Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Gaussian mixture model”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

265 records · Page 15Linked to original sources
PubMed2001

Performances of holographic gratings monitored by laser-induced phase separation in liquid mixtures.

We theoretically describe and experimentally explore the kinetics of holographic grating formation resulting from different laser-induced phase separation mechanisms. Our method makes use of two interfering c.w. laser waves to quench binary mixtures in composition, and to optically trap the nucleated domains on the fringes. Essentially, two different processes can lead to these variations in concentration: electrostriction and thermodiffusion. The former originates from induced dipolar couplings in a field gradient; as photopolymerization, this is a local process which is essentially sensitive to the q=q(0) Fourier mode forced by the fringe modulation. The latter corresponds to a variation in composition driven by a small thermal gradient; as solvent evaporation and thermal heating techniques, it is nonlocal and behaves as 1/q(2) because of its dissipative origin. By making experiments in both cases, we show that this q dependence on excitation has a strong influence on the performance of holographic gratings. While in the first case reflectivity saturates because the phase transition is confined by the fringes which behave as separated optical boxes with "soft walls" which calibrate the droplet size, blurring is expected for fringe-trapped domains induced by a nonlocal phase transition because the transition is governed by the Gaussian shape of the pump beams, and nucleated domains can reach a much larger size than the fringe spacing. The good agreement observed with our general model clearly illustrates how to make the difference between local and nonlocal excitations, and offers a first step towards a unified description of holographic grating formation monitored by phase transitions.

Biophysics↗
PubMed1997

The phase behavior of lipid monolayers containing pulmonary surfactant protein C studied by fluorescence light microscopy.

Three compounds of the pulmonary surfactant--dipalmitoylphosphatidylcholine (DPPC), dipalmitoylphosphatidylglycerol (DPPG), and the surfactant associated protein C (SP-C)--were spread at the air-water interface of a Langmuir trough as a model system to mimic the properties of natural surfactant. Fluorescence microscopical images of the film formed at the interface were obtained during compression using a fluorescence dye bound covalently either to phosphatidylcholine or to SP-C. The images were quantified using statistical methods in respect to relative areas and relative fluorescence intensities of the domains found. In the early stage of compression, film pressure rose slightly and was accompanied by a phase separation which could be recognized in the images by the formation of bright and dark domains. On further compression, after a steep increase of film pressure, a plateau region of constant film pressure started abruptly. During compression in the plateau region, fluorescence intensity of the bright domain formed in the early stage of compression increased. The increasing fluorescence intensity, the non-Gaussian intensity distribution of the bright domain, and the small mean molecular area of the film in the plateau region gave rise to the assumption that multilayer structures were formed in the late stage of compression. The formation of the multilayer structures was fully reversible in repeated compression-expansion cycles including the plateau region of the phase diagram. The ability of lipid/SP-C mixtures to form reversible multilayer structures during compression may be relevant to stability in lungs during expiration and inhalation.

1,2-Dipalmitoylphosphatidylcholine↗
PubMed2001

Characterization of the radical product formed from the reaction of nitric oxide with the spin trap 3,5-dibromo-4-nitrosobenzene sulfonate.

Previously, 3,5-dibromo-4-nitrosobenzene sulfonate (DBNBS) has been used in combination with electron paramagnetic resonance (EPR) spectrometry to trap nitric oxide (NO(*)). The reaction between DBNBS and NO(*) yields a radical product which gives rise to an EPR signal consisting of three lines with an A(N) = 0.96 mT, but the structure of this product is unknown. A two-stage high-performance liquid chromatography fractionation was performed to isolate the radical product from the other components in the DBNBS/NO(*) reaction mixture. The fractions containing the radical product were identified by the presence of the three-line EPR signal, and then these fractions were analyzed by negative ion fast atom bombardment-mass spectrometry (FAB-MS). Collectively, the FAB-MS data suggested that the radical product is the monosodium electrostatic complex with the dianion, bis(2,6-dibromo-4-sulfophenyl) nitroxyl. Analysis of the Gaussian and Lorentzian linewidths of the EPR signal suggested that bis(2,6-dibromo-4-sulfophenyl) nitroxyl molecules may group together to form micelles. Further studies also indicated that significant amounts of nitrogen and nitrate were produced during the reaction between DBNBS and NO(*). A reaction scheme consistent with these results is presented.

