SEARCH · PubMed Health

Results for “machine learning prediction”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 541 records · Page 30Linked to original sources

PubMed2002

Evaluating machine learning approaches for aiding probe selection for gene-expression arrays.

MOTIVATION: Microarrays are a fast and cost-effective method of performing thousands of DNA hybridization experiments simultaneously. DNA probes are typically used to measure the expression level of specific genes. Because probes greatly vary in the quality of their hybridizations, choosing good probes is a difficult task. If one could accurately choose probes that are likely to hybridize well, then fewer probes would be needed to represent each gene in a gene-expression microarray, and, hence, more genes could be placed on an array of a given physical size. Our goal is to empirically evaluate how successfully three standard machine-learning algorithms-naïve Bayes, decision trees, and artificial neural networks-can be applied to the task of predicting good probes. Fortunately it is relatively easy to get training examples for such a learning task: place various probes on a gene chip, add a sample where the corresponding genes are highly expressed, and then record how well each probe measures the presence of its corresponding gene. With such training examples, it is possible that an accurate predictor of probe quality can be learned. RESULTS: Two of the learning algorithms we investigate-naïve Bayes and neural networks-learn to predict probe quality surprisingly well. For example, in the top ten predicted probes for a given gene not used for training, on average about five rank in the top 2.5% of that gene's hundreds of possible probes. Decision-tree induction and the simple approach of using predicted melting temperature to rank probes perform significantly worse than these two algorithms. The features we use to represent probes are very easily computed and the time taken to score each candidate probe after training is minor. Training the naïve Bayes algorithm takes very little time, and while it takes over 10 times as long to train a neural network, that time is still not very substantial (on the order of a few hours on a desktop workstation). We also report the information contained in the features we use to describe the probes. We find the fraction of cytosine in the probe to be the most informative feature. We also find, not surprisingly, that the nucleotides in the middle of the probes sequence are more informative than those at the ends of the sequence.

Algorithms↗

PubMed2026

Genome-wide association, polygenic risk scores, and machine learning for chronic post-surgical pain risk stratification: A UK biobank study.

Chronic post-surgical pain is a prevalent and debilitating complication following surgery, representing a clinical challenge. Despite the established heritability of pain phenotypes, large-scale genetic studies remain limited. This study aimed to identify genetic variants associated with chronic post-surgical pain, develop polygenic risk scores, and integrate these with clinical features for risk prediction. UK Biobank data from 47,836 participants (2490 cases and 45,346 controls) were split into training (80%; n = 38,268) and validation (20%; n = 9568) sets prior to analysis. A genome-wide association study was conducted on the training set only, across 19 million variants, and polygenic risk scores were constructed and integrated with clinical features in a logistic regression framework. Two close, rare, imputed signals crossed the genome-wide significance threshold but lacked local linkage-disequilibrium support, while 220 variants crossed the suggestive threshold. In the held-out validation set, cases had higher mean polygenic risk scores than controls (0.138 vs. -0.021; Cohen's d = 0.16, p < 0.001). A logistic regression model integrating clinical features and polygenic risk scores achieved an area under the curve of 0.639 (95% CI: 0.583-0.693), higher than models using either feature set alone. The polygenic risk score for chronic post-surgical pain was among the most important predictors. Risk stratification revealed the top quartile had 3.84-fold higher odds of chronic post-surgical pain than the bottom quartile (95% CI: 2.00-7.37). These findings suggest a possible modest genetic contribution to chronic post-surgical pain. Polygenic risk scores may complement clinical factors in surgical risk stratification. PERSPECTIVE: Chronic post-surgical pain may have a modest genetic contribution. This UK Biobank study identified over 220 variants at suggestive significance and constructed a polygenic risk score that was significantly elevated in cases. A combined clinical-genomic model achieved a 3.84-fold difference in odds across predicted-risk quartiles.

Chronic post-surgical pain↗

PubMed2025

CSGL: chemical synthesis graph learning for molecule representation.

