Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Feature selection”

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.

At least 1,549 records · Page 86Linked to original sources
PubMed2004

Quantitative histopathological analysis of cervical intra-epithelial neoplasia sections: methodological issues.

OBJECTIVES: As part a Program Project to evaluate emerging optical technologies for cervical neoplasia, our group is performing quantitative histopathological analysis of biopsies from 1,800 patients. Several methodological issues have arisen with respect to this analysis: (1) Finding the most efficient way to compensate for staining intensity variation with out losing diagnostic information; (2) Assessing the inter- and intra-observer variability of the semi-interactive data collection; and (3) the use of non-overlapping cells from the intermediate layer only. METHODS: Non-overlapping quantitatively stained nuclei were selected from 280 samples with histopathological characteristics of normal (199), koilocytosis (37), CIN 1 (18), CIN 2 (10) and CIN 3 (16). Linear discriminant analysis was used to assess the diagnostic information in three different feature sets to evaluate and compare staining intensity normalization methods. Selected feature values and summary scores were used to evaluate intra- and inter-observer variability. RESULTS: The features normalized by the internal subset of the imaged cells had the same discriminatory power as those normalized by the control cells and by both normalization methods seem to have additional discriminatory power over the set of features which do not require normalization. The use of the internal subset decreased the image acquisition time by approximately 50% at each center, respectively. The intra- and inter-observer variability was of a similar size. Good performance was obtained by measuring the intermediate layer only. CONCLUSION: The use of intensity normalization from a subset of the imaged non-overlapping intermediate layer cells works as well as or better than any of the other methods tested and provides a significant timesaving. Our intra- and inter-observer variability do not seem to affect the diagnostic power of the data. Although this must be tested in a larger data set, the use of intermediate layer cells only may be acceptable when using quantitative histopathology.

Cell Nucleus↗
PubMed1995

Selectivity of fatty acid mobilization: a general metabolic feature of adipose tissue.

This study extends our earlier work (T. Raclot and R. Groscolas. J. Lipid Res. 34: 1515-1526, 1993), which showed that, under norepinephrine-stimulated lipolysis, fatty acids of rat retroperitoneal fat cells are selectively mobilized. The present study examines whether this selective mobilization of fatty acids 1) is based on their proportions in adipose tissue, 2) is a metabolic feature common to all adipose tissues, and/or 3) depends on the lipolysis-stimulating agent. Rat fat cells with two markedly different fatty acid compositions were isolated from four white adipose tissues and treated with three lipolytic agents. Fatty acid composition of in vitro released free fatty acids was compared with that of fat cell triacylglycerols, the ratio of percent in free fatty acid to percent in triacylglycerol being defined as the relative mobilization rate (RMR). The RMR of individual fatty acids was related to their molecular structure. It increased exponentially with unsaturation for a given chain length and decreased with increasing chain length for a given unsaturation. The selectivity of fatty acid mobilization was similar regardless of the fatty acid composition of adipose tissue, the tissue location, and the lipolytic agent used. Under conditions of stimulated lipolysis, the selectivity of fatty acid mobilization is therefore a general metabolic feature of adipose tissue. Fatty acids with 16-20 carbon atoms and 4 or 5 double bonds had the highest RMR (from 1.4 to > 5), whereas fatty acids with 20-22 carbon atoms and 0 or 1 double bond had the lowest RMR (from 0.3 to 0.7). For the other fatty acids, RMR was close to unity.(ABSTRACT TRUNCATED AT 250 WORDS)

Adenosine Deaminase↗
PubMed1990

Cysteinyl proteinases and their selective inactivation.

