selected genes suggest that these methods have isolated a number of potential genes involved in breast cancer biology, etiology and potentially novel therapeutics.

Human brain development is a continuum governed by differential gene expression. Therefore, we proceeded to identify genes selectively expressed in the developing brain. Using differential display and library screening, a novel rat cDNA, neuronatin, was identified and used to screen a human fetal brain cDNA library. Human neuronatin cDNA was isolated and sequenced. The cDNA was 1159 bp long and corresponded in size to the 1.25 kb message detected on Northern analysis. Neuronatin mRNA was selectively expressed in human brain during fetal development, but became repressed in adulthood. When studied in the rat, neuronatin mRNA first appeared at mid-gestation in association with the onset of neurogenesis, becoming most pronounced later in development when neuroepithelial proliferation and neuroblast commitment are manifest, and declined postnatally coinciding with the completion of neurogenesis. The deduced protein has two distinct domains, a hydrophobic N-terminal and basic C-terminal rich in arginine residues. Both the amino acid sequence and secondary structure of this amphipathic polypeptide exhibited homology to PMP1 and phospholamban, members of the "proteolipid' class of proteins which function as regulatory subunits of membrane channels. The neuronatin gene, 3973 bases long, contains in its 5'-flanking region a neural restrictive silencer element which may govern neuronspecific expression. Based on screening a somatic cell hybrid panel, neuronatin gene was assigned to chromosome-20. And, using deletion constructs of chromosome-20 and fluorescence in situ hybridization, neuronatin was localized to chromosome-20q11.2-12. In conclusion, neuronatin is a novel human gene that is developmentally regulated and expressed in the brain. The deduced protein is a proteolipid that may function as a unique regulator of ion channels during brain development. The definitive localization of neuronatin to human chromosome 20q11.2-12 provides the basis to investigate this gene as a candidate in neuro-developmental diseases that may also map to this region.

In this paper, we compare the accuracy of classification for different cancers, based on gene microarray expression data. For this reason, we have used a combination between filter selection methods and clustering algorithms to select relevant features, in each cancer dataset, for gene classification. Our effort is carried out in two steps. First, we survey the effect of the selection methods, on the classification accuracy for cancers, by comparing the performances evaluated by different classifiers. The considered selection methods in this paper are SNR, ReliefF, Correlation Coefficient, Mutual Information, T-Statistics, Fisher, Max relevance Min redundancy, and Principal component analysis. We evaluated the performances of each selection method by the use of the K Nearest Neighbor, Support Vector Machine, Linear Discriminant Analyses, Decision tree for classification and Naïve Bayes classifier for a supervised classification task. As a second step, we preceded the selection step ...

Kinesin super family protein 2A (KIF2A) is an ATP-dependent microtubule destabilizer, that belongs to the kinesin-13 family. It is highly expressed in juvenile brains, but its postnatal function has not been determined due to mortality in KIF2A-deficient mice. In addition, KIF2A is a causal gene in human Malformation of Cortical Development (MCD) with epilepsy, but the contribution of KIF2A to its pathogenesis is not yet understood due to the small number of human patients. In this study, we first analyzed the prenatal function of KIF2A using kif2a-KO mice. These mice were pale, breathed irregularly, and exhibited frequent twitching in addition to malformations resembling those in kif2a-mutated human patients. To determine the post-migration function of KIF2A, we generated newly tamoxifen-inducible conditional knockout mice. Despite successful neuronal migration, all offspring displayed hyper-activity, weight loss, severe cortical/hippocampal-focus epilepsy, and died. Interestingly, KIF2A was highly expressed in excitatory axons, in cortical neurites in layers II/III/V and in dentate mossy fibers (MF). In the cortex, the loss of KIF2A generated aberrant axon sprouting in those layers. In the hippocampus, the loss of KIF2A resulted in the development of hippocampal sclerosis, MF sprouting, and recurrent excitatory circuits resembling but different from TLE. These phenotypes developed without excitation. cKO granule cells exhibited failed axon/dendrite determination, and developed multiple short axons throughout the entire molecular layer. These results suggest that KIF2A is crucial postnatally for establishing accurate neuronal circuits, in addition to its role in prenatal cell survival, migration, and axon elongation.