MOTIVATION: Molecule representation learning (MRL) translates molecules into a real vector space, serving as input to downstream tasks in biology, chemistry, and computer science. This article introduces a chemical synthesis graph learning (CSGL) framework, which enhances MRL by considering both the atomic structures of molecules and their roles in chemical reactions through a hierarchical graph representation. Specifically, molecules are first modeled based on their molecular graphs, which capture atomic-level structural information. They are then further refined using a chemical synthesis graph, where nodes represent reactant and product molecule sets, and edges encode chemical transformations between reactants and products (e.g. changes in molecular structures). CSGL optimizes molecular embeddings of reactant and product nodes in a fashion that ensures the embeddings conform to a chemical balance constraint. RESULTS: Experimental results show that our method CSGL achieves strong performance on a variety of tasks, including product prediction, reaction classification, and molecular property prediction. AVAILABILITY AND IMPLEMENTATION: https://github.com/li-2023/CSGL.

Machine Learning↗

PubMed2002

Self-optimizing MPC of melt temperature in injection moulding.

The parameters in plastic injection moulding are highly nonlinear and interacting. Good control of plastic melt temperature for injection moulding is very important in reducing operator setup time, assuring consistent product quality, and preventing thermal degradation of the melt. Step response testing was performed on the barrel heating zones on an industrial injection moulding machine (IMM). The open loop responses indicated a high degree of process coupling between the heating zones. From these experimental step responses, a multiple-input-multiple-output model predictive control strategy was developed and practically implemented. The requirement of negligible overshoot is important to the plastics industry for preventing material overheating and wastage, and reducing machine operator setup time. A generic learning and self-optimizing MPC methodology was developed and implemented on the IMM to control melt temperature for any polymer to be moulded on any machine having different electrical heater capacities. The control performance was tested for varying setpoint trajectories typical of normal machine operations. The results showed that the predictive controller provided good control of melt temperature for all zones with negligible oscillations, and, therefore, eliminated material degradation and extended machine setup time.

Computer Simulation↗

PubMed2006

Predicting transmembrane beta-barrels and interstrand residue interactions from sequence.

Transmembrane beta-barrel (TMB) proteins are embedded in the outer membrane of Gram-negative bacteria, mitochondria, and chloroplasts. The cellular location and functional diversity of beta-barrel outer membrane proteins (omps) makes them an important protein class. At the present time, very few nonhomologous TMB structures have been determined by X-ray diffraction because of the experimental difficulty encountered in crystallizing transmembrane proteins. A novel method using pairwise interstrand residue statistical potentials derived from globular (nonouter membrane) proteins is introduced to predict the supersecondary structure of transmembrane beta-barrel proteins. The algorithm transFold employs a generalized hidden Markov model (i.e., multitape S-attribute grammar) to describe potential beta-barrel supersecondary structures and then computes by dynamic programming the minimum free energy beta-barrel structure. Hence, the approach can be viewed as a "wrapping" component that may capture folding processes with an initiation stage followed by progressive interaction of the sequence with the already-formed motifs. This approach differs significantly from others, which use traditional machine learning to solve this problem, because it does not require a training phase on known TMB structures and is the first to explicitly capture and predict long-range interactions. TransFold outperforms previous programs for predicting TMBs on smaller (<or=200 residues) proteins and matches their performance for straightforward recognition of longer proteins. An exception is for multimeric porins where the algorithm does perform well when an important functional motif in loops is initially identified. We verify our simulations of the folding process by comparing them with experimental data on the functional folding of TMBs. A Web server running transFold is available and outputs contact predictions and locations for sequences predicted to form TMBs.

Amino Acid Sequence↗

PubMed2006

Weighted quality estimates in machine learning.

MOTIVATION: Machine learning methods such as neural networks, support vector machines, and other classification and regression methods rely on iterative optimization of the model quality in the space of the parameters of the method. Model quality measures (accuracies, correlations, etc.) are frequently overly optimistic because the training sets are dominated by particular families and subfamilies. To overcome the bias, the dataset is usually reduced by filtering out closely related objects. However, such filtering uses fixed similarity thresholds and ignores a part of the training information. RESULTS: We suggested a novel approach to calculate prediction model quality based on assigning to each data point inverse density weights derived from the postulated distance metric. We demonstrated that our new weighted measures estimate the model generalization better and are consistent with the machine learning theory. The Vapnik-Chervonenkis theorem was reformulated and applied to derive the space-uniform error estimates. Two examples were used to illustrate the advantages of the inverse density weighting. First, we demonstrated on a set with a built-in bias that the unweighted cross-validation procedure leads to an overly optimistic quality estimate, while the density-weighted quality estimates are more realistic. Second, an analytical equation for weighted quality estimates was used to derive an SVM model for signal peptide prediction using a full set of known signal peptides, instead of the usual filtered subset.