The affinity-labeling of cysteinyl proteinases may now be carried out with a number of peptide-derived reagents with selectivity, particularly for reactions carried out in vitro. These reagents have been described with emphasis on their selectivity for cysteine proteinases and lack of action on serine proteinases, the most likely source of side reactions among proteinases. Perhaps a crucial feature of this selectivity is an enzyme-promoted activation due to initial formation of a hemiketal, which may destabilize the reagent. Prominent among the reagent types that have this class selectivity are the peptidyl diazomethyl ketones, the acyloxymethyl ketones, the peptidylmethyl sulfonium salts, and peptidyl oxides analogous to E-64. The need for specific inhibitors capable of inactivating the target enzyme in intact cells and animals is inevitably pushing the biochemical application of these inhibitors into more complex molecular environments where the possibilities of competing reactions are greatly increased. In dealing with the current state and potential developments for the in vivo use of affinity-labeling reagents of cysteine proteinases, the presently known variety of cysteinyl proteinases had to be considered. Therefore this chapter has, at the same time, attempted to survey these proteinases with respect to specificity and gene family. The continual discovery of new proteinases will increase the complexity of this picture. At present the lysosomal cysteine proteinases cathepsins B and L and the cytoplasmic calcium-dependent proteinases are reasonable goals for a fairly complete metabolic clarification. The ability of investigators to inactivate individual members of this family in vivo, possibly without complications due to concurrent inactivation of serine proteinases by improvements in reagent specificity, is increasing. Among the cysteine proteinases, at least those of the papain super family, hydrophobic interactions in the S2 and S3 subsites are important and some specificity has been achieved by taking advantage of topographical differences among members of this group. Some of this has probably involved surface differences removed from the regions involved in proteolytic action. The emerging cysteine proteinases include some which, in contrast to the papain family, have a pronounced specificity in S1 for the binding of basic side chains, familiar in the trypsin family of serine proteinases. At least a potential conflict with serine proteinases can be avoided by choice of a covalent bonding mechanism. The departing group region, has not been exploited. As a sole contributor to binding, this region may be rather limited as a source of specificity.(ABSTRACT TRUNCATED AT 400 WORDS)

Affinity Labels↗
PubMed1999

Simplified multiplet pattern HSQC-TOCSY experiment for accurate determination of long-range heteronuclear coupling constants.

A new two-dimensional pulse sequence for accurately determining heteronuclear coupling constants is presented. It is derived from HSQC and HECADE techniques with B0 gradient coherence selection. The main feature of the proposed method is spectra with only one component of the IS doublet; i.e., the final result is equivalent to a selective broadband excitation of either Salpha or Sbeta spin states and a preservation of these states during the entire experiment. The effect is obtained by an appropriate combination of in- and antiphase coherences. It is demonstrated that heteronuclear single-bond as well as long-range coupling constants and their relative signs are readily evaluated. The proposed sequence is equally or less sensitive to a variation of heteronuclear one-bond couplings than previously published, closely related sequences. The new method is applied to a peptide sample for determination of 3JN, Hbeta.

Arginine Vasopressin↗
PubMed2005

Improved classification of medical data using abductive network committees trained on different feature subsets.

This paper demonstrates the use of abductive network classifier committees trained on different features for improving classification accuracy in medical diagnosis. In an earlier publication, committee members were trained on different subsets of the training set to ensure enough diversity for improved committee performance. In situations characterized by high data dimensionality, i.e. a large number of features and a relatively few training examples, it may be more advantageous to split the feature set rather than the training set. We describe a novel approach for tentatively ranking the features and forming subsets of uniform predictive quality for training individual members. The abductive network training algorithm is used to select optimum predictors from the feature set at various levels of model complexity specified by the user. Using the resulting tentative ranking, the features are grouped into mutually exclusive subsets of approximately equal predictive power for training the members. The approach is demonstrated on three standard medical diagnosis datasets (breast cancer, heart disease, and diabetes). Three-member committees trained on different feature subsets and using simple output combination methods reduce classification errors by up to 20% compared to the best single model developed with the full feature set. Results are compared with those reported previously with members trained through splitting the training set. Training abductive committee members on feature subsets of approximately equal predictive power achieves both diversity and quality for improved committee performance. Ensemble feature subset selection can be performed using GMDH-based learning algorithms. The approach should be advantageous in situations characterized by high data dimensionality.

