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.

SourceResearch connectionFrom yearThrough year

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

At least 775 records · Page 43Linked to original sources

PubMed2026

Benchmark of biomarker identification and prognostic modeling methods on diverse censored data.

The practices of identifying biomarkers and developing prognostic models using genomic data has become increasingly prevalent. Such data often features characteristics that make these practices difficult, namely high dimensionality, correlations between predictors, and sparsity. Many modern methods have been developed to address these problematic characteristics while performing feature selection and prognostic modeling, but a large-scale comparison of their performances in these tasks on diverse right-censored time to event data (aka survival time data) is much needed. We have compiled many existing methods, including some machine learning methods, several which have performed well in previous benchmarks, primarily for comparison in regards to variable selection capability, and secondarily for survival time prediction on many synthetic datasets with varying levels of sparsity, correlation between predictors, and signal strength of informative predictors. For illustration, we have also performed multiple analyses on a publicly available and widely used cancer cohort from The Cancer Genome Atlas using these methods. We evaluated the methods through extensive simulation studies in terms of the false discovery rate, F1-score, concordance index, Brier score, root mean square error, and computation time. Of the methods compared, CoxBoost and the Adaptive LASSO performed well in all metrics, and the LASSO and elastic net excelled when evaluating concordance index and F1-score. The Benjamini-Hoschberg and q-value procedures showed volatile performances in controlling the false discovery rate. Some methods' performances were greatly affected by differences in the data characteristics. With our extensive numerical study, we have identified the best performing methods for a plethora of data characteristics using informative metrics. This will help cancer researchers in choosing the best approach for their needs when working with genomic data.

Humans↗

PubMed2004

Genetic algorithms for classification of olfactory stimulants.

We have developed and tested a genetic algorithm (GA) for pattern recognition, which identifies molecular descriptors that optimize the separation of the activity classes of olfactory stimulants in a plot of the two or three largest principal components of the data. Because principal components maximize variance, the bulk of the information encoded by these descriptors is about differences between olfactory classes in the dataset. In addition, the GA focuses on those classes and or samples that are difficult to classify as it trains using a form of boosting to modify the fitness landscape. Boosting minimizes the problem of convergence to a local optimum, because the fitness function of the GA is changing as the population is evolving toward a solution. Over time, compounds that consistently classify correctly are not as heavily weighted in the analysis as compounds that are difficult to classify. The pattern recognition GA learns its optimal parameters in a manner similar to a neural network. The algorithm integrates aspects of both strong and weak learning to yield a "smart" one-pass procedure for feature selection and classification.

Algorithms↗

PubMed2005

A hybrid neural network system for prediction and recognition of promoter regions in human genome.

This paper proposes a high specificity and sensitivity algorithm called PromPredictor for recognizing promoter regions in the human genome. PromPredictor extracts compositional features and CpG islands information from genomic sequence, feeding these features as input for a hybrid neural network system (HNN) and then applies the HNN for prediction. It combines a novel promoter recognition model, coding theory, feature selection and dimensionality reduction with machine learning algorithm. Evaluation on Human chromosome 22 was approximately 66% in sensitivity and approximately 48% in specificity. Comparison with two other systems revealed that our method had superior sensitivity and specificity in predicting promoter regions. PromPredictor is written in MATLAB and requires Matlab to run. PromPredictor is freely available at http://www.whtelecom.com/Prompredictor.htm.

Computational Biology↗

PubMed2005

A novel hybrid GA/RBFNN technique for protein sequences classification.

A novel hybrid genetic algorithm (GA)/radial basis function neural network (RBFNN) technique, which selects features from the protein sequences and trains the RBF neural network simultaneously, is proposed in this paper. Experimental results show that the proposed hybrid GA/RBFNN system outperforms the BLAST and the HMMer.

Algorithms↗

PubMed2005

Novel experimental design for steady-state processes: a systematic Bayesian approach for enzymes, drug transport, receptor binding, continuous culture and cell transport kinetics.

