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 19 recordsLinked to original sources
PubMed2026

Uncovering heterogeneous effects via localized feature selection.

Identifying features that interact to trigger disease, while accounting for heterogeneity across diverse populations, is essential for the development of precision and targeted medicine. Despite the availability of vast and complex health-related datasets, most existing works focus on identifying disease-associated features at the population level or within a few subpopulations, often overlooking individual-level heterogeneity within these groups. To address this limitation, we propose a framework that utilizes localized test statistics to identify disease-associated features tailored to individual profiles. Our method leverages the recently developed knockoffs methodology to control the noise level of the selection set so that the results are replicable. Moreover, it allows for the discovery of hidden heterogeneous effects within the data, as demonstrated in an application to single-cell RNA sequencing data for Alzheimer's disease. By aggregating localized feature selection results, our framework also enables powerful population-level feature selection. Our framework provides a powerful tool for exploratory studies of precision medicine, offering the potential to generate novel hypotheses for confirmatory biological experiments.

Alzheimer Disease↗
PubMed2025

A novel and robust feature selection method with FDR control for omics-wide association analysis.

Omics-wide association analysis is a very important tool for medicine and human health study. However, the modern omics data sets collected often exhibit the high-dimensionality, unknown distribution response, unknown distribution features and unknown complex association relationships between the response and its explanatory features. Reliable association analysis results depend on an accurate modeling for such data sets. Most of the existing association analysis methods rely on the specific model assumptions and lack effective false discovery rate (FDR) control. To address these limitations, the paper firstly applies a single index model for omics data. The model shows robust performance in allowing the relationships between the response variable and linear combination of covariates to be connected by any unknown monotonic link function, and both the random error and the covariates can follow any unknown distribution. Then based on this model, the paper combines rank-based approach and symmetrized data aggregation approach to develop a novel and robust feature selection method for achieving fine-mapping of risk features while controlling the false positive rate of selection. The theoretical results support the proposed method and the analysis results of simulated data show the new method possesses effective and robust performance for all the scenarios. The new method is also used to analyze the two real datasets and identifies some risk features unreported by the existing finds.

Humans↗
PubMed2026

engGNN: a dual-graph neural network for omics-based disease classification and feature selection.

Omics data, such as transcriptomics, proteomics, and metabolomics, provide critical insights into disease mechanisms and clinical outcomes. However, their high dimensionality, small sample sizes, and intricate biological networks pose major challenges for reliable prediction and meaningful interpretation. Graph neural networks offer a promising way to integrate prior knowledge by encoding feature relationships as graphs. Yet, existing methods typically rely solely on either an externally curated feature graph or a data-driven generated graph, which limits their ability to capture complementary information. To address this, we propose the external and generated Graph Neural Network (engGNN), a dual-graph framework that jointly leverages both external biological networks and data-driven generated graphs. Specifically, engGNN constructs a biologically informed undirected feature graph from established network databases and complements it with a directed feature graph derived from tree-ensemble models. This dual-graph design produces more comprehensive representations, thereby improving predictive performance and interpretability. Through extensive simulation studies and real-world applications to three independent gene expression datasets, engGNN consistently demonstrates strong classification performance compared with competitive baselines. Beyond classification, engGNN provides feature- and source-level interpretability, enabling biologically meaningful analyses such as pathway enrichment analysis. Taken together, these results highlight engGNN as a robust, flexible, and interpretable framework for disease classification and biomarker discovery in high-dimensional omics contexts.

Graph Neural Networks↗
PubMed1987

Comparison of the ID3 algorithm versus discriminant analysis for performing feature selection.

Having obtained disappointing results in a small medical data set despite the fact that our data seemed to be well suited for induction via ID3, we decided to compare the performance of ID3 to discriminant analysis. Performance was gauged by the percentage of correct classification in a second, independent data set. Examples were obtained from a cardiology project on the accuracy of auscultation. There were 107 examples in the first data set and 67 cases in the second. We found that ID3 and discriminant analysis performed equally poorly, with ID3 classifying only 60% of the second set correctly and discriminant analysis classifying 66% of the second set correctly. Also, the ID3 probability statistic for estimating the accuracy of ID3 for classifying further cases was markedly optimistic compared to our actual second data set results. Moreover, with an increase in sample size, ID3 seemed to break down, producing a large, complex decision tree of dubious generality, whereas discriminant analysis, with a larger sample size, used more independent variables but maintained its first set accuracy. These data suggest that there is a need for more sophisticated algorithms than ID3, even at the risk of giving up some computational efficiency.

Adult↗
PubMed1976

Elicited imitation of selected features of two american English dialects in Head Start children.

Three measures were used to check the bidialectal imitative facility of 100 black, white, and Spanish-speaking Head Start children. In general, blacks and Spanish-speaking subjects performed more accurately on black English markers than on Standard English markers and whites, the reverse. When the children did make an error on the feature marker they usually substituted the opposing dialectal marker. Blacks and Spanish-speaking subjects were more apt to be accurate on the total sentence when it was given in black English. Several explanations are offered for group similarities and differences.