pair linkage analysis of 24 families with pre-eclampsia, we confirm a susceptibility locus on chromosome 10q22.1 in Dutch females: a multipoint non-parametric linkage score of 3.6 near marker D10S1432 was obtained. Haplotype analysis showed a parent-of-origin effect: maximal allele sharing in the affected sibs was found for maternally derived alleles in all families, but not for the paternally derived alleles. As matrilineal inheritance suggests the presence of maternally expressed imprinted genes, while imprinting operates predominantly in (extra)embryonic tissues, all genes (n 5 132) known on 10q22 between GATA121A08 and D10S580 were screened for seven sequence-related features associated with imprinting and subsequently tested for expression in first trimester placenta. Placental expression of genes selected in this way (n 5 55) was compared with expression in androgenetic placentas of identical gestational age. Two regions on 10q22 were identified with developmentally co-repressed genes with non-random chromosomal distribution. Interestingly, these two clusters, near CTNNA3 and KCNMA1 and each containing five genes with down-regulated expression in androgenetic placentas, coincided with the regions with maximal maternal allele sharing seen in the pre-eclamptic sisters. Our linkage and expression data are compatible with the concept that pre-eclampsia involves maternally expressed imprinted genes that operate in the first trimester placenta.

Nous avons proposé un algorithme original de Fouille de Données, LICORN, afin d'inférer des relations de régulation coopérative à partir de données d'expression. LICORN donne de bons résultats s'il est appliqué à des données de levure, mais le passage à l'échelle sur des données plus complexes (e.g., humaines) est difficile. Dans cet article, nous proposons une extension de LI-CORN afin qu'il puisse gérer une contrainte de co-régulation adaptative. Une évaluation préliminaire sur des données de transcriptome de tumeurs de vessie montre que les réseaux significatifs sont obtenus à l'aide d'une contrainte de corégulation adaptative de manière beaucoup plus efficace, et qu'ils ont des performances de prédiction équivalentes voire meilleures que celles obtenues par LICORN.

L'inférence de signatures de facteurs de transcription à partir des données puces à ADN a déjà été étudié dans la communauté bioinformatique. La principale difficulté à résoudre est de trouver un ensemble d'heuristiques pertinentes, afin de contrôler la complexité de résolution de ce problème NP-difficile. Nous proposons dans cet article une solution heuristique alternative à celles utilisées dans les approches bayésiennes, fondée sur la recherche de motifs fréquents maximaux dans une matrice discrétisée issue des données numériques de puces ADN. Notre méthode

I would like to thank a number of people for their help, encouragement and support during the undertaking and writing up of this work. Firstly, my thanks go to both my supervisors Professor David Humber and Professor Robert Cheke. Without their availability and trust I would not have been able to start or finish this PhD. To Professor Humber for his enthusiasm in research, the stimulating discussion and for the constructive criticism that motivated me throughout my work and reminded me how exciting science can be. To Professor Cheke for his constant support, patience, kindness and help in overcoming any obstacle I came across during my research. To my family in Italy, especially to my parents for encouraging me to follow my aspirations even when it meant being thousands of miles apart. To my sister Daniela and her husband Bruno for looking after our family when I was far away and for ever being so supportive and proud. To Camilla and Alessandra, for being the most wonderful nieces an aunty can hope for. To my amazing friends,