We demonstrate that a Bayesian approach (the use of prior knowledge) to the design of steady-state experiments can produce major gains quantifiable in terms of information, productivity and accuracy of each experiment. Developing the use of Bayesian utility functions, we have used a systematic method to identify the optimum experimental designs for a number of kinetic model data sets. This has enabled the identification of trends between kinetic model types, sets of design rules and the key conclusion that such designs should be based on some prior knowledge of the kinetic model. We suggest an optimal and iterative method for selecting features of the design such as the substrate range, number of measurements and choice of intermediate points. The final design collects data suitable for accurate modelling and analysis and minimises the error in the parameters estimated. It is equally applicable to enzymes, drug transport, receptor binding, microbial culture and cell transport kinetics.

Bayes Theorem↗

PubMed2002

What role, if any, for laparoscopic surgery in Crohn's disease of the hindgut?

An outsider to the field of surgery would probably take it for granted that surgeons have a highly developed rationale for choosing a laparoscopic approach to Crohns disease. After all, an increasing number of surgeons are performing laparoscopic surgery for Crohns disease as witnessed by several articles published in the 1990s (Table). In fact this is not quite true. Most papers are case reports or series without controls, capable only of suggesting feasibility. Furthermore, comparison studies often feature selection flaws, and therefore beg the question of whether laparoscopic surgery should or not be considered as standard care. An attempt is made herein to give readers a concise insight of the evidence available in the English language literature. It does not pretend to offer a comprehensive review of the topic rather, it highlights some relevant issues, and then outlines what role, if any, laparoscopic surgery should play in Crohn's disease. There are at least 6 categories for discussion.

Colitis↗

PubMed2000

Familial clustering of arterial blood pressure, HDL cholesterol, and pro-insulin but not of insulin resistance and microalbuminuria in siblings of patients with type 2 diabetes.

OBJECTIVE: To test the hypothesis that selected abnormalities cluster in type 2 diabetic families. Offspring of patients with type 2 diabetes have a 40-60% chance of developing type 2 diabetes and an increased frequency of impaired glucose tolerance (IGT) or unknown diabetes. These offspring also show metabolic abnormalities of type 2 diabetes, such as insulin resistance, high insulin and pro-insulin, low HDL cholesterol levels, arterial hypertension, and microalbuminuria. RESEARCH DESIGN AND METHODS: We studied 87 families including at least one type 2 diabetic patient, i.e., 87 probands and 146 siblings; 60 spouses of probands with no family history of diabetes were compared with siblings. Familial clustering was evaluated by 2 methods: concordance of siblings and probands for a given abnormality (method 1) and intraclass correlation coefficients of values within each family (method 2). RESULTS: At oral glucose tolerance testing, 24 siblings had type 2 diabetes, 31 siblings had IGT, and 14 spouses had IGT (P = 0.0012 vs. siblings). With method 1, familial clustering occurred for microalbuminuria, insulin resistance, arterial hypertension, HDL cholesterol and pro-insulin levels; with method 2, familial clustering was observed for the same variables except for microalbuminuria. With both method 1 and 2, familial clustering for insulin resistance disappeared, whereas familial clustering for arterial blood pressure, HDL cholesterol, and pro-insulin remained after correction for BMI; after further restriction of analysis to probands and to siblings with normal glucose tolerance, familial clustering for pro-insulin was observed only with method 2. CONCLUSIONS: These data indicate that siblings of diabetic patients are at high risk for selected features of type 2 diabetes.

Albuminuria↗

PubMed1997

Comparative physiology of acoustic and allied central analyzers.

To exploit comparisons among classes of vertebrates and invertebrates, and between higher and lower levels of the brain, and between modalities, some important needs and opportunities for new research into the way central processing of acoustic input takes place are pointed out. Most of these are suggested by unfamiliar results on fish and reptiles that call for new controls in mammalian experiments as well as more systematic study of nonmammalian taxa. Three frameworks or basic agendas are outlined; i) systematic comparison of dynamical properties to acoustic variables including especially repetition at different rates and the related states of expectation; ii) comparison of response measures, including especially sequences such as oscillations and measures of assembly cooperativity such as synchrony, coherence and bicoherence; and iii) comparison of auditory subsystems, including especially modal categories such as complex feature selective regions and small sets. Some recent and some new results are summarized on human acoustic and non-mammalian event related potentials (ERPs) in response to expectations. When a regular and frequent standard stimulus is omitted, the omitted stimulus potential (OSP) after conditioning with low repetition rates (long ISIs--1-3 s) is a slow, broad positivity P600-900), previously known. With high rates (ISI < 1 s), a new form of response appears, with fast components (P22), different dynamics and less dependence on attention. Slow and fast OSPs each show a constant peak latency after the due-time of the missing stimulus, as though a temporal expectation has been learned. Unlike the visual OSP we have reported earlier, both fast and slow can occur together in the 1-2 Hz range. Very few conditioning stimuli suffice to create the "expectation" that causes an OSP--only two for the slow type. These and more familiar ERPs, considered in human subjects to index cognitive events, need to be compared in other classes of vertebrates and invertebrates.

