Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Principal component”

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 955 records · Page 53Linked to original sources
PubMed1982

Fatness and fat patterning among athletes at the Montreal Olympic Games, 1976.

Six skinfold measurements for male and female athletes (N=456) at the 1976 Montreal Olympic Games were analyzed to identify principal components of fatness and anatomical distribution of fat, i.e., fat patterning. As in non-athletes, two principal components were evident among the athletes. All skinfolds were correlated positively with the first component, which was termed fatness, while extremity fat measurements were correlated positively and trunk measurements were correlated negatively with the second principal component, which was termed an extremity/trunk ratio component. The two principal components accounted for about 85% of the variance. The first component was related to control variables in order of descending contribution to its variance as follows: sex (21-31%), sport (19%), ethnicity (3%), and age (1-3%). Likewise, the second component (extremity/trunk ratio) was related to the control variables: sex (20-35%), age (4-7%), ethnicity (2%), and sport (2%). Fatness is more influenced by sport and by inference training than is the anatomical distribution or patterning of fat on the extremities relative to the trunk. The latter characteristic may be more dependent on biological or environmental factors unrelated to sport and training.

Adipose Tissue↗
PubMedDate not supplied

Component analysis in multivariate aging research.

A method of components analysis, related to principal components analysis, is described for applications in multivariate cross-sectional or longitudinal data. The method may be useful to researchers seeking to reduce the number of observed variables through scale construction. The method creates component variables as weighted sums of the observed variables using weights that are identical across groups and occasions. The statistical and conceptual properties of these components are discussed. The method is contrasted with traditional principal components analysis and factor analysis. An application of the method is presented using longitudinal WAIS and WAIS-R data.

Cross-Sectional Studies↗
PubMed1994

Multivariate analyses of elemental hair concentrations from a medieval Nubian population.

Elemental hair concentrations were obtained from 168 mummified individuals recovered during excavations of cementeries S and R (A.D. 550-1450), at Kulubnarti, Republic of the Sudan (Van Gerven et al., 1981). Concentrations of calcium (Ca), magnesium (Mg), strontium (Sr), manganese (Mn), zinc (Zn), iron (Fe), and copper (Cu) were determined using inductively coupled plasma emission spectrometry (ICP) (Sandford, 1984; Sandford et al., 1983). Following univariate statistical reanalyses of these data (Sandford and Kissling, 1993a, b), we applied principal component analysis and multidimensional scaling to clarify their multivariate structure. Similar underlying associations were revealed in the two cemeteries. The first principal component, consisting of Mg, Ca, Sr, and Mn, may reflect inherent chemical similarities coupled with specific dietary factors (e.g., vegetation intakes) and physiological processes (e.g., bone remodeling). The second principal component, a contrast between Fe and Zn, may be due to their competitive relationship during absorption. The third principal component consists of Cu alone (in cemetery S), and Cu and Fe (in cemetery R), suggesting utilization of animal protein and an underlying synergism between Cu and Fe, respectively. Multidimensional scaling substantiates a three-dimensional model for describing elemental covariation. While interpretation of the first two dimensions was analogous to those of the first two principal components, the third dimension may represent antagonism between pairs of elements during absorption, transport and/or utilization (cemetery R: Cu vs. Zn; cemetery S: Cu vs. Zn, Fe vs. Mn). While these results provide the most persuasive evidence to date that elemental hair concentrations from this population reflect chiefly biogenic processes, isolation of diagenetic and exogenous effects requires further investigation through controlled studies.

Adolescent↗
PubMed2006

Spatiotemporal features of severe air pollution in northern Taiwan.

