SEARCH · PubMed Health
Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Sequence comparison, a vital research tool in computational biology, is based on a simple O(n2) algorithm that easily maps to a linear array of processors. This paper reviews and compares high-performance sequence analysis on general-purpose supercomputers and single-purpose reconfigurable, and programmable co-processors. The difficulty of comparing hardware from published performance figures is also noted.
Relative potency is a measure that has been used for many years to summarize the comparison of dose-response curves in parallel line bioassays. When response curves for two preparations are not parallel the traditional definition of relative potency no longer applies. We review the concept of relative potency and show that, in some situations, it can be given meaning for non-parallel curves as the ratio of biological activity in full strength assay preparations. Under an assumption that non-parallel curves result from the competition of mixtures of antigens for receptor binding sites, estimation of relative potency for non-parallel curves can be accomplished. We show that estimation of models for both parallel curve and response attenuation situations may be accomplished within the framework of generalized linear models. This estimation depends on the ability to deal with non-linear parameters appearing in the link function, and an iterative algorithm depending on direct parameter updates is outlined. The topics discussed are illustrated with the analysis of data from two immunoassays conducted with veterinary vaccines. The models developed here depend in an essential way on the assumption of response attenuation by competing antigens. Our methods may not be appropriate for non-parallel curves caused by other phenomena.
We propose a new approach for fabricating more sophisticated combinatorial chemistry libraries via split synthesis and evaluate its potential through extensive simulation. Our algorithmically intensive method promises to reduce the time and materials costs of synthesizing libraries which are (1) too large to synthesize economically by sequential or parallel synthesis, (2) too long or irregular for conventional split synthesis generation techniques, and (3) not used in sufficient quantity to justify the setup costs of array makers. It also encourages the design of more focused and interesting libraries than are typically constructed using split synthesis. Our algorithms automate the design of efficient synthesis procedures for motif-based libraries which are too complex to design by hand. Our software allows the user to select the most desirable tradeoff between minimizing the number of steps in the synthesis process and containing the combinatorial explosion of the number of compounds synthesized.
Methods for estimating required sample size for comparing two population means have been published. Most involve the use of complicated formulae and tables. These methods are limited to comparing two groups. Although techniques exist to determine sample sizes for comparing more than two groups, they are intrinsically far more complicated. A simple linear nomogram is proposed as a solution to these problems, and its use is illustrated with examples of parallel group, ordered parallel group and factorial designs.
In this work, we report the development of a novel QSAR technique combining genetic algorithms and neural networks for selecting a subset of relevant descriptors and building the optimal neural network architecture for QSAR studies. This technique uses a neural network to map the dependent property of interest with the descriptors preselected by the genetic algorithm. This technique differs from other variable selection techniques combining genetic algorithms to neural networks by two main features: (1) The variable selection search performed by the genetic algorithm is not constrained to a defined number of descriptors. (2) The optimal neural network architecture is explored in parallel with the variable selection by dynamically modifying the size of the hidden layer. By using both artificial data and real biological data, we show that this technique can be used to build both classification and regression models and outperforms simpler variable selection techniques mainly for nonlinear data sets. The results obtained on real data are compared to previous work using other modeling techniques. We also discuss some important issues in building QSAR models and good practices for QSAR studies.
We study asymmetric stochastic networks from two points of view: combinatorial optimization and learning algorithms based on relative entropy minimization. We show that there are non trivial classes of asymmetric networks which admit a Lyapunov function L under deterministic parallel evolution and prove that the stochastic augmentation of such networks amounts to a stochastic search for global minima of L. The problem of minimizing L for a totally antisymmetric parallel network is shown to be associated to an NP-complete decision problem. The study of entropic learning for general asymmetric networks, performed in the non equilibrium, time dependent formalism, leads to a Hebbian rule based on time averages over the past history of the system. The general algorithm for asymmetric networks is tested on a feed-forward architecture.