Machine-learning algorithms have been widely used in breast cancer diagnosis to help pathologists and physicians in the decision-making process. However, the high dimensionality of genetic data makes the classification process a challenging task. In this paper, we propose a new optimized wrapper gene selection method that is based on a nature-inspired algorithm (simulated annealing (SA)), which will help select the most informative genes for breast cancer prediction. These optimal genes will then be used to train the classifier to improve its accuracy and efficiency. Three supervised machine-learning algorithms, namely, the support vector machine, the decision tree, and the random forest were used to create the classifier models that will help to predict breast cancer. Two different experiments were conducted using three datasets: Gene expression (GE), deoxyribonucleic acid (DNA) methylation, and a combination of the two. Six measures were used to evaluate the performance of the proposed algorithm, which include the following: Accuracy, precision, recall, specificity, area under the curve (AUC), and execution time. The effectiveness of the proposed classifiers was evaluated through comprehensive experiments. The results demonstrated that our approach outperformed the conventional classifiers as expected in terms of accuracy and execution time. High accuracy values of 99.77%, 99.45%, and 99.45% have been achieved by SA-SVM for GE, DNA methylation, and the combined datasets, respectively. The execution time of the proposed approach was significantly reduced, in comparison to that of the traditional classifiers and the best execution time has been reached by SA-SVM, which was 0.02, 0.03, and 0.02 on GE, DNA methylation, and the combined datasets respectively. In regard to precision and specificity, SA-RF obtained the best result of 100 on GE dataset. While SA-SVM attained the best recall result of 100 on GE dataset.

Je tiens tout d’abord à remercier mes encadrants. Après avoir grandemment contribué à me faire venir enseigner à l’Université, Eric Matzner-Løber m’a fait confiance pour ce travail de recherche; je tiens à lui témoigner toute ma reconnaissance et mes remerciements à son égard dépassent très largement le cadre professionnel. Arnaud Guyader a accepté sans hésitation de m’aider au moment même où les difficultés à venir s’annonçaient grandes. Je le remercie infiniment pour son écoute et sa très grande disponibilité; ce manuscrit lui doit beaucoup. Enfin, rendons à César ce qui appartient à César: l’idée de la méthode présentée ici est tout droit sortie de l’esprit prodigieux de Nicolas Hengartner. En m’invitant deux fois à Los Alamos ces dernières années, il m’a donné le privilège de travailler à ses côtés. Je suis désormais au moins sûr d’une chose: le génie et l’enthousiasme sont liés! Je tiens à remercier chaleureusement Cécile Durot et Sylvain Sardy pour l’intérêt qu’ils ont porté à...

Androgens play an important role in controlling the growth of the normal prostate gland and in the pathogenesis of benign prostate hyperplasia, and prostate cancer. Although testosterone is the main androgen secreted from the testes, dihydrotestosterone (DHT), a more potent androgen converted from testosterone by 5α-reductase isozymes, type I and II, is the major androgen in the prostate cells. The aim of this study is to investigate the cellular and molecular effects of dutasteride, a potent inhibitor of 5α-reductase type I and type II, in androgenresponsive (LNCaP) and androgen-unresponsive (DU145) human prostate cancer(PCa) cell lines. The expression pattern of 190 genes, selected on the basis of their proved or potential role in prostate cancerogenesis related to androgen signalling, were analysed using a low density home-made oligoarray (AndroChip 2). Our results show that dutasteride reduces cell viability and cell proliferation in both cell lines tested. AndroChip 2 gene signature identified in LNCaP a total of 11 genes differentially expressed (FC≥±1.5). Eight of these genes, were overexpressed and three were underexpressed. Overexpressed genes included genes encoding for proteins involved in biosynthesis and metabolism of androgen (HSD17B1; HSD17B3;CYP11B2), androgen receptor and androgen receptor co-regulators (AR;CCND1), and signal transduction(ERBB2; V-CAM; SOS1) whereas, underexpressed genes (KLK3; KLK2; DHCR24) were androgen-regulated genes (ARGs). No differentially expressed genes were scored in DU145. Microarray data were confirmed by quantitative real-time PCR assay (QRT-PCR). These data offer a selective genomic signature for dutasteride treatment in prostate epithelial cells and provide important insights in prostate cancer pathophysiology.