Algorithms↗
PubMedDate not supplied

Comparative fatty acid selectivity of lipases in esterification reactions with glycerol and diol analogues in organic media.

Reaction selectivity of Pseudomonas cepacia, Rhizomucor miehei, and Candida antarctica B lipases was assessed in multicompetitive esterification reaction mixtures containing an homologous series of n-chain even carbon number fatty acid (FA; C4-C18) substrates and a single alcohol cosubstrate (glycerol, 1,2-propanediol (1,2-PD), or 1, 3-propanediol (1,3-PD)) in tert-butyl methyl ether at water activity of 0.69 or 0.90 and a reaction temperature of 35 degrees C. For P. cepacia lipase, the ordinal patterns of FA selectivities observed were, with glycerol, C8 > C10, C6, C16 > other FA; with 1,2-PD and 1, 3-PD, C16 > C8 > C14 > other FA. For R. miehei lipase, the ordinal patterns of FA selectivities observed were, with glycerol, C8 > C12 > C10, C14 > other FA; with 1,2-PD and 1,3-PD, C8 > C12 > other FA. For C. antarctica B lipase, the ordinal patterns of FA selectivities observed were, with glycerol, C8 > C10, C6, C12 > other FA; with 1, 2-PD, C8 > C10, C6 > other FA; and with 1,3-PD, C8 > C10 > C6 > other FA. The differences in selectivity among FA ranged up to 16-fold, depending upon the lipase and alcohol cosubstrate used. These findings represent intrinsic and substrate-modulated features of FA selectivities that are of particular relevance to the use of lipases for acylglycerol synthesis reactions.

Burkholderia cepacia↗
PubMed2001

Characterization of beta(1)-selectivity, adrenoceptor-G(s)-protein interaction and inverse agonism of nebivolol in human myocardium.

Intrinsic activity and beta(1)-selectivity are important features of beta-blockers in the treatment of patients with coronary syndromes and heart failure. In human myocardium, intrinsic activity and beta(1)-selectivity of the novel beta-adrenoceptor antagonist nebivolol have not yet been determined. The study examines intrinsic activity, beta-adrenoceptor-G-protein coupling and beta(1)-selectivity of nebivolol and bisoprolol in human ventricular myocardium. Furthermore, intrinsic activity of both compounds is compared to the one of bucindolol, carvedilol and metoprolol in human atrial myocardium. Radioligand binding studies ([(125)I]-lodocyanopindolol) were performed on membrane preparations of human failing and nonfailing myocardium and on COS-7 cells transfected with human beta(1)- and beta(2)-adrenoceptors, respectively. Functional experiments were carried out on isolated muscle preparations of human left ventricular and right atrial myocardium from failing and nonfailing hearts. Radioligand binding studies reveal 3 - 4 fold beta(1)-selectivity for nebivolol and 16 - 20 fold beta(1)-selectivity for bisoprolol in human myocardium. In COS-7-cells, beta(1)-selectivity is 3 fold for nebivolol and 15 fold for bisoprolol. Neither the binding of nebivolol nor of bisoprolol is affected by the presence of guanylylimidodiphosphate (Gpp(NH)p). Nebivolol and bisoprolol exert similar inverse agonist activity in human ventricular as well as atrial myocardium. In atrial myocardium, inverse agonism of both compounds is higher compared to bucindolol, equal to carvedilol and lower compared to metoprolol. Favourable haemodynamic effects of nebivolol in humans are not due to beta(1)-selectivity or partial agonist activity of this agent. Other mechanisms, i.e. the production of nitric oxide, may thus be responsible for its unique haemodynamic profile.

Adrenergic beta-Agonists↗
PubMed1990

Predictive value of the four good prognostic features in DSM-III-R schizophreniform disorder.