BACKGROUND, AIMS AND SCOPE: This research attempted to identify the dominant factors simultaneously affecting the airborne concentrations of five air pollutants with principal component analysis and to determine the meteorologically related parameters that cause severe air-pollution events. According to the definition of subPSI and PSI values through the U.S. EPA, the historical raw data of five criteria air pollutants, SO2, CO, O3, PM10 and NO2, were calculated as daily subPSI values. In addition to the airborne concentrations, this study simultaneous collected the surface meteorological parameters of the Taipei meteorological station, established by the Central Weather Bureau. METHODS: Principal component analysis was conducted to screen severe air pollution scenarios for five air pollutants: SO2, CO, O3, PM10 and NO2. The concentrations of various air pollutants measured at 17 air-quality stations in northern Taiwan from 1995 to 2001 were transformed into daily subPSI values. The correlation analysis of the five air pollutants and four meteorological parameters (wind speed, temperature, mixing height and ventilation rate) were included in this research. After screening severe air pollution scenarios, this study recognized the synoptic patterns easily causing the severe air-pollution events. RESULTS AND DISCUSSION: Analytical results showed that the eigenvalues of the first two principal components for SO2, CO, O3, PM10 and NO2 were greater than 1. The first component of five air pollutants explained 64, 64, 67, 76 and 63% of subPSI variance for SO2, CO, O3, PM10 and NO2, respectively. Only the correlation coefficient of NO2 and CO had statistically significant positive values (0.82); other pollutant pairs presented medium (0.4 to 0.7) or low (0 to 0.4) positive values. The correlation coefficients for air pollutants and three meteorological parameters (wind speed, mixing height and ventilation index) were medium or low negative values. In northern Taiwan, spring was most likely induced high concentrations and the component scores of the first component for SO2, CO, PM10 and NO2; summer was the worst season that caused high O3 episodes. Consequently, the analytical results of factor loadings for the first principal component and emission inventory of various sources revealed that mobile sources were dominant factors affecting ambient air quality in northern Taiwan. CONCLUSION: According to the results of principal component analysis for the five air pollutants, the first two of 17 components were cited as major factors and explained 71% of subPSI variance. Based on the inventory of NOx emissions and the isopleth diagram of factor loading for the first component, mobile sources in the southwest Taipei City accounted for the highest factor loading values and emission inventory values. Synoptic analysis and principal component analysis demonstrated that three types of weather patterns (high-pressure recirculation, prefrontal warm sector and the southwesterly wind system) easily caused the severe air-pollution scenarios. In summary, if severe air-pollution days occurred, the average meteorological parameters experienced adverse conditions for diffusing air pollutants; that is, the average values of wind speed, mixing height and ventilation index were lower than 2.1 ms(-1), 360 m and 800 m2s(-1), respectively. If one of the three synoptic patterns were to occur in combination with adverse meteorological conditions, severe air-pollution events would be developed. RECOMMENDATION AND OUTLOOK: By utilizing synoptic patterns, this work found three weather systems easily caused severe air-pollution events over northern Taiwan. Analytical results showed, respectively, the wind speed and mixing height were less than 2.1 m/s and 360 m during severe air-pollution events.

Air Movements↗
PubMed1993

Quantifying periodic activity in central pattern generators: the crayfish swimmeret.

Experimental recordings from neurons are generally difficult to quantify. Spike patterns and membrane potential cycles are often compared qualitatively and are hence limited by human resolution and judgement. This paper describes a statistical method for quantifying individual spike patterns and membrane potential fluctuations in oscillating systems, allowing quantification of successive 'individual' bursts. The neuronal activities of the network can therefore be subject to more extensive analyses and comparisons. Principal-component analysis was used to extract the principal components from a sequence of bursts recorded in vitro from the crayfish (Pacifasticus leniusculus) swimmeret. The principal component coefficients were correlated with intuitive biological concepts such as burst displacement, number of spikes in a burst, and burst width. The bursting of both interneurones and motor neurones were found to be modulated by some unidentified oscillator that displaced the bursts in time. The amplitude of the time displacements was significant, i.e., of the order of the bursting period (1.5 s), and the period of the oscillation was approximately 200 bursts. The strength of the method is that this feature was detected without any prior suspicion of its existence.

Action Potentials↗
PubMed2006

Factors influencing osteological changes in the hands and fingers of rock climbers.

