Roland Linder

Microarray Data Classified by Artificial Neural Networks

Systems biology has enjoyed explosive growth in both the number of people participating in this area of research and the number of publications on the topic. The field of systems biology encompasses the in silico analysis of high-throughput data as provided by DNA or protein microarrays. Along with the increasing availability of microarray data, attention is focused on methods of analyzing the expression rates. One important type of analysis is the classification task, for example, distinguishing different types of cell functions or tumors. Recently, interest has been awakened toward artificial neural networks (ANN), which have many appealing characteristics such as an exceptional degree of accuracy. Nonlinear relationships or independence from certain assumptions regarding the data distribution are also considered. The current work reviews advantages as well as disadvantages of neural networks in the context of microarray analysis. Comparisons are drawn to alternative methods. Selected solutions are discussed, and finally algorithms for the effective combination of multiple ANNs are presented. The development of approaches to use ANN-processed microarray data applicable to run cell and tissue simulations may be slated for future investigation.

ACMD: A Practical Tool for Automatic Neural Net Based Learning

Although neural networks have many appealing properties, yet there is neither a systematic way how to set up the topology of a neural network nor how to determine its various learning parameters. Thus an expert is needed for fine tuning. If neural network applications should not be realisable only for publications but in real life, fine tuning must become unnecessary. We developed a tool called ACMD (Approximation and Classification of Medical Data) that is demonstrated to fulfil this demand. Moreover referring to six medical classification and approximation problems of the PROBEN1 benchmark collection this approach will be shown even to outperform fine tuned networks.

A new neural network approach classifies olfactory signals with high accuracy

Artificial neural networks (ANN) become more significant in signal processing. Because ANN still have some drawbacks we developed a new neural network tool named ACMD considering several methods of resolution, existing ones as well as new ones. In order to demonstrate the capabilities of ACMD in the field of food quality, we classified signals from an electronic nose smelling different types of edible oil and honey. The accuracies achieved by ACMD were evidently greater than the accuracies obtained by ANN trained by other well-known methods. As a conclusion it seems to be worthwhile considering sophisticated ANN strategies like those integrated in ACMD.

The capabilities of artificial neural networks in body composition research

Abstract.When estimating in vivo body composition or combining such estimates with other results, multiple variables must be taken into account (e. g. binary attributes such as gender or continuous attributes such as most biosignals). Standard statistical models, such as logistic regression and multivariate analysis, presume well-defined distributions (e. g. normal distribution); they also presume independence among all inputs and only linear relationships, yet rarely are these requirements met in real life. As an alternative to these models, artificial neural networks can be used. In the present work, we describe the pre-processing and multivariate analysis of data using neural network techniques, providing examples from the medical field and making comparisons with classic statistical approaches. We also address the criticisms raised regarding neural network techniques and discuss their potential improvement.

A study of subharmonics in connected speech material.

Subharmonics are often observed in running speech spectrograms that are difficult to quantify. This study investigates the relationship between rough voice quality and the presence-and amount-of subharmonics in connected speech material in a group of 35 male and 35 female speakers with voice pathology. Spectrum analysis was undertaken in 145 pathologic voices, of which 77 had subharmonics in connected speech. Only 34% of 70 subjects under investigation developed subharmonics in both connected speech and sustained phonations. All voices were judged as perceptually rough. The results of this study indicate that male and female voices with subharmonics do not differ in examined acoustic characteristics except for the modal fundamental frequency (F(0)). A small but significant difference in roughness scores between genders seems to be caused by the confounding perception of factors not related to acoustic measurements of F(0) and number of subharmonics. The degree of roughness has a significant relationship with the irregularity index, percentage of low F(0) estimates, and, to a lesser extent, power of subharmonics.

Artificial neural network analysis: a novel application for predicting site-specific bone mineral density

Abstract.Dual X-ray absorptiometry (DXA), which is the most commonly used method for the diagnosis and followup of human bone health, is known to produce accurate estimates of bone mineral density (BMD). However, high costs and problems with availability may prevent its use for mass screening. The objective of the present study was to estimate BMD values for healthy persons and those with conditions known to be associated with BMD, using artificial neural networks (ANN). An ANN was used to quantitatively estimate site-specific BMD values in comparison with reference values obtained by DXA (i. e. BMDspine, BMDpelvis, and BMDtotal). Anthropometric measurements (i. e. sex, age, weight, height, body mass index, waist-to-hip ratio, and the sum of four skinfold thicknesses) were fed to the ANN as independent input variables. The estimates based on four input variables were generated as output and were generally identical to the reference values for all studied groups. We believe the ANN is a promising approach for estimating and predicting site-specific BMD values using simple anthropometric measurements.