DSM-III-R divides schizophreniform disorder into 2 subtypes with and without good prognostic features. The 4 prognostic features have been selected based on the literature, and the presence of at least 2 should indicate a good prognosis. The predictive value of the good prognostic features was tested in a sample of 16 untreated patients with DSM-III-R schizophreniform disorder with known long-term outcome based on personal follow-up examination. No correlation between the presence of 2 or more features and favorable outcome was observed. Confusion, disorientation or perplexity at the height of the psychotic episode was the only feature consistently associated with a favorable outcome in this sample. The introduction of good prognostic features of schizophreniform disorder by DSM-III-R has been done without due consideration of the methodological problems of prediction research.

Follow-Up Studies↗
PubMed1999

Focal fibrosis: a common breast lesion diagnosed at imaging-guided core biopsy.

OBJECTIVE: Focal fibrosis is a benign breast lesion commonly diagnosed by imaging-guided core biopsy. The goal of this study is to determine the frequency of focal fibrosis diagnosed at core biopsy and to describe its imaging features. MATERIALS AND METHODS: A consecutive series of 894 imaging-guided breast core biopsies were reviewed, and all cases of focal fibrosis were selected. The imaging features of each lesion were characterized. All lesions had been reviewed during radiologic-histologic review sessions to assess for accurate needle positioning and concordant results. Follow-up imaging and histologic data were reviewed to document lesion stability. RESULTS: Focal fibrosis was diagnosed in 80 (8.9%) of 894 imaging-guided core biopsies: 20 (8.7%) of 229 sonographically guided biopsies and 60 (9.0%) of 665 mammographically guided biopsies. Of 75 mammographically visible lesions, 39 (52%) were masses, 29 (39%) were densities, and seven (9.3%) were clusters of calcifications. Thirty-five hypoechoic lesions were visualized on sonography: 29 (80%) were oval, and six (17%) were irregularly shaped. Six (21%) of the 28 oval masses showed posterior enhancement, four (14%) posterior shadowing, and 19 (68%) neither feature. Fifty-two (65%) of 80 patients with focal fibrosis had routine imaging follow-up; all had stable findings (mean follow-up period, 27 months). No false-negative cases were identified. CONCLUSION: Focal fibrosis most commonly appears as an enlarging solid mass or developing density on mammography or as an oval mass on sonography. Our data suggest that focal fibrosis accounts for 9% of lesions that undergo imaging-guided core biopsy and that the diagnosis can be accurately reached using imaging-guided biopsy.

Adult↗
PubMed2003

Three-dimensional quantitative structure-activity relationship studies on selected MT1 and MT2 melatonin receptor ligands: requirements for subtype selectivity and intrinsic activity modulation.

The three-dimensional quantitative structure-activity relationship comparative molecular field analysis (3D-QSAR CoMFA) approach was applied to some classes of melatonin (MLT) membrane receptor ligands, with the principal aim of exploring the correlation between their steric features and MT(2)-selective antagonism. Binding data obtained from cloned MT(1) and MT(2) receptor subtypes were used to develop 3D-QSAR models for agonists and for antagonists at the two receptor subtypes, looking for the structural requirements for receptor subtype selectivity. In particular, we superposed the compounds showing antagonist activity, or very low intrinsic activity at the GTPgammaS test, following the hypothesis that the occupation of an additional pocket positioned out of the plane of MLT is one of the major determinants for MT(2) selectivity; the statistical models obtained confirmed this hypothesis. Structure-intrinsic activity relationship studies, applied to a set of compounds homogeneously tested, allowed the identification of the structural features whose modulation shifts the behavior from that of the agonist to that of the antagonist. The pocket out of the plane of MLT was identified as one of the key features for obtaining selective MT(2) antagonists. The reliability of our statistical models was further confirmed by the correct prediction of the pharmacological behavior of some N-substituted melatonin derivatives, which were prepared and tested on cloned receptor subtypes.

Ligands↗
PubMed1999

The perception of visual images encoded in musical form: a study in cross-modality information transfer.