This study examines the osteological changes in the hands and fingers of rock climbers that result from intense, long-term mechanical stress placed on these bones. Specifically, it examines whether rock climbing leads to metacarpal and phalange modelling in the form of increased cortical thickness as well as joint changes associated with osteoarthritis. This study also attempts to identify specific climbing-related factors that may influence these changes, including climbing intensity and frequency of different styles of climbing. Radiographs of both hands were taken for each participant and were scored for radiographic signs of osteoarthritis using an atlas method. Total width and medullary width were measured directly on radiographs using digital calipers and used to calculate cross-sectional area and second moment of area based on a ring model. We compared 27 recreational rock climbers and 35 non-climbers for four measures of bone strength and dimensions (cross-sectional area, second moment of area, total width and medullary width) and osteoarthritis. A chi-squared test for independence was used to compare climber and non-climber osteoarthritis scores. For each measure of bone strength climbers and non-climbers were compared using a manova test. Significant manova tests were followed by principal components analysis (PCA) and individual anova tests performed on principal components with eigenvalues greater than one. A second PCA was performed on the climber subsample and the first principal component was then used as the dependent variable in linear regression variable selection procedures to determine which climbing-related variables affect bone thickness. The results suggest that climbers are not at an increased risk of developing osteoarthritis compared with non-climbers. Climbers, however, do have greater cross-sectional area as well as second moment of area. Greater total width, but not meduallary width, indicates that additional bone is deposited subperiosteally. The strength of the finger and hand bones are correlated with styles of climbing that emphasize athletic difficulty. Significant predictors include the highest levels achieved in bouldering and sport climbing.

Adult↗
PubMedDate not supplied

Interrelationships among certain measures of growth and maturation rate in boys during adolescence.

Relationships among ages at attaining 17 or 21 indices of maturity were considered in a longitudinal sample of 177 Polish boys examined at annual intervals from 1961 to 1972. Maturity indicators included ages at peak velocity for stature, sitting height, leg length and weight; ages at attaining 80%, 90%, 95% and 99% of adult stature; ages at attaining the median skeletal maturity scores (TW-2) characteristic of chronological ages 11, 12, 13, 14, and 15 years; and ages at attaining stages II and IV of genital and public hair development. Age at initiation of the stature spurt (take-off) and ages at eruption of 14, 20 and 26 permanent teeth were ascertained for only 111 boys. All intercorrelations among the developmental indicators were positive. Ordering the correlation matrix gave three clusters: (1) a large central group including age at take-off and ages at all peak velocities, at genital and pubic hair stages II and IV, at attaining 90%, 95% and 99% of adult stature, and at the later stages of skeletal maturity; (2) indices of the tempo of maturation during prepubertal and/or early pubertal stages; and (3) ages at attaining a given number of permanent teeth. Results of a principal components analysis of the ages indicated two principal components, the first accounting for about 77% of the sample variance and the second for about 12%. The first principal component is apparently a general maturity factor, while the second apparently relates to the rate of skeletal maturity during pre-adolescence.

Adolescent↗
PubMed2001

Statistical analysis of hyperspectral data from two Swedish lakes.

CASI data has been collected from two lakes in Sweden. In this paper, some statistical properties of CASI spectral data have been discussed. Principal component analysis is used for assessing the dimensionality of the data and the principal components were used for making chlorophyll maps. The quality of the reconstruction of the spectra from the principal components was demonstrated. Examples of the accuracy of the radiative transfer code 6S in atmospheric correction applications have been given. Furthermore, the widths and positions of the spectral bands based on the studied dataset were proposed for chlorophyll mapping. Robustness aspects of regression models have been discussed. Algorithms derived from one lake have been used to map water quality parameters in another lake. Algorithms based on principal components, as well as algorithms based on image bands, have been used.

Algorithms↗
PubMed1991

Accuracy of assessment of cardiac vagal tone by heart rate variability in normal subjects.