Algorithms↗

PubMed2005

Prediction of the phenotypic effects of non-synonymous single nucleotide polymorphisms using structural and evolutionary information.

MOTIVATION: There has been great expectation that the knowledge of an individual's genotype will provide a basis for assessing susceptibility to diseases and designing individualized therapy. Non-synonymous single nucleotide polymorphisms (nsSNPs) that lead to an amino acid change in the protein product are of particular interest because they account for nearly half of the known genetic variations related to human inherited diseases. To facilitate the identification of disease-associated nsSNPs from a large number of neutral nsSNPs, it is important to develop computational tools to predict the phenotypic effects of nsSNPs. RESULTS: We prepared a training set based on the variant phenotypic annotation of the Swiss-Prot database and focused our analysis on nsSNPs having homologous 3D structures. Structural environment parameters derived from the 3D homologous structure as well as evolutionary information derived from the multiple sequence alignment were used as predictors. Two machine learning methods, support vector machine and random forest, were trained and evaluated. We compared the performance of our method with that of the SIFT algorithm, which is one of the best predictive methods to date. An unbiased evaluation study shows that for nsSNPs with sufficient evolutionary information (with not <10 homologous sequences), the performance of our method is comparable with the SIFT algorithm, while for nsSNPs with insufficient evolutionary information (<10 homologous sequences), our method outperforms the SIFT algorithm significantly. These findings indicate that incorporating structural information is critical to achieving good prediction accuracy when sufficient evolutionary information is not available. AVAILABILITY: The codes and curated dataset are available at http://compbio.utmem.edu/snp/dataset/

Algorithms↗

PubMed2005

Computational searches for splicing signals.

The removal of introns from pre-mRNA requires as an initial event the accurate molecular recognition of the proper exon-intron borders. It is now evident that RNA sequence elements in addition to the consensus splice site sequences themselves are required for this recognition. Genomic analyses have contributed to the definition of these elements as exonic and intronic splicing enhancers and silencers, comprising what has been called the "splicing code." Many computational methods have been brought to bear in such studies. We describe here some of the methods we have used to discover functional splicing signals. What these methods have in common is a comparison of sequences in and around exons to sequences found elsewhere in the genome. We have especially made use of comparisons to "pseudo exons," intronic sequences resembling exons by virtue of being bounded by sequences indistinguishable from splice sites. Two computational strategies are emphasized: (1) the use of a machine learning technique in which a computational algorithm, a support vector machine, is first trained on known examples and then used to predict sequences associated with splicing; and (2) straight statistical analysis of differences between regions associated with exons and other regions in the genome. In most cases, the predictions made using these methods have been validated by subsequent empirical tests. An attempt has been made to make this description understandable by researchers unfamiliar with computational practice and to include practical references to specific databases and programs.

Alternative Splicing↗

PubMed1996

Nonlinear quantitative structure-activity relationship for the inhibition of dihydrofolate reductase by pyrimidines.

A novel method for quantitative structure-activity relationship (QSAR) analysis is presented. The method, which does not assume any particular functional form for the QSAR, develops nonlinear relationships between parameters describing a set of molecules and the activity of the molecules. For the QSAR of the inhibition of Escherichia coli dihydrofolate reductase by 2,4-diamino-5-(substituted benzyl)pyrimidines, the method compares favorably to other nonlinear methods. Cross-validation trials demonstrate that the predictive ability is as accurate as other methods, and the method is simpler and faster than neural network and machine-learning methods. Consequently, its implementation is much easier, and interpretation of the generated QSAR is more straightforward.

Escherichia coli↗

PubMed2002

Biological information: making it accessible and integrated (and trying to make sense of it).