22q11 deletion syndrome (22q11DS) is a developmental anomaly caused by a microdeletion on human chromosome 22q11. Although mouse models indicate that Tbx1 is the gene responsible for the syndrome, the phenotypic spectrum of del22q11 patients is complex suggesting that gene-gene and gene-environment interactions are operative in delineating the pathogenesis of 22q11DS. In order to study the regulatory effects of 22q11 haploinsufficiency during development, the expression pattern of the orthologous MM16 genes was analysed in total embryos at different stages (from 4.5 dpc to 14.5 dpc; corresponding to pharyngeal development) by using a low-density oligonucleotide microarray (the "22q11DS-chip"). This microarray consists of 39 mouse genes orthologous to the 22q11 human ones and 29 mouse target genes selected on the basis of their potential involvement in biological pathways regarding 22q11 gene products. Expression level filtering and statistical analysis identified a set of genes that was consistently differentially expressed (FC N ± 2) during specific developmental stages. These genes show a similar profile in expression (overexpression or underexpression). Quantitative real-time PCR analyses showed an identical expression pattern to that found by microarrays. A bioinformatic screening of regulative sequence elements in the promoter region of these genes, revealed the existence of conserved transcription factor binding sites (TFBSs) in co-regulated genes which are functionally active at 4.5, 8.5 and 14.5 dpc. These data are likely to be helpful in studying developmental anomalies detected in del22q11 patients.

Cystic fibrosis (CF) is caused by mutations in the gene encoding the cystic fibrosis conductance transmembrane regulator (CFTR). Symptoms are pancreatic insufficiency, chronic obstructive lung disease, liver disease, chronic sinusitis and infertility in male patients. The phenotypic variability may be explained only in part by the more than 1200 CFTR mutations, which are grouped into six different classes, according to their effect on the protein ranging from a severe (no synthesis or blocked processing) to mild mutation (altered conductance or reduced synthesis). However, it is now accepted that other genes (CF modifiers) influence the phenotypic spectrum of the disease. In order to identify CF modifier genes, we built a low-density home-made oligoarray containing 144 genes selected according to biochemical criteria and eval-uated their expression in two CF bronchial epithelial cell lines (CuFi1 F508del/F508del; CuFi3 F508del/R553X). If we consider both cell lines, 38 genes (26.3%) show an altered expression pattern with a threshold >±1.5. Of these 38 genes, 12 are altered in CuFi1, and 26 in CuFi3. Some of these genes share the same expression pattern in both cell lines, while others have a different behaviour. These results were validated by a QRT-PCR assay (R 2 CuFi1=0.81 and R 2 CuFi3=0.91). These data could suggest that the presence of a class I allele (R553X) determines a more profound alteration of gene expression pattern than the presence of a class II allele (F508del). The identification of the genes altered by a specific CF mutation could lead to the development of a pharmacological approach specific for different CFTR genotypes.

In the classification of Mass Spectrometry (MS) proteomics data, peak detection, feature selection, and learning classifiers are critical to classification accuracy. To better understand which methods are more accurate when classifying data, some publicly available peak detection algorithms for Matrix assisted Laser Desorption Ionization Mass Spectrometry (MALDI-MS) data were recently compared; however, the issue of different feature selection methods and different classification models as they relate to classification performance has not been addressed. With the application of intelligent computing, much progress has been made in the development of feature selection methods and learning classifiers for the analysis of high-throughput biological data. The main objective of this paper is to compare the methods of feature selection and different learning classifiers when applied to MALDI-MS data and to provide a subsequent reference for the analysis of MS proteomics data. We compared a well-known method of feature selection, Support Vector Machine Recursive Feature Elimination (SVMRFE), and a recently developed method, Gradient based Leaveone-out Gene Selection (GLGS) that effectively performs microarray data analysis. We also compared several learning classifiers including K-Nearest Neighbor Classifier (KNNC), Naïve Bayes Classifier (NBC), Nearest Mean Scaled Classifier (NMSC), uncorrelated normal based quadratic Bayes Classifier recorded as UDC, Support Vector Machines, and a distance metric