Acoustic Stimulation↗

PubMed1987

Employer perspectives of HMO prescription drug benefits.

A survey of employers was undertaken to ascertain the potential role of the prescription drug component in HMO marketing strategies. The study focused on (1) the importance of the drug component in making a HMO attractive to employers and (2) employer evaluations of selected features of HMO drug programs pertaining to: sources of supply, copayment amounts, and quality assurance/cost containment initiatives. While not unimportant to employers, the prescription drug component was of relatively less importance than some other medical services (e.g., outpatient medical, hospital). Employers did perceive differences among alternative features of HMO drug programs.

Analysis of Variance↗

PubMed2026

Discovery and validation of a multi-protein panel for predicting non-fatal major adverse cardiovascular events in diabetic kidney disease.

OBJECTIVE: To identify plasma protein biomarkers associated with incident non-fatal major adverse cardiovascular events (MACE) in diabetic kidney disease (DKD) patients. RESEARCH DESIGN AND METHODS: We analyzed 317 DKD patients from the UK Biobank. Plasma proteomics and clinical data (demographics, metabolism, renal function) were integrated. In an exploratory discovery phase, three sequential Cox regression models (crude, socio-demographic-adjusted, socio-demographic-metabolic adjusted) screened non-fatal MACE-associated proteins. To prevent information leakage, the cohort was then randomly split into training (70%) and testing (30%) sets; machine-learning feature selection, hyperparameter optimization, and final model development were performed exclusively within the training set. The associated proteins were input into the four-step machine-learning pipeline (LASSO-Cox, random survival forest, Boruta, XGBoost-Cox). Predictive performance was validated using Kaplan-Meier survival analyses, longitudinal trajectory modeling, and ROC benchmarking. An interactive web application was deployed for clinical implementation. RESULTS: Of 1,463 plasma proteins, 561 were associated with non-fatal MACE across Cox models, with 14 overlapping proteins. Nine core proteins (ANG, IL1R1, CXCL14, ESAM, PTGDS, HAVCR1, FGFR2, IGSF8, CCL3) were validated: ANG showed the strongest non-fatal MACE association (HR = 3.88, 95%CI 2.33-6.48, p<0.001), and all high-expression groups had elevated non-fatal MACE risk. GO/KEGG enrichment highlighted inflammatory-immune pathways like positive regulation of MAPK cascade, Cytokine-cytokine receptor interaction and PI3K-Akt signaling pathway as key mechanisms. The model integrating proteins, demographic factors, and clinical variables achieved the highest predictive performance across non-fatal MACE (AUC = 0.768), myocardial infarction (MI) (0.808), and stroke (0.816) outcomes, with superior stability in cross-validation. CoxBoost + Elastic Net framework was selected as the optimal framework via benchmarking of 101 algorithms. The model demonstrated favorable calibration in high-risk patients and yielded positive net clinical benefit across decision thresholds of 5% to 45%. The web tool (https://jiangli2941.github.io/MACE-prediction-v2/) enables input of 28 variables, outputs non-fatal MACE risk status, risk probability, and highlights abnormal indicators. CONCLUSION: Plasma proteomics combined with machine learning identifies robust non-fatal MACE predictors in DKD.

Humans↗

PubMed2026

Integrated multi-omics profiling identifies aging-related molecular signatures and convergent interferon signaling in systemic lupus erythematosus.