This study demonstrates the ability of blind (previously sighted) and blindfolded (sighted) subjects in reconstructing and identifying a number of visual targets transformed into equivalent musical representations. Visual images are deconstructed through a process which selectively segregates different features of the image into separate packages. These are then encoded in sound and presented as a polyphonic musical melody which resembles a Baroque fugue with many voices, allowing subjects to analyse the component voices selectively in combination, or separately in sequence, in a manner which allows a subject to patch together and bind the different features of the object mentally into a mental percept of a single recognizable entity. The visual targets used in this study included a variety of geometrical figures, simple high-contrast line drawings of man-made objects, natural and urban scenes, etc., translated into sound and presented to the subject in polyphonic musical form.

Humans↗
PubMed1980

[Variability of a Cephalosporium acremonium culture for 2 quantitative traits: antibiotic formation and proteolytic activity].

Exposure of Cephalosporium acremonium, strain 1435, to N-nitrozo-N-methylbiuret resulted in a changed variation coefficient with respect to two quantitative features--the antibiotic production and proteolytic activity. Correlation between the variation coefficient and mutagen exposure time was different for every feature. Positive correlation was found in variation with respect to the antibiotic production and proteolytic activity in populations of various Cephalosporium acremonium strains chosen with regard to one or two of the above features. The level and form of the correlation in variation of the above features in populations changed during selection. Selection according to the two quantitative features resulted in an increased correlation coefficient.

Acremonium↗
PubMed2001

CaT1 manifests the pore properties of the calcium-release-activated calcium channel.

The calcium-release-activated Ca2+channel, ICRAC, is a highly Ca2+-selective ion channel that is activated on depletion of either intracellular Ca2+ levels or intracellular Ca2+ stores. The unique gating of ICRAC has made it a favourite target of investigation for new signal transduction mechanisms; however, without molecular identification of the channel protein, such studies have been inconclusive. Here we show that the protein CaT1 (ref. 4), which has six membrane-spanning domains, exhibits the unique biophysical properties of ICRAC when expressed in mammalian cells. Like ICRAC, expressed CaT1 protein is Ca2+ selective, activated by a reduction in intracellular Ca2+ concentration, and inactivated by higher intracellular concentrations of Ca2+. The channel is indistinguishable from ICRAC in the following features: sequence of selectivity to divalent cations; an anomalous mole fraction effect; whole-cell current kinetics; block by lanthanum; loss of selectivity in the absence of divalent cations; and single-channel conductance to Na+ in divalent-ion-free conditions. CaT1 is activated by both passive and active depletion of calcium stores. We propose that CaT1 comprises all or part of the ICRAC pore.

Animals↗
PubMed2004

Categorization in the monkey hippocampus: a possible mechanism for encoding information into memory.

The mammalian hippocampus processes sensory information into memory. The neurobiological basis of this representation, as well as the type of information that is encoded, is central to understanding how memories are formed. Normally, there is an infinite amount of information that could be encoded for any given stimulus. Thus, the question arises as to how the hippocampus selects and encodes features of a given stimulus. Here, we show that neurons in the hippocampus of the monkey appear to categorize types of visual stimuli presented in a delayed-match-to-sample memory task. By extracting unique combinations of features, these category cells are able to encode aspects of behaviorally important images instead of encoding all visual details. The subject is then able to rapidly select an appropriate response to that stimulus when distracting stimuli are presented simultaneously, thereby facilitating performance. Moreover, across animals, this specific type of encoding differed considerably. Just as in humans, different monkeys attended to and selected different aspects of the same stimulus image, most likely reflecting different histories, strategies, and expectations residing within individual hippocampal networks.

Animals↗
PubMed1979

Some models of genetic selection.

This paper begins with a description of the classical theory of viability selection in which probabilities that individuals of various genotypes survive are in proportions that do not change with time and are independent of population structure. Salient features of viability selection with one and two loci are reviewed. This theory is intimately connected with the usual theory of mass selection in quantitative genetics. It is well known that the mean of the relative viabilities does not necessarily increase if there is viability selection at more than one locus. It also turns out that if there is selection for fecundity with one locus, the mean fecundity may steadily decrease or oscillate rather than increase. This and the fact that a Hardy-Weinberg structure may no longer exist at any stage of life may have a bearing on predicting progress from artificial selection on reproductive characters. Classical viability selection theory does not completely describe natural selection. Other possibilities are discussed. Among these is the density and frequency dependent selection induced when the population lives in a limited habitat. Implications in quantitative genetics are discussed.