The correlations of 11 indexes of heart rate variability were examined with pharmacologically determined cardiac vagal tone in 15 normal subjects at supine rest. After sympathetic influences by intravenous propranolol were eliminated, RR interval variability was measured for 10 minutes under controlled respiration (0.25 Hz), and cardiac vagal tone was determined as the decrease in mean RR interval following complete vagal blockade with atropine. Time domain indexes (standard deviation, coefficient of variance and mean successive difference) correlated strongly with vagal tone (r = 0.87, 0.81 and 0.92, respectively; p less than 0.001 for all). The same was true for frequency domain indexes for the high-frequency (0.25 Hz) component calculated both by autoregressive spectrum analysis (square root of power and coefficient of component variance) and by fast Fourier transform (mean amplitude) (r = 0.91, 0.85 and 0.86, respectively; p less than 0.0001 for all). However, frequency domain indexes for the low-frequency spectral component (0.03 to 0.15 Hz) correlated less strongly (r = 0.69, 0.55 and 0.70, respectively), and the fraction of power [power/(total power greater than 0.03 Hz)] of both components showed no correlation. Principal component analysis showed that the first 6 indexes with strong correlations contained solely the first principal component closely related to vagal tone, whereas the remaining 5 indexes also contained the second component unrelated to vagal tone. These results indicate that most of the time and frequency domain analyses in use provides an accurate and common measure of cardiac vagal tone at rest.

Adult↗
PubMed2001

Toxicity of uranium mine-receiving waters to caged fathead minnows, Pimephales promelas.

Larval fathead minnows (Pimephales promelas) were placed at four exposure sites for 7 days in each of five lakes surrounding the Key Lake uranium mine in northern Saskatchewan, Canada. Fish placed in lakes receiving Mo-contaminated mill effluent demonstrated higher mortalities than those placed in lakes receiving Ni-contaminated mine-dewatering effluent, which was not significantly different from reference sites. No significant differences were detected in fish growth among the study lakes because of the high (90%) mortality in Fox and Unknown lakes. Principal components analysis characterized exposure sites by total and dissolved metal concentration. Stepwise multiple regression of fish mortality on principal components generated from total metal data revealed that principal component 1 could account for 84% of the variance associated with fish mortality. Careful examination of the metals that correlated strongly with principal component 1 and with fish mortality suggested that dietary Se toxicity probably resulted in the differential fathead minnow mortality observed among study lakes.

Animals↗
PubMed2000

A principal components-based method for the detection of neuronal activity maps: application to optical imaging.

We present a novel analysis technique for the extraction of neuronal activity patterns from functional imaging data. We illustrate this technique on data from optical imaging. Optical imaging of the mammalian visual cortex probe the patterns in which the neuronal responses to various aspects of the visual world, such as orientation and color, are spatially organized within the cortex. Recovering these patterns from the image data is a challenging problem as the neuronal response signal is extremely weak in comparison to the background vegetative processes (e.g., circulation and respiration). The proposed technique obtains the neuronal activity pattern using a combination of principal component analysis and statistical significance testing. The performance of this method is compared with the results of existing analysis techniques. The comparison shows the new method to be more sensitive than previous methods.

Animals↗
PubMed2004

A genomic predictor of oral squamous cell carcinoma.

OBJECTIVES/HYPOTHESIS: The objective was to identify a genomic profile that predicts the likelihood of oral squamous cell carcinoma compared with normal oral mucosa in unknown tissue samples. STUDY DESIGN: Using a training set of tissue samples that were histologically classified as oral squamous cell carcinoma or normal mucosa, the authors used principal component analysis to develop a genomic predictor for oral squamous cell carcinoma. On a separate test set of unclassified samples, the authors used the predictor to classify the samples, then evaluated the performance of the predictor using histological diagnosis. METHODS: The authors used a data set consisting of messenger RNA extracted from 29 oral squamous cell carcinoma and 19 normal oral mucosa tissue samples and hybridized to Affymetrix oligonucleotide microarrays containing probe sets for 7070 genes and expressed sequence tags. The samples were divided into a training set of 15 oral squamous cell carcinoma and 10 normal samples and a test set consisting of the remaining samples. Using principal component analysis on the training set, the authors found a composite gene expression vector (principal component vector), which they used to compute likelihood ratios for oral squamous cell carcinoma on the test set. By calculating the contribution of each gene to the principal component vector, the authors identified genes with the greatest predictive value. RESULTS: Using the likelihood ratio, the authors correctly classified all 23 samples in the test set as either oral squamous cell carcinoma or normal. The authors found that many of the most predictive genes are known to be markers of squamous cell carcinoma or normal mucosa. CONCLUSION: Principal component analysis can be used with genomic microarray data to correctly predict the presence of oral squamous cell carcinoma in unknown tissue samples.