The availability of the genome sequences of human and mouse, human sequence variation data and other large genetic data sets will lead to a revolution in understanding of the human machine and the treatment of its diseases. The success of the international genome sequencing consortiums shows what can be achieved by well coordinated large scale public domain projects and the benefits of data access to all. It is already clear that the availability of this sequence is having a huge impact on research worldwide. Complete genome sequences provide a framework to pull all biological data together such that each piece has the potential to say something about biology as a whole. Biology is too complex for any organisation to have a monopoly of ideas or data, so the collection, analysis and access to this data can be contributed to by research institutes around the world. However, although it is possible for all this data to be accessible to all through the internet, the more organisations provide data or analysis separately, the harder it becomes for anyone to collect and integrate the results. To address these problems of intergration of data, open standards for biological data exchange, such as the 'Distributed Annotation System' (DAS) are being developed and bioinformatics (Dowell et al., 2001) as a whole is now being strongly driven by the open source software (OSS) model for collaborative software development (Hubbard and Birney, 1999). The leading provider of human genome annotation, the Ensembl project (http://www.ensembl.org), is entirely an OSS project and has been widely adopted by academic and commerical organisations alike (Hubbard et al., 2002). Accurate automatic annotation of features such as genes in vertebrate genomes currently relies on supporting evidence in the form of homologies to mRNAs, ESTs or protein. However, it appears that sufficient high quality experimentally curated annotation now exists to be used as a substrate for machine learning algorithms to create effective models of biological signal sequences (Down and Hubbard, 2002). Is there hope for ab initio prediction methods after all?

Chromosome Mapping↗

PubMed2000

The 2000 Olympic Games of protein structure prediction; fully automated programs are being evaluated vis-à-vis human teams in the protein structure prediction experiment CAFASP2.

In this commentary, we describe two new protein structure prediction experiments being run in parallel with the CASP experiment, which together may be regarded as the 2000 Olympic Games of structure prediction. The first new experiment is CAFASP, the Critical Assessment of Fully Automated Structure Prediction. In CAFASP, the participants are fully automated programs or Internet servers, and here the automated results of the programs are evaluated, without any human intervention. The second new experiment, named LiveBench, follows the CAFASP ideology in that it is aimed towards the evaluation of automatic servers only, while it runs on a large set of prediction targets and in a continuous fashion. Researchers will be watching the 2000 protein structure prediction Olympic Games, to be held in December, in order to learn about the advances in the classical 'human-plus-machine' CASP category, the fully automated CAFASP category, and the comparison between the two.

Amino Acid Sequence↗

PubMed2006

The forecast of the postoperative survival time of patients suffered from non-small cell lung cancer based on PCA and extreme learning machine.

In this paper, a new effective model is proposed to forecast how long the postoperative patients suffered from non-small cell lung cancer will survive. The new effective model which is based on the extreme learning machine (ELM) and principal component analysis (PCA) can forecast successfully the postoperative patients' survival time. The new model obtains better prediction accuracy and faster convergence rate which the model using backpropagation (BP) algorithm and the Levenberg-Marquardt (LM) algorithm to forecast the postoperative patients' survival time can not achieve. Finally, simulation results are given to verify the efficiency and effectiveness of our proposed new model.

Algorithms↗

PubMed2025

FluxRETAP: a REaction TArget Prioritization genome-scale modeling technique for selecting genetic targets.

MOTIVATION: Metabolic engineering is rapidly evolving as a result of new advances in synthetic biology tools and automation platforms that enable high throughput strain construction, as well as the development of machine learning tools (ML) for biology. However, selecting genetic engineering targets that effectively guide the metabolic engineering process is still challenging. ML can provide predictive power for synthetic biology, but current technical limitations prevent the independent use of ML approaches without previous biological knowledge. RESULTS: Here, we present FluxRETAP, a simple and computationally inexpensive method that leverages the prior mechanistic knowledge embedded in genome-scale models for suggesting targets for genetic overexpression, downregulation or deletion, with the final goal of increasing the production of a desired metabolite. This method can provide a list of desirable engineering targets that can be combined with current ML pipelines. FluxRETAP captured 100% of reaction targets experimentally verified to improve Escherichia coli isoprenol production, 50% of targets that experimentally improved taxadiene production in E. coli and ∼60% of genetic targets from a verified minimal constrained cut-set in Pseudomonas putida, while providing additional high priority targets that could be tested. Overall, FluxRETAP is an efficient algorithm for identifying a prioritized list of testable genetic and reaction targets. AVAILABILITY AND IMPLEMENTATION: FluxRETAP is implemented in python and released under the creative commons license. The implementation and code are freely available at: https://github.com/JBEI/FluxRETAP.