BACKGROUND: Systemic lupus erythematosus (SLE) is characterized by chronic immune activation and molecular alterations that overlap with aging-related biological processes. However, how these alterations are organized across molecular layers and whether they converge on shared regulatory networks remain incompletely understood. METHODS: We performed an integrative multi-omics analysis combining in-house proteomic and phosphoproteomic data from 130 patients with SLE and 90 healthy controls (HCs) and publicly available transcriptomic datasets comprising 1,461 SLE patients. Proteins and phosphorylation sites were annotated using established aging-related gene resources. Differential protein abundance and phosphorylation changes were analyzed across disease-status and disease-activity comparisons. Nominal P-value thresholds were used for exploratory feature selection, whereas FDR-adjusted P values were used to assess robustness after multiple-testing correction. Kinase-substrate enrichment, transcription factor annotation, and cell-type-resolved transcriptomic comparison were used to explore potential regulatory programs. RESULTS: We identified 128 nominally altered proteins annotated to aging-related biological processes, including genomic instability, mitochondrial dysfunction, and epigenetic alterations. Phosphoproteomic analysis revealed 36 nominally altered phosphorylation sites, including previously unreported sites in IFI16 (S153, S780) and PKCδ (S507, S664). Clustering analysis demonstrated heterogeneous protein co-regulation patterns across disease states. Kinase activity inference suggested altered activity of TBK1 and IKKβ. TF analysis further highlighted STAT1, RELA, and PML as potential central nodes within the inferred regulatory network. Notably, these multi-omic alterations were not randomly distributed but showed convergence toward shared signaling pathways, particularly those related to interferon responses. CONCLUSIONS: This integrative multi-omics study identifies inflammatory and interferon-dominated molecular alterations in SLE PBMCs that overlap with aging-related biological processes and converge on shared regulatory networks. These findings provide a hypothesis-generating framework for investigating the intersection between chronic immune activation and aging-related molecular remodeling in SLE.

Humans↗

PubMed2005

Eye movements are functional during face learning.

In a free viewing learning condition, participants were allowed to move their eyes naturally as they learned a set of new faces. In a restricted viewing learning condition, participants remained fixated in a single central location as they learned the new faces. Recognition of the learned faces was then tested following the two learning conditions. Eye movements were recorded during the free viewing learning condition, as well as during recognition. The recognition results showed a clear deficit following the restricted viewing condition, compared with the free viewing condition, demonstrating that eye movements play a functional role during human face learning. Furthermore, the features selected for fixation during recognition were similar following free viewing and restricted viewing learning, suggesting that the eye movements generated during recognition are not simply a recapitulation of those produced during learning.

Eye Movements↗

PubMed1992

Salience of stimulus and response features in choice-reaction tasks.

A pattern of differential reaction time (RT) benefits obtained in spatial-precuing tasks has been attributed to translation processes that operate on mental codes formed to represent the stimulus and response sets. According to the salient-features coding principle, the codes are based on the salient stimulus and response features, with RTs being fastest when the two sets of features correspond. Three experiments are reported in which the stimulus and response sets were manipulated using Gestalt grouping principles. In the first two experiments, stimuli and responses were grouped according to spatial proximity, whereas in the last experiment, they were grouped according to similarity. With both types of manipulations, the grouping of the stimulus set systematically affected the pattern of precuing benefits. Thus, in these experiments, the organization of the stimulus set was the primary determinant of the features selected for coding the stimulus and response sets in the translation process.

Adult↗

PubMed2000

Metaphor comprehension: a computational theory.

Metaphor comprehension involves an interaction between the meaning of the topic and the vehicle terms of the metaphor. Meaning is represented by vectors in a high-dimensional semantic space. Predication modifies the topic vector by merging it with selected features of the vehicle vector. The resulting metaphor vector can be evaluated by comparing it with known landmarks in the semantic space. Thus, metaphorical prediction is treated in the present model in exactly the same way as literal predication. Some experimental results concerning metaphor comprehension are simulated within this framework, such as the nonreversibility of metaphors, priming of metaphors with literal statements, and priming of literal statements with metaphors.

Algorithms↗

PubMed2006

Tumor classification based on DNA copy number aberrations determined using SNP arrays.