Alleles↗
PubMed1993

Feature processing during high-rate auditory selective attention.

Auditory event-related brain potentials (ERPs) and reaction times were analyzed in a selective attention task in which subjects attended to tone pips presented at high rates (interstimulus intervals [ISIs] of 40-200 msec). Subjects responded to infrequent target tones of a specified frequency (250 or 4000 Hz) and location (left or right ear) that were louder than otherwise identical tones presented randomly to the left and right ears. Negative difference (Nd) waves were isolated by subtracting ERPs to tones with no target features from ERPs to the same tones when they shared target location, frequency, or both frequency and location cues. Nd waves began 60-70 msec after tone onset and lasted until 250-350 msec after tone onset, even for tones with single attended cues. The duration of Nd waves exceeded the ISIs between successive tones, implying that several stimuli underwent concurrent analysis. Nd waves associated with frequency processing had scalp distributions different from those associated with location processing, implying that the features were analyzed in distinct cortical areas. Nd waves specific to auditory feature conjunction were isolated. These began at latencies of 110-120 msec, some 30-40 msec after the Nds to single features. The relative timing of the different Nd waves suggests that auditory feature conjunction begins after a brief parallel analysis of individual features but before feature analysis is complete.

Acoustic Stimulation↗
PubMed2001

Size-based selection in rapid serial visual presentation.

Several recent studies [Farell, B., & Pelli, D. (1993). Can we attend to large and small at the same time? Vision Research, 33, 2757-72; Shih, S., & Sperling, G. (1996). Is there feature-based attentional selection in visual search? Journal of Experimental Psychology: Human Perception and Performance, 22, 758-79] have found that visual selection based on the size of stimuli is impossible when the stimuli are presented in rapid succession in overlapping positions (RSVP paradigm). In the present study effective size-based selection is demonstrated in several conditions with RSVP stimuli. Attention to specific size is highly efficient when stimuli are presented in a single location (at fixation point) and may be possible also with a few (2-4) locations. When overlapping small and large characters are presented without abrupt onsets, then selection by size is effective at least over six locations. The results are explained by certain mandatory properties of spatial attention.

Adult↗
PubMed2003

Rabbit immune repertoires as sources for therapeutic monoclonal antibodies: the impact of kappa allotype-correlated variation in cysteine content on antibody libraries selected by phage display.

The rabbit immune repertoire has long been a rich source of diagnostic polyclonal antibodies. Now it also holds great promise as a source of therapeutic monoclonal antibodies. On the basis of phage display technology, we recently reported the first humanization of a rabbit monoclonal antibody. The allotypic diversity of rabbit immunoglobulins prompted us to compare different rabbit immune repertoires for the generation and humanization of monoclonal antibodies that bind with strong affinity to antigens involved in tumor angiogenesis. In particular, we evaluated the diversity of unselected and selected chimeric rabbit/human Fab libraries that were derived from different kappa light chain allotypes. Most rabbit light chains have an extra disulfide bridge that links the variable and constant domains in addition to the two intrachain disulfide bridges shared with mouse and human kappa light chains. Here we evaluate the impact of this increased disulfide bridge complexity on the generation and selection of chimeric rabbit/human Fab libraries. We demonstrate that rabbits with mutant bas and wild-type parental b9 allotypes are excellent sources for therapeutic monoclonal antibodies. Featured among the selected clones with b9 allotype is a rabbit/human Fab that binds with a dissociation constant of 1nM to both human and mouse Tie-2, which will facilitate its evaluation in mouse models of human cancer. Examination of 228 new rabbit antibody sequences allowed for a comprehensive comparison of the LCDR3 and HCDR3 length diversity in rabbits. This study revealed that rabbits exhibit an HCDR3 length distribution more closely related to human antibodies than mouse antibodies.

Animals↗