This paper surveys the computational strategies followed to parallelise the most used software in the bioinformatics arena. The studied algorithms are computationally expensive and their computational patterns range from regular, such as database-searching applications, to very irregularly structured patterns (phylogenetic trees). Fine- and coarse-grained parallel strategies are discussed for these very diverse sets of applications. This overview outlines computational issues related to parallelism, physical machine models, parallel programming approaches and scheduling strategies for a broad range of computer architectures. In particular, it deals with shared, distributed and shared/distributed memory architectures.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
A two-dimensional pencil-beam algorithm is presented for the calculation of arc electron dose distributions in any plane that is perpendicular to the axis of rotation. The dose distributions are calculated by modelling the arced beam as a single broad beam defined by the irradiated surface of the patient. The algorithm is two-dimensional in that the anatomical cross section of the patient and the skin collimators are assumed identical in parallel planes outside the plane of calculation. The broad beam is modelled as a collection of strip beams, each strip beam being characterised by its planar fluence, mean projected angular direction and a root-mean-square spread about the mean direction. Using these parameters, the dose distribution is calculated using pencil-beam theory. Examples of strip-beam parameters and resulting dose distributions for patient geometries are presented. Features of the algorithm, which include (1) incorporation of pencil-beam theory for the calculation of dose in heterogeneous tissue, (2) run times of only about twice that of comparable-sized fixed electron fields and (3) the input requirement of only a single depth dose and four off-axis dose profiles of measured data, make the algorithm practical for clinical use.
These studies examine the distribution of single-cell responses in basophil preparations in the context of four events that may be associated with early activation by anti-immunoglobulin E (IgE) antibody and the bacterial peptide fMet-Leu-Phe (fMLP). In general, we measured the single-cell response distributions after challenge with a concentration of stimulus that resulted in an optimal response and compared this with the distribution that occurred after challenge with suboptimal concentrations of the same stimulus. The elevation in cytosolic calcium, as detected in Fura-2-labeled basophils, after challenge with anti-IgE or fMLP showed graded characteristics in that the distributions were unimodal under conditions of optimal or suboptimal challenge with little skewing from a normal distribution. Similarly, the up-regulation of the cell surface adhesion molecule CD11b, as determined by flow cytometry, showed graded unimodal increases after challenge with anti-IgE antibody at optimal and suboptimal concentrations. In addition, stimulation of basophils led to increased F-actin polymerization. After challenge with an optimal concentration of anti-IgE antibody, the F-actin content of basophils increased to a maximum between 10 and 15 min and returned to near prechallenge levels by 60 min. There was a close correlation between the maximum increase in F-actin content and histamine release regardless of the stimulus; anti-IgE antibody, fMLP, and phorbol ester (PMA) responses lay on the same regression line. The single-cell F-actin polymerization distributions were also unimodal and graded according to the magnitude of the histamine release response. During measurements of the calcium response under the microscope we noted that basophils underwent significant changes in morphology after challenge with any stimulus. These changes were related to both degranulation and nondegranulation events and could be quantitated by a series of image-processing algorithms, which are presented. The kinetics of the morphological change, measured as a change in cell perimeter, paralleled degranulation. Single-cell distributions of the morphologic changes were also unimodal under conditions of both optimal and suboptimal stimulation. Therefore, no evidence of all-or-nothing responses could be observed in the context of these four early activation events. In general, the response distributions resembled normal distributions at both optimal and suboptimal levels of stimulation, which indicated that single basophils responded in a graded manner.