Black or African American↗
PubMed2025

BISON: bi-clustering of spatial omics data with feature selection.

MOTIVATION: The advent of next-generation sequencing-based spatially resolved transcriptomics (SRT) techniques has reshaped genomic studies by enabling high-throughput gene expression profiling while preserving spatial and morphological context. Understanding gene functions and interactions in different spatial domains is crucial, as it can enhance our comprehension of biological mechanisms, such as cancer-immune interactions and cell differentiation in various regions. It is necessary to cluster tissue regions into distinct spatial domains and identify discriminating genes (DGs) that elucidate the clustering result, referred to as spatial domain-specific DGs. Existing methods for identifying these genes typically rely on a two-stage approach, which can lead to the phenomenon known as double-dipping. RESULTS: To address the challenge, we propose a unified Bayesian latent block model that simultaneously detects a list of DGs contributing to spatial domain identification while clustering these DGs and spatial locations. The efficacy of our proposed method is validated through a series of simulation experiments, and its capability to identify DGs is demonstrated through applications to benchmark SRT datasets. AVAILABILITY AND IMPLEMENTATION: The R/C++ implementation of BISON is available at https://github.com/new-zbc/BISON.

Software↗
PubMed1987

Infective endocarditis in children with congenital heart disease: comparison of selected features in patients with surgical correction or palliation and those without.

The diagnostic and prognostic features of 44 episodes of infective endocarditis in 42 children with congenital heart disease were reviewed. Endocarditis occurred in 18 patients who had not had surgical correction or palliation of the defect (non-operated group). There were 26 episodes in 24 patients who had been treated surgically (operated group) (16 open and eight closed cardiac operations). Endocarditis occurred soon after open heart surgery in eight patients and as a late complication in the other 16. It recurred in two patients (operated group). Invasive monitoring and low cardiac output were consistent features in those patients who had endocarditis soon after open heart surgery whereas dental treatment was a common feature in non-operated cases and after closed cardiac operations. Late cases of endocarditis after open heart surgery had various microbiological features that were not typical of infection after dental problems. Gram positive infections occurred in non-operated patients and in those who had had closed cardiac operations. The group that had open heart surgery had infections caused by Gram positive, Gram negative, and anaerobic bacteria and fungi. Fever, anaemia, leucocytosis, and positive blood cultures were the only consistent findings. Vegetations were seen in nine of 12 patients at cross sectional echocardiography. All 12 (four non-operated, one closed, and seven open cases) needed acute surgical treatment. The mortality from infective endocarditis was 17% for non-operated cases, 0% for those who had had closed heart surgery, and 50% for those who had had open heart surgery. Infective endocarditis after open heart surgery differs from that in the other subgroups in terms of microbiology, source of infection, and outcome and its early diagnosis depends on a thorough investigation of minimal symptoms and signs.

Adolescent↗
PubMed1978

Shedd's formulations concerning the hyperkinetic syndrome--an empirical test of selected features.

According to Shedd's (1968) formulations, one of the distinguishing characteristics of hyperkinetic children is the score pattern which they earn on different types of ability tests. Specifically, Shedd's work suggests that IQs from a picture vocabulary test, the WISC, and a drawing test should show the following order relationship: picture vocabulary greater than WISC greater than drawing. This hypothesis was investigated for 62 overactive children (47 boys and 15 girls) enrolled in classes for learning disabled. The findings support Shedd's since IQs from the Ammons' test were significantly higher than WISC IQs which were significantly higher than IQs from the Goodenough-Harris. The magnitudes of the mean differences in scores for the three tests were within the range indicated by Shedd and 51 of the 62 children showed the order of scores he specified.

Adolescent↗
PubMed1989

Effects of polymyxin B on selected features of equine carbohydrate overload.

Gram negative endotoxins play a contributory role in the syndrome which results from over consumption of carbohydrates by horses and ponies. Since the antibiotic polymyxin B exerts a direct anti-endotoxin effect by chemically modifying the active lipid A moiety of endotoxin, it might be expected to protect horses after carbohydrate overload and provide a new therapeutic and experimental tool for this condition. The present study was undertaken to evaluate the effect of polymyxin B on hemostatic, hemodynamic, acid-base, and clinical aspects of the syndrome resulting from carbohydrate overload. Experimentally-induced carbohydrate overload resulted in lactic acidosis, hypercoagulability, hypovolemic shock and lameness. Although there was a slight delay in the onset of clinical signs resulting from experimental carbohydrate overload in treated animals, polymyxin B administered iv at 2.5 mg/kg every 6 hr failed to significantly ameliorate the coagulopathy, acidosis, lameness and shock induced by alimentary carbohydrate overload.

Acid-Base Imbalance↗
PubMed1987

Mammography equipment: principles, features, selection.