Escherichia coli↗

PubMed2026

Senescent fibroblasts drive CD8+ T cell dysfunction in colorectal cancer via CD36-mediated lipid transfer and peroxidation.

BACKGROUND: Functional exhaustion of tumor-infiltrating CD8+ T cells represents a hallmark of colorectal cancer (CRC) immunosuppression, though its mechanistic drivers remain elusive. Given the established correlation between CRC progression and stromal senescence characterized by pathological lipid accumulation and impaired immunity, we investigated whether and how senescent fibroblasts actively regulate CD8+ T cell dysfunction. METHODS: Single-cell RNA sequencing (scRNA-seq) analysis was conducted to unveil the diverse fibroblast populations and the significant lipid metabolism changes between senescent fibroblasts and non-senescent fibroblasts in human CRC specimens and adjacent normal mucosa. Machine-learning identified senescent fibroblasts with a distinct gene signature. Cell-cell communication analysis was used to evaluate the interactions between senescent fibroblasts and CD8+ T cells in colorectal cancer. Co-culture experiments were conducted among senescent fibroblasts, CD8+ T cells and patient-derived organoids of CRC (CRC-PDOs), with the results evaluated with high-content imaging and propidium iodide/Hoechst 33,342 staining. Flow cytometry, ELISA and lipid pulse-chase with BODIPY FL C16 were performed to detect the alterations of CD8+ T cell cytotoxic function and metabolic status. AOM/DSS-induced CRC mouse model was used to conduct in vivo validation to evaluate whether senolytics could suppress CRC progression. Patients from the Cancer Genome Atlas colorectal cancer cohort were stratified into CD36-high and CD36-low groups by median expression, and drug sensitivity for GDSC2 compounds was predicted computationally using the oncoPredict R package. RESULTS: ScRNA-seq demonstrated the specific cell population presence and divergence of senescent fibroblasts between neoplastic and histologically normal adjacent cell clusters in CRC. Random Forest was employed for cell senescence classification. Feature importance analysis identified five genes as key contributors to the model’s decision process. Cell-cell communication analysis revealed enhanced interactions between senescent fibroblasts and CD8+ T cells in CRC. Co-culture of senescent fibroblasts significantly impaired the cytotoxic functions of CD8+ T cells on CRC-PDOs, which was reflected by the declined proportions of granzyme B (GZMB) + and interferon gamma (IFNγ) + CD8+ T cells and enhanced viability of CRC-PDOs. Mechanistically, the co-culture with senescent fibroblasts promoted the lipid shuttling into CD8+ T cells to induce lipid peroxidation and downstream impairment of cytotoxicity. Furthermore, the inhibition of CD36, the specific scavenger receptor for lipid uptake of CD8+ T cells, effectively suppressed lipid transfer and peroxidation thereby preserving the effector functions of CD8+ T cells and ultimately promoting tumor apoptosis. Complementarily, in vivo senolytic treatment significantly suppressed CRC progression in AOM-DSS CRC mouse models. Top 12 therapeutic agents were identified significantly enhanced predicted efficacy in CD36-high tumors. CONCLUSIONS: Our study identified a substantial population of senescent fibroblasts in human CRC through single cell transcriptomics, machine-learning and clinical biopsies. These senescent fibroblasts impair CD8+ T cell-mediated killing of CRC-PDOs via CD36-dependent lipid transfer, suggesting senolytic targeting of stromal cells as a promising immunotherapeutic strategy for CRC.

Colorectal Neoplasms↗

PubMed2006

On the nature of cavities on protein surfaces: application to the identification of drug-binding sites.