Biomarkers, Tumor↗
PubMed2001

Deciphering the genetic architecture of a multivariate phenotype.

A heritable multivariate quantitative phenotype comprises several correlated component phenotypes that are usually pleiotropically controlled by a set of major loci and environmental factors. One approach to decipher the genetic architecture of a multivariate phenotype, in particular to map the underlying loci, is to reduce the dimensionality of the data by means of a data reduction technique, such as principal component analysis. The extracted principal components are then analyzed in conjunction with marker data to map the underlying loci. We have examined the efficiency of this approach with and without taking into account the correlation structure of the multivariate phenotype when extracting principal components. We have assumed that genome-wide scan data on sibpairs are available for low-density (widely spaced) and high-density markers. Using extensive simulations, based on three models of the multivariate phenotype, we have shown that although ignoring the correlation structure of the multivariate phenotype does not have any serious impact on the efficiency of mapping the underlying trait loci in wide marker intervals, there is a significant adverse effect of this practice for fine-mapping. We, therefore, recommend that the correlation structure of the multivariate phenotype be carefully examined to decide on the strategy of extracting principal components for deciphering the genetic architecture of the multivariate phenotype.

Chromosome Mapping↗
PubMed1998

[A multiple variable analysis about the risk factors of breast cancer in women].

A comprehensive investigation to the risk factors of breast cancer was carried out in 146 pairs of women subjects, with the method of 1:1 matched case-control study, and of logistic regression analysis and principal component analysis. As a result, a regression equation which includes 5 independent variables has been established by logistic regression analysis; Ten principal components were extracted from 22 risk factors of breast cancer through principal component analysis, and the cumulative contribution of variances of the 10 principal components is 64.8%. The result combining the mono-factor analysis and the multiple variables analysis indicated that the principal risk factors of breast cancer are cadre occupation, the age of primiparity and late marriage, more intake of salted meat and sausage, passive smoking, having the history of family tumor, of breast benign tumor, and of the mastadenitis; but the more number of giving births, longer period of time of lactation, more vegetable intake, and taking more vitamin may have preventive effect for breast cancer.

Adult↗
PubMed2006

Rotation forest: A new classifier ensemble method.

We propose a method for generating classifier ensembles based on feature extraction. To create the training data for a base classifier, the feature set is randomly split into K subsets (K is a parameter of the algorithm) and Principal Component Analysis (PCA) is applied to each subset. All principal components are retained in order to preserve the variability information in the data. Thus, K axis rotations take place to form the new features for a base classifier. The idea of the rotation approach is to encourage simultaneously individual accuracy and diversity within the ensemble. Diversity is promoted through the feature extraction for each base classifier. Decision trees were chosen here because they are sensitive to rotation of the feature axes, hence the name "forest." Accuracy is sought by keeping all principal components and also using the whole data set to train each base classifier. Using WEKA, we examined the Rotation Forest ensemble on a random selection of 33 benchmark data sets from the UCI repository and compared it with Bagging, AdaBoost, and Random Forest. The results were favorable to Rotation Forest and prompted an investigation into diversity-accuracy landscape of the ensemble models. Diversity-error diagrams revealed that Rotation Forest ensembles construct individual classifiers which are more accurate than these in AdaBoost and Random Forest, and more diverse than these in Bagging, sometimes more accurate as well.

Algorithms↗
PubMed1991

[Factors related to serum lipid levels in women of middle and old age--a community study].