1. Three-phase and constant potential generators may offer advantages over single-phase generators. These include shorter exposure times with less chance for motion unsharpenss, more adequate exposure and possible lower dose, and longer tube life. 2. Control of tube voltage in increments of 1 kVp are desirable for screen-film mammography and 2 kVp for xeromammography. 3. Tube loading limitations should be capable of allowing an adequate mA and mAs output. A mammography unit with a lower mA output may require longer exposure times with consequent motion unsharpness and/or higher dose. The maximum mA output is limited by the power rating of the tube and not by the power rating of the generator, which is usually much higher. Thus, tube rating charts rather than generator power rating should be used to compare the mA output of different mammography units. 4. Phototimer capability can reduce exam time, facilitate proper exposure, and reduce radiation dose. Units with a greater number of density steps and three or more photocell locations, one of which is near the edge of the film holder, provide better phototimer results. 5. A molybdenum target, beryllium window tube with 0.03 mm molybdenum added filtration is presently preferred for screen-film mammography. A tungsten target tube with 1 to 2 mm added aluminum filtration is advised for xeromammography. 6. Effective local spot size for contact mammography should be in the range of 0.2 to 0.5 mm. However, high resolution also requires sufficiently long source-to-image-receptor distances. Using measured focal spot size and source-to-image-receptor distance, it has been shown that some mammography units do not fully utilize the resolution capabilities of some available screen-film systems. On the other hand, a focal spot that is too small may limit the field size and lead to uneven energy distribution in the x-ray field. 7. In order to perform standard and supplementary mammographic projections, the C arm assembly must be capable of being located at multiple positions within a 180 degrees rotation. Greater degrees of rotation are not required. 8. The heat dissipation capability should be adequate to accommodate the anticipated work-load in terms of number of patients and patient scheduling. A mammographic unit with an anode heat capacity of 90,000 heat units or more should be adequate for most practices. A mammography unit with a heat capacity of 20,000 heat units or less may require waiting time between exposures or studies in a high-volume mammography practice. 9. The compression plate should be made of plastic, usually 1 to 4 mm thick.(ABSTRACT TRUNCATED AT 400 WORDS)

Female↗
PubMed1990

Spatial patterns underlying population differences in the background EEG.

A method is described which can be used to extract common spatial patterns underlying the EEGs from two human populations. These spatial patterns account, in the least-squares sense, maximally for the variance in the EEGs from one population and minimally for the variance in the other population and therefore would seem to be optimal for quantitatively discriminating between the individual EEGs in the two populations. By using this method, it is suggested that the problems associated with the more common approach to discriminating EEGs, significance probability mapping, can be avoided. The method is tested using EEGs from a population of normal subjects and using the EEGs from a population of patients with neurologic disorders. The results in most cases are excellent and the misclassification which occurs in some cases is attributed to the nonhomogeneity of the patient population particularly. The advantages of the method for feature selection, for automatically classifying the clinical EEG, and with respect to the reference-free nature of the selected features are discussed.

Adolescent↗
PubMed1972

Comparison of body weights, organ weights and histological features of selected organs of gnotobiotic, conventional and isolator-reared contaminated pigs.

Twenty-seven pigs from three litters were used in a comparison of body weights, organ weights, and selected histological features of germfree, conventional and isolator-reared contaminated pigs. At three weeks of age conventional pigs were heavier than pigs of the other two groups. The mandibular lymphnodes, stomachs, and small intestines of contaminated pigs were significantly heavier than the same organs of germfree pigs. This difference was not found in superficial inguinal or prefemoral lymph nodes. Other statistically significant organ weight differences were found.Histologically, the lymph nodes of conventional and contaminated pigs were much more active than those of germfree pigs, although secondary nodules were occasionally found in lymph nodes of germfree pigs. Greater quantities of iron-containing pigment were found in the spleens of germfree pigs than in spleens of the other two groups. Hepatic interlobular septa were somewhat more developed in conventional pigs than in germfree or contaminated pigs.

Adrenal Glands↗
PubMed1985

The use of multidimensional perceptual models in the selection of sonar echo features.

The development of an accurate and efficient sonar-target classification system depends upon the identification of a set of signal features which may be used to discriminate important classes of signals. Feature selection can be facilitated through the identification of perceptual features used by human listeners in discriminating relevant sonar echoes. This study was conducted to establish a more reliable means of identifying perceptual features in terms of physical signal parameters as an initial step toward the development of an automatic sonar-target classification system. The results of an experiment involving eight subjects and six sonar echoes are presented. A model of the perceptual structure of these echoes was derived from subject similarity judgments using a multidimensional scaling (MDS) technique. It was found that three perceptual features accounted for the similarity judgments made by the human listeners. Echoes modified along candidate physical dimensions were employed to aid in the identification of perceptual dimensions in terms of physical signal parameters. The three perceptual features could be associated with signal parameters involving the amplitude envelope of the echoes.

Auditory Perception↗
PubMed1977

Two kinds of cognitive deficit associated with chronic schizophrenia.

The performance of 21 chronic schizophrenic patients was investigated on two tests of feature selection. It was found that patients with negative symptoms (muteness, withdrawal, etc.) were characterized by an extreme lack of persistence, but selected usual features; whereas patients with positive symptoms (hallucinations, delusions, etc.) had a normal degree of persistence, but selected unusual features.

Aged↗