This paper proposes an optimized classifier based on feature elimination (OCFE) for gene selection with combining two feature elimination methods, ReliefF and SVM-RFE. ReliefF algorithm is filter feature selection which rank the data by the importance of the data. SVM-RFE algorithm is a wrapper feature selection which wrapped the data and rank the data based on the weight of feature. With combining these two methods we get less error rate average, 0.3016138 for OCFE and 0.3096779 for SVM-RFE. The proposed method also get better accuracy with 70% for OCFE and 69% for SVM-RFE.

Selecting the most relevant factors from genetic profiles that can optimally characterize cellular states is of crucial importance in identifying complex disease genes and biomarkers for disease diagnosis and assessing drug efficiency. In this paper, we present an approach using a genetic algorithm for a feature subset selection problem that can be used in selecting the near optimum set of genes for classification of cancer data. In substantial improvement over existing methods, we classified cancer data with high accuracy with less features.

Motivation: Pre-selection of informative features for supervised classification is a crucial, albeit delicate, task. It is desirable that feature selection provides the features that contribute most to the classification task per se and which should therefore be used by any classifier later used to produce classification rules. In this article, a conceptually simple but computer-intensive approach to this task is proposed. The reliability of the approach rests on multiple construction of a tree classifier for many training sets randomly chosen from the original sample set, where samples in each training set consist of only a fraction of all of the observed features. Results: The resulting ranking of features may then be used to advantage for classification via a classifier of any type. The approach was validated using Golub et al. leukemia data and the Alizadeh et al. lymphoma data. Not surprisingly, we obtained a significantly different list of genes. Biological interpretation of t...

Microarray technology today has the ability of having the whole genome spotted on a single chip. It allows the biologist to inspect thousands of gene activities simultaneously. Machine learning approaches are suited and used to discovering the complex relationships between genes under controlled experimental conditions and classify microarray data by identifying a subset of informative genes embedded in a large data set that involves multiple classesand is infected with the high dimensionality noise. In this paper, a hybrid system integrates genetic algorithms and decision tree is proposed for genes expression analysis and prediction to their functionality for cancer classification. The learning capacity of decision trees used in the base learning systems is boosted by feature selection method. Experiments present preliminary results to demonstrate the capability of hybrid system to mine accurate classification rules for classifying prediction in comparable to traditional machine learning algorithms.

In gene expression microarray data analysis, selecting a small number of discriminative genes from thousands of genes is an important problem for accurate classification of diseases or phenotypes. The problem becomes particularly challenging due to the large number of features (genes) and small sample size. Traditional gene selection methods often select the top-ranked genes according to their individual discriminative power without handling the high degree of redundancy among the genes. Latest research shows that removing redundant genes among selected ones can achieve a better representation of the characteristics of the targeted phenotypes and lead to improved classification accuracy. Hence, we study in this paper the relationship between feature relevance and redundancy and propose an efficient method that can effectively remove redundant genes. The efficiency and effectiveness of our method in comparison with representative methods has been demonstrated through an empirical study using public microarray data sets.