In this article we introduce a new method for the identification and the accurate characterization of protein surface cavities. The method is encoded in the program SCREEN (Surface Cavity REcognition and EvaluatioN). As a first test of the utility of our approach we used SCREEN to locate and analyze the surface cavities of a nonredundant set of 99 proteins cocrystallized with drugs. We find that this set of proteins has on average about 14 distinct cavities per protein. In all cases, a drug is bound at one (and sometimes more than one) of these cavities. Using cavity size alone as a criterion for predicting drug-binding sites yields a high balanced error rate of 15.7%, with only 71.7% coverage. Here we characterize each surface cavity by computing a comprehensive set of 408 physicochemical, structural, and geometric attributes. By applying modern machine learning techniques (Random Forests) we were able to develop a classifier that can identify drug-binding cavities with a balanced error rate of 7.2% and coverage of 88.9%. Only 18 of the 408 cavity attributes had a statistically significant role in the prediction. Of these 18 important attributes, almost all involved size and shape rather than physicochemical properties of the surface cavity. The implications of these results are discussed. A SCREEN Web server is available at http://interface.bioc.columbia.edu/screen.

Binding Sites↗

PubMed2005

An SVM-based system for predicting protein subnuclear localizations.

BACKGROUND: The large gap between the number of protein sequences in databases and the number of functionally characterized proteins calls for the development of a fast computational tool for the prediction of subnuclear and subcellular localizations generally applicable to protein sequences. The information on localization may reveal the molecular function of novel proteins, in addition to providing insight on the biological pathways in which they function. The bulk of past work has been focused on protein subcellular localizations. Furthermore, no specific tool has been dedicated to prediction at the subnuclear level, despite its high importance. In order to design a suitable predictive system, the extraction of subtle sequence signals that can discriminate among proteins with different subnuclear localizations is the key. RESULTS: New kernel functions used in a support vector machine (SVM) learning model are introduced for the measurement of sequence similarity. The k-peptide vectors are first mapped by a matrix of high-scored pairs of k-peptides which are measured by BLOSUM62 scores. The kernels, measuring the similarity for sequences, are then defined on the mapped vectors. By combining these new encoding methods, a multi-class classification system for the prediction of protein subnuclear localizations is established for the first time. The performance of the system is evaluated with a set of proteins collected in the Nuclear Protein Database (NPD). The overall accuracy of prediction for 6 localizations is about 50% (vs. random prediction 16.7%) for single localization proteins in the leave-one-out cross-validation; and 65% for an independent set of multi-localization proteins. This integrated system can be accessed at http://array.bioengr.uic.edu/subnuclear.htm. CONCLUSION: The integrated system benefits from the combination of predictions from several SVMs based on selected encoding methods. Finally, the predictive power of the system is expected to improve as more proteins with known subnuclear localizations become available.

Algorithms↗

PubMed2006

STAR: predicting recombination sites from amino acid sequence.

BACKGROUND: Designing novel proteins with site-directed recombination has enormous prospects. By locating effective recombination sites for swapping sequence parts, the probability that hybrid sequences have the desired properties is increased dramatically. The prohibitive requirements for applying current tools led us to investigate machine learning to assist in finding useful recombination sites from amino acid sequence alone. RESULTS: We present STAR, Site Targeted Amino acid Recombination predictor, which produces a score indicating the structural disruption caused by recombination, for each position in an amino acid sequence. Example predictions contrasted with those of alternative tools, illustrate STAR'S utility to assist in determining useful recombination sites. Overall, the correlation coefficient between the output of the experimentally validated protein design algorithm SCHEMA and the prediction of STAR is very high (0.89). CONCLUSION: STAR allows the user to explore useful recombination sites in amino acid sequences with unknown structure and unknown evolutionary origin. The predictor service is available from http://pprowler.itee.uq.edu.au/star.

Algorithms↗

PubMed2006

Applying hybrid reasoning to mine for associative features in biological data.

We develop the means to mine for associative features in biological data. The hybrid reasoning schema for deterministic machine learning and its implementation via logic programming is presented. The methodology of mining for correlation between features is illustrated by the prediction tasks for protein secondary structure and phylogenetic profiles. The suggested methodology leads to a clearer approach to hierarchical classification of proteins and a novel way to represent evolutionary relationships. Comparative analysis of Jasmine and other statistical and deterministic systems (including Explanation-Based Learning and Inductive Logic Programming) are outlined. Advantages of using deterministic versus statistical data mining approaches for high-level exploration of correlation structure are analyzed.

Algorithms↗