Text mining, a race against time? An attempt to quantify possible variations in text corpora of medical publications throughout the years

BACKGROUND The continuous growth of medical sciences literature indicates the need for automated text analysis. Scientific writing which is neither unitary, transcending social situation nor defined by a timeless idea is subject to constant change as it develops in response to evolving knowledge, aims at different goals, and embodies different assumptions about nature and communication. The objective of this study was to evaluate whether publication dates should be considered when performing text mining. METHODS A search of PUBMED for combined references to chemokine identifiers and particular cancer related terms was conducted to detect changes over the past 36 years. Text analyses were performed using freeware available from the World Wide Web. TOEFL Scores of territories hosting institutional affiliations as well as various readability indices were investigated. Further assessment was conducted using Principal Component Analysis. Laboratory examination was performed to evaluate the quality of attempts to extract content from the examined linguistic features. RESULTS The PUBMED search yielded a total of 14,420 abstracts (3,190,219 words). The range of findings in laboratory experimentation were coherent with the variability of the results described in the analyzed body of literature. Increased concurrence of chemokine identifiers together with cancer related terms was found at the abstract and sentence level, whereas complexity of sentences remained fairly stable. CONCLUSIONS The findings of the present study indicate that concurrent references to chemokines and cancer increased over time whereas text complexity remained stable.

MicroRNA target prediction: theory and practice.

Abstract The present study is one of the few that includes tissue samples in the evaluation of target prediction algorithms designed to detect microRNA (miRNA) sequences that might interact with particular messenger RNA (mRNA) sequences. Twelve different target prediction tools were used to find miRNA sequences that might interact with CCL20 gene expression. Different algorithms predicted controversial miRNA sequences for CCL20 regulation due to a different weighting of parameters. Hsa-miR-21 and hsa-miR-145 suggested by four or more programs were chosen for further investigation. Possible real interaction of these miRNA sequences with CCL20 gene expression was monitored using luciferase assays and expression analyses of tissue samples of colorectal adenocarcinoma by either qRT-PCR or ELISA. Folding status of seed-binding sites in complete mRNA and 3′UTR of CCL20 was predicted. Prediction of miRNA expression was attempted based on CCL20 expression data. Eight of the target prediction tools forecasted a role for hsa-miR-21 and four mentioned hsa-miR-145 in CCL20 gene regulation. Laboratory experimentation showed that CCL20 may serve as a target of hsa-miR-21 but not hsa-miR-145. Expression of the molecules resulted in no clear assertion. Folding of seed-binding sites was predicted to be relatively constant for the complete mRNA and 3′UTR. Predicting miRNA expression based on target gene expression was impossible. This might be attributable to the fact that effects of miRNA activity may oscillate between gene product repression and activation. Additional systematic studies are needed to address this issue.

Virtual tissue. Quaoaring

Virtual tissue can be generated by employing various methods. First steps en route to virtual tissue may encompass the generation of virtual cells. One such approach termed Quaoaring was applied to produce artificial erythrocytes and these were both discocyte and echinocyte in shape. The results were subsequently compared with data gleaned from scanning electron microscopy and atomic force microscopy. Quaoaring has, however, proved to be unsuccessful in creating convincing objects, particularly those which should be echinocytic in appearance.ZusammenfassungEs gibt mannigfaltige Wege, virtuelles Gewebe herzustellen. Erste Schritte dorthin können über die Herstellung virtueller Zellen führen. Eine „Quaoaring“ genannte Methode wurde genutzt, um artifizielle Erythrozyten sowohl der Diskozyten- als auch der Echinozytenform zu erzeugen. Die Ergebnisse wurden mit Daten aus der Rasterelektronenmikroskopie und der „Atomic-Force-Mikroskopie“ verglichen. Quaoaring konnte hierbei insbesondere bei Objekten, die eine Echinozytenform haben sollten, nicht überzeugen.AbstractVirtual tissue can be generated by employing various methods. First steps en route to virtual tissue may encompass the generation of virtual cells. One such approach termed Quaoaring was applied to produce artificial erythrocytes and these were both discocyte and echinocyte in shape. The results were subsequently compared with data gleaned from scanning electron microscopy and atomic force microscopy. Quaoaring has, however, proved to be unsuccessful in creating convincing objects, particularly those which should be echinocytic in appearance.

Approximation dreidimensionaler Oberflächenmodelle der Lippen-Kiefer-Gaumen-Region durch aktive Polygonnetze

In dieser Arbeit wird ein Verfahren zur Rekonstruktion eines dreidimensionalen Oberflachenmodells aus Serienschnitten vorgestellt, die sich von Lippen-Kiefer-Gaumenspalten (kurz LKG) ableiten und insbesondere die in der chirurgischen Behandlung relevanten medizinischen Strukturen berucksichtigen. Die Intention ist, ein aktives Polygonnetz (deformierbares Oberflachennetz) zur Rekonstruktion der Struktur zu verwenden. Initial wird ein Netzprototyp als grobe Approximation der Struktur erstellt und dann mithilfe eines Gradient Vector Flow (GVF)-Kraftfeldes deformiert. Inharenter Aspekt der Deformation ist die Beachtung eines resultierenden uberschneidungsfreien Oberflachennetzes. Die rekonstruierte Oberflache eignet sich aufgrund ausgezeichneter Eigenschaften als Eingabe fur eine Kraftesimulation mittels Finite-Elemente-Methoden.