Benzenesulfonates↗
PubMed2002

A framework for ML estimation of parameters of (mixtures of) common reaction time distributions given optional truncation or censoring.

We present a framework for distributional reaction time (RT) analysis, based on maximum likelihood (ML) estimation. Given certain information relating to chosen distribution functions, one can estimate the parameters of these distributions and of finite mixtures of these distributions. In addition, left and/or right censoring or truncation may be imposed. Censoring and truncation are useful methods by which to accommodate outlying observations, which are a pervasive problem in RT research. We consider five RT distributions: the Weibull, the ex-Gaussian, the gamma, the log-normal, and the Wald. We employ quasi-Newton optimization to obtain ML estimates. Multicase distributional analyses can be carried out, which enable one to conduct detailed (across or within subjects) comparisons of RT data by means of loglikelihood difference tests. Parameters may be freely estimated, estimated subject to boundary constraints, constrained to be equal (within or over cases), or fixed. To demonstrate the feasibility of ML estimation and to illustrate some of the possibilities offered by the present approach, we present three small simulation studies. In addition, we present three illustrative analyses of real data.

Humans↗
PubMed1989

Fluorescence lifetime distributions of diphenylhexatriene-labeled phosphatidylcholine as a tool for the study of phospholipid-cholesterol interactions.

Fluorescence lifetimes of 1-palmitoyl-2-diphenylhexatrienylpro-pionyl-phosphatidylc hol ine in vesicles of palmitoyloleoyl phosphatidylcholine (POPC) (1:300, mol/mol) in the liquid crystalline state were determined by multifrequency phase fluorometry. On the basis of statistic criteria (chi 2red) the measured phase angles and demodulation factors were equally well fitted to unimodal Lorentzian, Gaussian, or uniform lifetime distributions. No improvement in chi 2red could be observed if the experimental data were fitted to bimodal Lorentzian distributions or a double exponential decay. The unimodal Lorentzian lifetime distribution was characterized by a lifetime center of 6.87 ns and a full width at half maximum of 0.57 ns. Increasing amounts of cholesterol in the phospholipid vesicles (0-50 mol% relative to POPC) led to a slight increase of the lifetime center (7.58 ns at 50 mol% sterol) and reduced significantly the distributional width (0.14 ns at 50 mol% sterol). Lifetime distributions of POPC-cholesterol mixtures containing greater than 20 mol% sterol were within the resolution limit and could not be distinguished from monoexponential decays on the basis of chi 2red. Cholesterol stabilizes and rigidifies phospholipid bilayers in the fluid state. Considering its effect on lifetime distributions of fluorescent phospholipids it may also act as a membrane homogenizer.

Cholesterol↗
PubMed2005

Mechanisms of solvolyses of acid chlorides and chloroformates. Chloroacetyl and phenylacetyl chloride as similarity models.