Accuracy in quantitative real-time polymerase chain reaction (qPCR) requires the use of stable endogenous controls. Normalization with multiple reference genes is the gold standard, but their identification is a laborious task, especially in species with limited sequence information. Coffee (Coffea ssp.) is an important agricultural commodity and, due to its economic relevance, is the subject of increasing research in genetics and biotechnology, in which gene expression analysis is one of the most important fields. Notwithstanding, relatively few works have focused on the analysis of gene expression in coffee. Moreover, most of these works have used less accurate techniques such as northern blot assays instead of more accurate techniques (e.g., qPCR) that have already been extensively used in other plant species. Aiming to boost the use of qPCR in studies of gene expression in coffee, we uncovered reference genes to be used in a number of different experimental conditions. Using two distinct algorithms implemented by geNorm and Norm Finder, we evaluated a total of eight candidate reference genes (psaB, PP2A, AP47, S24, GAPDH, rpl39, UBQ10, and UBI9) in four different experimental sets (control versus drought-stressed leaves, control versus droughtstressed roots, leaves of three different coffee cultivars, and four different coffee organs). The most suitable combination of reference genes was indicated in each experimental set for use as internal control for reliable qPCR data normalization. This study also provides useful guidelines for reference gene selection for researchers working with coffee plant samples under conditions other than those tested here.

This paper presents an informative gene set selection approach to tumor diagnosis based on the Distance Sensitive Rival Penalized Competitive Learning (DSRPCL) algorithm and redundancy analysis. Since the DSRPCL algorithm can allocate an appropriate number of clusters for an input dataset automatically, we can utilize it to classify the genes (expressed by the gene expression levels of all the samples) into certain basic clusters. Then, we apply the post-filtering algorithm to each basic gene cluster to get the typical and independent informative genes. In this way we can obtain a compact set of informative genes. To test the effectiveness of the selected informative gene set, we utilize the support vector machine (SVM) to construct a tumor diagnosis system based on the express profiles of its genes. It is shown by the experiments that the proposed method can achieve a higher diagnosis accuracy with a smaller number of informative genes and less computational complexity in comparison with the previous ones.

Selection of significant genes via expression patterns is an important problem in microarray data processing. In this article, we propose and study a new method for selecting relevant genes obtained by spectral biclustering and based on similarity between genes and eigenvectors. The proposed algorithm can select a much smaller gene subset to make accurate predictions. The unsupervised gene selection method suggested in this paper is demonstrated on two microarray cancer data sets, i.e., the Lymphoma and the Liver cancer data sets. In both examples, our method is able to identify two-gene combinations which can lead to prediction with very high accuracy.

Selective Alzheimer's disease indicator-1 (seladin-1) is a novel gene with antiapoptotic activity that is down-regulated in vulnerable brain regions in Alzheimer's disease. This gene encodes 3-␤-hydroxysterol ⌬-24-reductase (DHCR24), which converts desmosterol into cholesterol. In the adrenal cortex, increased expression of seladin-1/DHCR24, which appears to be modulated by ACTH, has been recently reported in cortisolsecreting adenomas, compared with the adjacent atrophic tissue. In our study, we measured the expression level of seladin-1/DHCR24 in cortisol-(n ‫؍‬ 18) and aldosterone-secreting (n ‫؍‬ 16) adrenocortical adenomas, in carcinomas (n ‫؍‬ 17), and in normal adrenal glands (n ‫؍‬ 8) by quantitative real-time RT-PCR. The amount of seladin-1/DHCR24 mRNA was signifi-cantly reduced in carcinomas (total RNA, 2.5 ؎ 0.8 pg/g) compared with the other groups (P < 0.01). Western blot analysis confirmed the mRNA results. Similarly, in adrenal malignancies, significantly reduced levels of expression of the ACTH receptor gene were found. In the adrenal cancer cell line H295R and in primary cultures from adrenocortical cells, ACTH (1 nM) and forskolin (10 M) effectively increased seladin-1/DHCR24 expression, confirming that seladin-1/DHCR24 is modulated by the ACTH/cAMP-driven pathway. In summary, this is the first demonstration that seladin-1/DHCR24 expression is reduced in adrenal cancer, suggesting that it might be viewed as a new potential marker of adrenal malignancies. (