High-density single nucleotide polymorphism (SNP) array is a recently introduced technology that genotypes more than 10,000 human SNPs on a single array. It has been shown that SNP arrays can be used to determine not only SNP genotype calls, but also DNA copy number (DCN) aberrations, which are common in solid tumors. In the past, effective cancer classification has been demonstrated using microarray gene expression data, or DCN data derived from comparative genomic hybridization (CGH) arrays. However, the feasibility of cancer classification based on DCN aberrations determined by SNP arrays has not been previously investigated. In this study, we address this issue by applying state-of-the-art classification algorithms and feature selection algorithms to the DCN aberration data derived from a public SNP array dataset. Performance was measured via leave-one-out cross-validation (LOOCV) classification accuracy. Experimental results showed that the maximum accuracy was 73.33%, which is comparable to the maximum accuracy of 76.5% based on CGH-derived DCN data reported previously in the literature. These results suggest that DCN aberration data derived from SNP arrays is useful for etiology-based tumor classification.

Algorithms↗

PubMed1984

Experimental infection with Rickettsia mooseri and antibody response of adult and newborn laboratory rats.

Quantitative studies of selected features of peripherally induced Rickettsia mooseri (= R. typhi) infection in Rattus norvegicus-derived white laboratory rats revealed a unique association between microbe and amplifying vertebrate host which appears to be especially conducive to maintenance of the enzootic cycle. Both adult and newborn (1-3 days old) rats were highly susceptible to percutaneous infection (ID50 = approximately 1 organism), but neither showed signs of disease or died even when inoculated with 10(4)-10(5) plaque-forming units. Gain in body weight of infected newborn rats was indistinguishable from that of uninfected newborn rats over the first 3 weeks of life. The course of the systemic infection, as measured by the rise and fall of R. mooseri titers in blood, brain and kidney and the serum antibody response, was almost identical in adult and newborn rats. Thus, despite their immaturity in certain immunological processes, newborn rats controlled postnatal R. mooseri infection about as well as did adult rats. The rickettsemic period of about 10 days corresponds to the period of infectivity of inoculated rats for fleas. Rickettsiae were not isolated from blood, brain or kidneys by methods employed for more than 4-5 weeks after infection. Serum antirickettsial antibodies persisted for at least 60 weeks postinfection, i.e., longer than the usual life span of rats in nature and, hence, are a valid measure of the cumulative experience of rat populations with R. mooseri infection.

Age Factors↗

PubMedDate not supplied

Overdose with chloral hydrate: a pharmacological and therapeutic review.

The purpose of this review is to highlight the toxicity of chloral hydrate and to review the management of overdoses with chloral hydrate. Three patients are presented in whom life-threatening cardiac arrhythmias dominated the clinical presentation. These arrhythmias were resistant to standard antiarrhythmic therapy. Also, we have reviewed selected features in eight patients who took overdoses of chloral hydrate who were admitted to an intensive care unit between 1981 and 1988. The pharmacology and toxicology of chloral hydrate are discussed with particular reference to the cardiac arrhythmias that are seen with overdosage. A proposed management scheme is detailed, including intravenously-administered propranolol as the preferred first-line antiarrhythmic agent. A case may be made for the discontinuation of the usage of chloral hydrate.

Adult↗

PubMed2002

Detection and discrimination of relative spatial phase by V1 neurons.

Edge-like and line-like features result from spatial phase congruence, the local phase agreement between harmonic components of a spatial waveform. Psychophysical observations and models of early visual processing suggest that human visual feature detectors are specialized for edge-like and line-like phase congruence. To test whether primary visual cortex (V1) neurons account for such specificity, we made tetrode recordings in anesthetized macaque monkeys. Stimuli were drifting equal-energy compound gratings composed of four sinusoidal components. Eight congruence phases (one-dimensional features) were tested, including line-like and edge-like waveforms. Many of the 137 single V1 neurons (recorded at 45 sites) could reliably signal phase congruence by any of several response measures. Across neurons, the preferred spatial feature had only a modest bias for line-like waveforms. Information-theoretic analysis showed that congruence phase was temporally encoded in the frequency band present in the stimuli. The most sensitive neurons had feature discrimination thresholds that approached psychophysical levels, but typical neurons were substantially less sensitive. In single V1 neurons, feature discrimination exhibited various dependences on the congruence phase of the reference waveform. Simple cells were over-represented among the most sensitive neurons and on average carried twice as much feature information as complex cells. However, the distribution of the indices of optimal tuning and discrimination of relative phase was indistinguishable in simple and complex cells. Our results suggest that phase-sensitive pooling of responses is required to account for human psychophysical performance, although variation in feature selectivity among nearby neurons is considerable.

Action Potentials↗