[reaction: see text] Rate constants and product selectivities (S = ([ester product]/[acid product]) x ([water]/[alcohol solvent]) are reported for solvolyses of chloroacetyl chloride (3) at -10 degrees C and phenylacetyl chloride (4) at 0 degrees C in ethanol/ and methanol/water mixtures. Additional kinetic data are reported for solvolyses in acetone/water, 2,2,2-trifluoroethanol(TFE)/water, and TFE/ethanol mixtures. Selectivities and solvent effects for 3, including the kinetic solvent isotope effect (KSIE) of 2.18 for methanol, are similar to those for solvolyses of p-nitrobenzoyl chloride (1, Z = NO(2)); rate constants in acetone/water are consistent with a third-order mechanism, and rates and products in ethanol/ and methanol/water mixtures can be explained quantitatively by competing third-order mechanisms in which one molecule of solvent (alcohol or water) acts as a nucleophile and another acts as a general base (an addition/elimination reaction channel). Selectivities increase for 3 as water is added to alcohol. Solvent effects on rate constants for solvolyses of 3 are very similar to those of methyl chloroformate, but acetyl chloride shows a lower KSIE, and a higher sensitivity to solvent-ionizing power, explained by a change to an S(N)2/S(N)1 (ionization) reaction channel. Solvolyses of 4 undergo a change from the addition/elimination channel in ethanol to the ionization channel in aqueous ethanol (<80% v/v alcohol). The reasons for change in reaction channels are discussed in terms of the gas-phase stabilities of acylium ions, calculated using Gaussian 03 (HF/6-31G(d), B3LYP/6-31G(d), and B3LYP/6-311G(d,p) MO theory).

Journal Article↗
PubMed2005

Robust multi-scale clustering of large DNA microarray datasets with the consensus algorithm.

MOTIVATION: Hierarchical and relocation clustering (e.g. K-means and self-organizing maps) have been successful tools in the display and analysis of whole genome DNA microarray expression data. However, the results of hierarchical clustering are sensitive to outliers, and most relocation methods give results which are dependent on the initialization of the algorithm. Therefore, it is difficult to assess the significance of the results. We have developed a consensus clustering algorithm, where the final result is averaged over multiple clustering runs, giving a robust and reproducible clustering, capable of capturing small signal variations. The algorithm preserves valuable properties of hierarchical clustering, which is useful for visualization and interpretation of the results. RESULTS: We show for the first time that one can take advantage of multiple clustering runs in DNA microarray analysis by collecting re-occurring clustering patterns in a co-occurrence matrix. The results show that consensus clustering obtained from clustering multiple times with Variational Bayes Mixtures of Gaussians or K-means significantly reduces the classification error rate for a simulated dataset. The method is flexible and it is possible to find consensus clusters from different clustering algorithms. Thus, the algorithm can be used as a framework to test in a quantitative manner the homogeneity of different clustering algorithms. We compare the method with a number of state-of-the-art clustering methods. It is shown that the method is robust and gives low classification error rates for a realistic, simulated dataset. The algorithm is also demonstrated for real datasets. It is shown that more biological meaningful transcriptional patterns can be found without conservative statistical or fold-change exclusion of data. AVAILABILITY: Matlab source code for the clustering algorithm ClusterLustre, and the simulated dataset for testing are available upon request from T.G. and O.W.

Algorithms↗
PubMed1995

Permeability of dimyristoyl phosphatidylcholine/dipalmitoyl phosphatidylcholine bilayer membranes with coexisting gel and liquid-crystalline phases.

The passive permeation of glucose and a small zwitterionic molecule, methyl-phosphoethanolamine, across two-component phospholipid bilayers (dimyristoyl phosphatidylcholine (DMPC)/dipalmitoyl phosphatidylcholine (DPPC) mixtures) exhibit a maximum when gel domains and fluid domains coexist. The permeability data of the two-phase bilayers cannot be fitted to single-rate kinetics, but are consistent with a Gaussian distribution of rate constants. In pure DMPC and DPPC as well as in their mixtures, at the temperature of the maximum excess heat capacity, the logarithm of the average permeability rate constants are linearly correlated with the mole fraction of DPPC in the total system. In addition, in the 50:50 mixture, the excess heat capacity values as well as the apparent fractions of interfacial lipid correlate with the logarithm of the excess permeabilities in the two-phase region. These results suggest that small polar molecules can cross the membrane at the interface between gel and fluid domains at a much faster rate than through the homogeneous phases; the acyl chains located at the domain interface experience lateral density fluctuations that are inversely proportional to their average length, and large enough to allow rapid transmembrane diffusion of the solute molecules. The distribution of the permeability rate constants may reflect temporal and spatial fluctuations of the lipid composition at the phase boundaries.