Selective Alzheimer's disease indicator-1 (seladin-1) is a novel gene with antiapoptotic activity that is down-regulated in vulnerable brain regions in Alzheimer's disease. This gene encodes 3-␤-hydroxysterol ⌬-24-reductase (DHCR24), which converts desmosterol into cholesterol. In the adrenal cortex, increased expression of seladin-1/DHCR24, which appears to be modulated by ACTH, has been recently reported in cortisolsecreting adenomas, compared with the adjacent atrophic tissue. In our study, we measured the expression level of seladin-1/DHCR24 in cortisol-(n ‫؍‬ 18) and aldosterone-secreting (n ‫؍‬ 16) adrenocortical adenomas, in carcinomas (n ‫؍‬ 17), and in normal adrenal glands (n ‫؍‬ 8) by quantitative real-time RT-PCR. The amount of seladin-1/DHCR24 mRNA was signifi-cantly reduced in carcinomas (total RNA, 2.5 ؎ 0.8 pg/g) compared with the other groups (P < 0.01). Western blot analysis confirmed the mRNA results. Similarly, in adrenal malignancies, significantly reduced levels of expression of the ACTH receptor gene were found. In the adrenal cancer cell line H295R and in primary cultures from adrenocortical cells, ACTH (1 nM) and forskolin (10 M) effectively increased seladin-1/DHCR24 expression, confirming that seladin-1/DHCR24 is modulated by the ACTH/cAMP-driven pathway. In summary, this is the first demonstration that seladin-1/DHCR24 expression is reduced in adrenal cancer, suggesting that it might be viewed as a new potential marker of adrenal malignancies. (

The unique characteristic of a biological system in nature is described by its DNA through the transcription of its genes. This is like a memory map that each cell of an organism contains. An artificial embryonic cell contains a similar memory map where the specific &#39;gene&#39; it executes determines the functionality of the cell. An electronic system is then constructed by a large number of identical cells where each possesses a different behaviour. The cells collectively however determine the characteristic of the target system. This paper proposes a new, variable size memory map based on a novel gene selection algorithm that no longer uses the hitherto common address decoding approach to access the cell&#39;s gene it will execute. Instead it applies the principle of cyclic metamorphic gene selection of the artificial DNA memory. A further benefit of the approach is that through genetic operators or variable memory space environment for enhanced behaviour the functionality of t...

Smoking remains a global health crisis, contributing to addiction and diverse diseases through complex biological mechanisms. This study explores the hypothesis that smoking induces epigenetic modifications and alters bidirectional neurobiological pathways, perpetuating addiction and disease progression. Leveraging a dataset of 55,692 individuals with 27 health metrics, we analyze associations between smoking status and physiological markers (e.g., lipid profiles, blood pressure, liver enzymes) to infer potential epigenetic and neurobiological mediators. Preliminary data reveal significant correlations between smoking and elevated triglycerides, LDL cholesterol, and liver function markers, suggesting systemic inflammation and oxidative stress as plausible intermediaries. We propose a methodology integrating bioinformatics and systems biology to map smoking-associated phenotypic changes to epigenetic loci (e.g., DNA methylation) and neural circuits (e.g., dopaminergic pathways). This work aims to bridge clinical observations with molecular mechanisms, offering insights into personalized interventions targeting smoking's epigenetic "fingerprints" and their neurobiological consequences.

The increasing dimensionality of gene expression data poses significant challenges in cancer classification, particularly in colon cancer. This study presents a novel filtering approach (FA) and a gene classifier (GC) to enhance gene selection and classification accuracy. Utilizing a dataset of 62 samples, our methods integrate statistical measures and machine learning classifiers, achieving classification accuracies of 96% and 97%, respectively. The FA effectively filters out noise and redundancy, allowing for accurate predictions with a minimal subset of genes, while the GC leverages multiple classifiers for optimal performance. These findings underscore the importance of robust feature selection in improving cancer diagnostics and suggest potential applications in personalized medicine. By addressing the limitations of existing methodologies, our work lays the groundwork for future research in cancer genomics, emphasizing the need for adaptive strategies to handle complex datasets.