Serum lipid levels of 168 women ranging in age from 30 to 69 were measured at a mass health screening in an urban community in Saitama prefecture. Degree of obesity, skinfold thickness and blood pressure were measured and the relation between these physical measurements to serum lipid levels was studied. The results are as follows: (1) Mean total cholesterol level, degree of obesity, systolic blood pressure and diastolic blood pressure increased linearly with advancing age. (2) Simple correlation analysis disclosed that the level of total cholesterol was positively correlated to age, degree of obesity, systolic blood pressure and triglyceride level. HDL-cholesterol level showed a significantly negative correlation to degree of obesity, skinfold thickness and triglyceride level. Degree of obesity was positively correlated to skinfold thickness, systolic blood pressure, diastolic blood pressure, total cholesterol and triglyceride level, but was negatively correlated to HDL-cholesterol. Systolic blood pressure had a positive correlation with age, degree of obesity, diastolic blood pressure, total cholesterol and triglyceride level. (3) Principal component analysis, when performed with the variables age, degree of obesity, skinfold thickness, systolic blood pressure, diastolic blood pressure, level of total cholesterol, triglyceride and HDL-cholesterol, showed that subjects could be divided into the group having obesity with low HDL-cholesterol (first principal component), the hypertensive group (second principal component), and the group having hyperlipidemia with advancing age (third principal component). (4) Multiple regression analysis was also carried out, taking HDL-cholesterol level as the dependent variable, and age, degree of obesity, systolic blood pressure, total cholesterol and triglyceride level as independent variables. Degree of obesity and triglyceride level were negatively related to HDL-cholesterol.

Adult↗
PubMed2006

Mass spectrometry-based metabolic profiling reveals different metabolite patterns in invasive ovarian carcinomas and ovarian borderline tumors.

Metabolites are the end products of cellular regulatory processes, and their levels can be regarded as the ultimate response of biological systems to genetic or environmental changes. We have used a metabolite profiling approach to test the hypothesis that quantitative signatures of primary metabolites can be used to characterize molecular changes in ovarian tumor tissues. Sixty-six invasive ovarian carcinomas and nine borderline tumors of the ovary were analyzed by gas chromatography/time-of-flight mass spectrometry (GC-TOF MS) using a novel contamination-free injector system. After automated mass spectral deconvolution, 291 metabolites were detected, of which 114 (39.1%) were annotated as known compounds. By t test statistics with P < 0.01, 51 metabolites were significantly different between borderline tumors and carcinomas, with a false discovery rate of 7.8%, estimated with repeated permutation analysis. Principal component analysis (PCA) revealed four principal components that were significantly different between both groups, with the highest significance found for the second component (P = 0.00000009). PCA as well as additional supervised predictive models allowed a separation of 88% of the borderline tumors from the carcinomas. Our study shows for the first time that large-scale metabolic profiling using GC-TOF MS is suitable for analysis of fresh frozen human tumor samples, and that there is a consistent and significant change in primary metabolism of ovarian tumors, which can be detected using multivariate statistical approaches. We conclude that metabolomics is a promising high-throughput, automated approach in addition to functional genomics and proteomics for analyses of molecular changes in malignant tumors.

Cluster Analysis↗
PubMed2005

Correlation of KIT and platelet-derived growth factor receptor alpha mutations with gene activation and expression profiles in gastrointestinal stromal tumors.

Activating mutations of KIT and platelet-derived growth factor receptor alpha (PDGFRA) are known to be alternative and mutually exclusive genetic events in the development of gastrointestinal stromal tumors (GISTs). We examined the effect of the mutations of these two genes on the gene expression profile of 22 GISTs using the oligonucleotide microarray. Mutations of KIT and PDGFRA were found in 17 cases and three cases, respectively. The remaining two cases had no detectable mutations in either gene. The mutation status of KIT and PDGFRA was directly related to the expression levels of activated KIT and PDGFRA, and was also related to the different expression levels of activated proteins that play key roles in the downstream of the receptor tyrosine kinase III family. To evaluate the impact of mutation status and the importance of the type of mutation in gene expression and clinical features, microarray-derived data from 22 GISTs were interpreted using a principal component analysis (PCA). Three relevant principal component representing mutation of KIT, PDGFRA and chromosome 14q deletion were identified from the interpretation of the oligonucleotide microarray data with PCA. After supervised analysis, there was at least a two fold difference in expression between GISTs with KIT and PDGFRA mutation in 70 genes. Our findings demonstrate that mutations of KIT and PDGFRA affect differential activation and expression of some genes, and can be used for the molecular classification of GISTs.

Adult↗