1,2-Dipalmitoylphosphatidylcholine↗
PubMed2003

Electrophoretic signal comparison applied to mRNA differential display analysis.

Gene expression analysis by electrophoretic methods is currently limited by the labor-intensive visual evaluation of the electrophoretic signal profiles. For this purpose, we present a flexible approach to computer-assisted comparison of quantitative electrophoretic patterns between multiple expression signals. Gaussian curves are first fitted to the complex peak mixtures, and the resulting approximate signals are then aligned and compared on a peak-by-peak basis with respect to specific patterns defined by the investigator. The rationale of the method is to produce a compressed list of exceptional expression patterns quantified by a set of associated numeric features. A score value is attached to each pattern in such a way that large values identify the most potential findings to be focused on in visual analysis instead of the vast amount of original electrophoretic results. The validity of the method is demonstrated by analyzing a large set of electrophoretic data from mRNA differential display experiments monitoring changes in gene expression patterns in human colonic carcinoma. The automated identification of variously defined gene expression patterns agrees well with the visual evaluation of the same electropherograms. The general comparison approach may also be found useful with other gene expression profiling instruments.

Adenocarcinoma↗
PubMed2005

A normalization strategy applied to HiCEP (an AFLP-based expression profiling) analysis: toward the strict alignment of valid fragments across electrophoretic patterns.

BACKGROUND: Gene expression analysis based on comparison of electrophoretic patterns is strongly dependent on the accuracy of DNA fragment sizing. The current normalization strategy based on molecular weight markers has limited accuracy because marker peaks are often masked by intense peaks nearby. Cumulative errors in fragment lengths cause problems in the alignment of same-length fragments across different electropherograms, especially for small fragments (< 100 bp). For accurate comparison of electrophoretic patterns, further inspection and normalization of electrophoretic data after fragment sizing by conventional strategies is needed. RESULTS: Here we describe a method for the normalization of a set of time-course electrophoretic data to be compared. The method uses Gaussian curves fitted to the complex peak mixtures in each electropherogram. It searches for target ranges for which patterns are dissimilar to the other patterns (called "dissimilar ranges") and for references (a kind of mean or typical pattern) in the set of resultant approximate patterns. It then constructs the optimal normalized pattern whose correlation coefficient against the reference in the range achieves the highest value among various combinations of candidates. We applied the procedure to time-course electrophoretic data produced by HiCEP, an AFLP-based expression profiling method which can detect a slight expression change in DNA fragments. We obtained dissimilar ranges whose electrophoretic patterns were obviously different from the reference and as expected, most of the fragments in the detected ranges were short (< 100 bp). The normalized electrophoretic patterns also agreed well with reference patterns. CONCLUSION: The normalization strategy presented here demonstrates the importance of pre-processing before electrophoretic signal comparison, and we anticipate its usefulness especially for temporal expression analysis by the electrophoretic method.

Algorithms↗
PubMed2004

Optimization of an independent component analysis approach for artifact identification and removal in magnetoencephalographic signals.