Four groups of mdx mice (deflazacort, high dose of 1.5 mg/kg and low dose of 0.75 mg/kg; prednisone, 1.0 mg/kg; and a placebo) were examined in a double-blind protocol. The experiments tested the hypothesis that infrared spectroscopy can distinguish among gastrocnemius muscle tissues derived from dystrophic animals (n = 22) from different treatment groups and from control muscle tissue (n = 23). Results showed that muscle, inflamed muscle, and tendon can be distinguished on the basis of their infrared absorption patterns. Distinctions among the spectra of the four treatment groups were sought with automated pattern-recognition methods. These classification methods, based either on spectral regions (900-1,500 cm-1) or on principal-component analysis, were in close agreement, assigning 15 or 16, respectively, of 22 mdx spectra to the correct treatment group. Both trials cleanly separated the high-dose deflazacort from the placebo group of muscles, whereas the prednisone and low-dose deflazacort groups were persistently confused in these classifications. Changes in the histology of muscle inflammation paralleled the spectral-classification results. Thus the proposed method, combining infrared spectroscopy with pattern-recognition algorithms, can distinguish treatment effects on muscle tissue. Specific spectral features characteristic of tissue type, disease progression, and treatment effects are not yet elucidated.
We performed high-resolution CT scans on 30 wrists in 27 patients with either previous surgical intercarpal fusion or known or suspected scaphoid abnormalities. Most examinations used the same protocol: contiguous 1.5-mm axial sections parallel to the long axis of the body of the scaphoid, small reconstruction circle, and bone reconstruction algorithm. Fifteen patients were evaluated for union of previous surgical carpal fusions; eight patients were evaluated for healing of scaphoid fractures; and the other four patients had miscellaneous disorders, including a midcarpal dislocation with a scaphoid fracture. High-resolution CT clearly showed whether the carpal fusions were united. In addition, metallic fixation devices were easy to locate and did not significantly degrade image quality. Osseous union of healing scaphoid fractures was reliably assessed. We conclude that high-resolution CT of the wrist is a useful method for evaluating surgical carpal fusions and various disorders of the scaphoid.
At present, the exchange of information between clinicians and pathologists are limited to mail, telefax and telephone. The bi-directional video communication using at least two digitized phone lines (ISDN) is an additional medium for video-conference and for image exchange. The MECOM system (Medical Communication) has been set up in the department of pathology of Aixla-Chapelle Hospital (Germany) and has routinely used for two years between 4 local hospitals. Now it is mainly used for staff discussions between pathologists and practitioners involved in different oncological fields. Mecom can transmit any kind of high resolution images (x rays films, scanner images, histologic slides...) which are useful or necessary during medical staff discussions. This system requires only two phone lines to be operative. However it can turn on up to twelve parallel ISDN lines which allow to transmit 768 kbytes/sec. It manages the SG3, HVQ and H320 compression algorithms and allows the transmission of still images at a resolution varying from 256/240 to 512/480 pixels. Mecom is a modulable system which adapts to the needs. It has the following advantages. Emergency inter-disciplinary conferences. Real time images, discussion between a pathologist and an expert to confirm histo-pathological diagnosis. Abolition of distances. That is specially useful or necessary in countries with low medical population and poor transportation network.
One of the benefits of imaging neuronal activity is the capability of resolving spatial patterns in the x-y plane. With optical sectioning microscopy, the 3-dimensional (3-D) structure may also be studied without physical deformation by serially moving the focal plane of the microscope through the volume of interest along the focal axis. However, each image is blurred by contributions from neighboring planes. This degradation is most severe for low numerical aperture lenses and large amounts of defocus. In this study, an image restoration method using the optical properties of an aberration-free, defocused optical system has been developed for improving optical signals from voltage-sensitive dyes. Deblurring based on the optical transfer function (OTF) of the system was applied on two test sets of serially sectioned images: (1) fluorescent beads and (2) in vivo rat cerebellar cortex stained with the voltage-sensitive dye RH795. This method was shown to reduce significantly the out-of-focus contribution to the images, improving the spatial resolution not only in the x-y plane, but also the z axis. The algorithms were then applied to optical signals obtained by stimulation of the cerebellar surface. Optical signals having a distinct beam-like pattern were evoked and recorded over depths ranging from 0 to 300 microns prior to deblurring. Application of the deblurring algorithm reduced the depth of cerebellar cortex over which the optical signals were observed. In agreement, field potential recordings of the evoked parallel fiber volley and post-synaptic components were restricted to a narrow range of depths similar to the deblurred optical images. Removal of out-of-focus information is an essential step in the serial sectioning of central nervous system structures for neuronal imaging and 3-D reconstruction.
The aim of this work was to investigate the accuracy of dose predicted by a Batho power law correction, and two models which account for electron range: A superposition/convolution algorithm and a Monte Carlo algorithm. The results of these models were compared in phantoms with cavities and low-density inhomogeneities. An idealized geometry was considered with inhomogeneities represented by regions of air and lung equivalent material. Measurements were performed with a parallel plate ionization chamber, thin TLDs (thermoluminescent dosimeters) and film. Dose calculations were done with a generalized Batho model, the Pinnacle collapsed cone convolution model (CCC), and the Peregrine Monte Carlo dose calculation algorithm. Absolute central axis and off axis dose data at various depths relative to interfaces of inhomogeneities were compared. Our results confirm that for a Batho correction, dose errors in the calculated depth dose arise from the neglect of electron transport. This effect increases as the field size decreases, as the density of the inhomogeneity decreases, and with the energy of incident photons. The CCC calculations were closer to measurements than the Batho model, but significant discrepancies remain. Monte Carlo results agree with measurements within the measurement and computational uncertainties.
OBJECTIVE: Left-ventricular ejection fraction (LVEF) can be computed from gated myocardial perfusion SPECT studies using quantitative algorithms. The purpose of this study was to compare the LVEF obtained using the conventional high-resolution parallel-hole collimator (HRC) to the Cardiofocal collimator (CFC) (Siemens Medical Systems, Hoffman Estates, IL) using a quantitative LVEF program. METHODS: Thirty-four patients (15 men, 19 women; mean age = 62 y) had either treadmill or pharmacological stress testing with 25-30 mCi 99mTc sestamibi injected at peak stress. Conventional gated SPECT imaging was performed approximately 30 min poststress, first with the HRC collimator, then with the CFC, using the same acquisition parameters on a single-head gamma camera. Traditional (TRAD) determination of LVEF using planar gated blood pool and/or cardiac catherization also was obtained for each patient. RESULTS: The correlation in LVEF between the CFC and HRC acquisitions was excellent, r = 0.99. The correlation between CFC and TRAD LVEF was good, r = 0.95, as was the HRC and TRAD correlation, r = 0.97. The mean LVEF value for HRC was slightly less than TRAD (54% vs. 55.4%), while the CFC mean LVEF was higher (62% vs. 55.4%). Although CFC LVEF correlated well with HRC, mean LVEF value using CFC was higher than HRC. CONCLUSION: The choice of collimator may alter the LVEF obtained from gated SPECT perfusion studies.
Genetic linkage calculations can be time consuming, even on a fast computer. The ability to collect large family pedigrees has increased the magnitude of linkage computations. Sequential genetic algorithms have many successful applications in very different domains, but they have a main drawback in their utilization. Evaluations are very time-consuming, e.g., a pedigree consisting of 55 nodes takes about 70 min on a DEC-Alpha processor and about 270 min on a 166 MHz Pentium for certain likelihood calculations. This time increases exponentially with the increase in the size of the pedigree. In order to solve these shortcomings and to study new models of higher efficiency and efficacy, parallel platforms are being used for genetic programs. LINKAGE is a software package for performing genetic likelihood calculations; FASTLINK is an improved, faster version of it. This paper provides a parallel implementation of the "Linkmap" program (one of the four programs in LINKAGE/FASTLINK) for a heterogeneous environment, using a static and a dynamic strategy for task allocation. It was found that the increased performance by the dynamic strategy was close to the estimated maximum speed up.