OBJECTIVE: To propose a noise reduction procedure for magnetoencephalography (MEG) signals introducing an automatic detection system of artifactual components (ICs) separated by an independent component analysis (ICA) algorithm, and a control cycle on reconstructed cleaned data to recovery part of non-artifactual signals possibly lost by the blind mechanism. METHODS: The procedure consisted of three main steps: (1) ICA for blind source separation (BSS); (2) automatic detection method of artifactual components, based on statistical and spectral ICs characteristics; (3) control cycle on 'discrepancy,' i.e. on the difference between original data and those reconstructed using only ICs automatically retained. Simulated data were generated as representative mixtures of some common brain frequencies, a source of internal Gaussian noise, power line interference, and two real artifacts (electrocardiogram=ECG, electrooculogram=EOG), with the adjunction of a matrix of Gaussian external noise. Three real data samples were chosen as representative of spontaneous noisy MEG data. RESULTS: In simulated data the proposed set of markers selected three components corresponding to ECG, EOG and the Gaussian internal noise; in real-data examples, the automatic detection system showed a satisfactory performance in detecting artifactual ICs. 'Discrepancy' control cycle was redundant in simulated data, as expected, but it was a significant amelioration in two of the three real-data cases. CONCLUSIONS: The proposed automatic detection approach represents a suitable strengthening and simplification of pre-processing data analyses. The proposed 'discrepancy' evaluation, after automatic pruning, seems to be a suitable way to render negligible the risk of loose non-artifactual activity when applying BSS methods to real data. SIGNIFICANCE: The present noise reduction procedure, including ICA separation phase, automatic artifactual ICs selection and 'discrepancy' control cycle, showed good performances both on simulated and real MEG data. Moreover, application to real signals suggests the procedure to be able to separate different cerebral activity sources, even if characterized by very similar frequency contents.

Algorithms↗
PubMed2006

Automatic measurement of the nonlinearity parameter B/A in liquid media.

Using a 19-MHz focused Gaussian beam generated by a LiNbO(3) plate with an inverted polarization layer coupled to an acoustic lens, a method to automatically measure the nonlinearity parameter B/A in liquid media is presented. A liquid sample is poured into a cylindrical cell set with a glass tube on the lens surface. The focusing source emits burst waves. A tungsten rod immersed in the sample so that the top surface is normal to the beam axis behaves as a perfect reflector. The position of the reflector to maximize the amplitude receiving the returned wave determines the sound speed c of the sample. The amplitude measurement for the waves returned from the lens-sample interface and the reflector surface leads to the determination of the density rho and attenuation coefficient alpha, respectively. Then the B/A is finally derived by summarizing the measured result of the 38-MHz second harmonic amplitude contained in the sound returned back from the reflector, together with rho, c and alpha. All these are automatically processed in a short time less than 2 min with the performance of computer controlled instruments.

Algorithms↗
PubMed1992

Determination of fluid and gel domain sizes in two-component, two-phase lipid bilayers. An electron spin resonance spin label study.

The average sizes of fluid and gel domains in the two-component, two-phase system formed from mixtures of dimyristoyl phosphatidylcholine and distearoyl phosphatidylcholine were determined from an analysis of the electron spin resonance spectral lineshapes of a dimyristoyl phosphatidylcholine-nitroxide spin label as a function of spin label concentration. The ratio, R, of the intensities measured at two magnetic field strengths was found to be diagnostic of a statistical distribution of spin labels in disconnected domains. R is defined as V'/2Vpp, where Vpp is the maximum intensity and V' is the intensity at a position in the wings of a first derivative electron spin resonance line that is a constant multiple of the peak-to-peak linewidth. The intensity ratio for Gaussian or Voigt lineshapes is less than or equal to the value for a Lorentzian lineshape. The intensity ratio was found to be greater than the value for a Lorentzian line when spectra from disconnected domains containing a statistical distribution of spin labels undergoing spin-spin interactions were summed. The intensity ratio, R, calculated by spectral simulations as a function of the average number of labels per domain, N, was found to increase to a maximum with increasing N and then to decrease. The dependence on spin label concentration of the experimentally measured intensity ratios paralleled this predicted behavior. A method is presented to calculate the average number of lipids per fluid or gel domain based on a knowledge of R, and of the distribution of the spin label between the fluid and gel phases determined from the phase diagram. The results demonstrate that the number of lipids per domain increases linearly from a fixed number of nucleation sites, as the fraction of the phase that is disconnected increases. At any given mole fraction of the particular phase, the gel domains are bigger than the fluid domains because they have a lower nucleation density. The results also suggest that the disconnected domains are, in most cases, nonrandomly distributed in the plane of the bilayer.

Dimyristoylphosphatidylcholine↗