SEARCH · PubMed Health

Results for “Prediction Algorithms”

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 127 records · Page 7Linked to original sources

PubMed2025

Proteomics-enabled learning machine algorithms enhance the prediction of cardiovascular diseases in patients with type 2 diabetes mellitus.

BACKGROUND AND AIMS: Estimating the risk of cardiovascular disease (CVD) complications in type 2 diabetes mellitus (T2DM) patients is critical in the medical decision-making process. This study aimed to use a machine learning technique combined with proteomics to develop personalized models for predicting CVD in patients with T2DM. METHODS AND RESULTS: In total, 874 patients with T2DM and 2,920 Olink proteins obtained from the UK Biobank were used in this study. Proteins were screened using Cox regression and LASSO regression. A basic model containing clinical features and a full model combining proteome and clinical features were constructed using the random survival forest algorithm. The area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate the predictive performance of the models and compare them with other CVD predictive models. Compared with the basic model, the full model performed better in predicting CVD, with time-dependent AUCs of 0.81 (3 years), 0.74 (5 years) and 0.74 (10 years) (0.77, 0.69 and 0.67). We calculated the risk scores of the Framingham, ASCVD and Score2-Diabetes models. The results revealed that the prediction performance of the full model was also better than that of the abovementioned models. In terms of differentiation accuracy, the results of the net reclassification improvement index and integrated discrimination improvement index showed that the full model can identify high-risk individuals more accurately (accuracy rate: 79% vs. 69%). CONCLUSIONS: Proteomics can be used to predict cardiovascular complications in diabetic patients. It is also necessary to consider the applicability of the model due to the limitations of the sample size and the constraints of proteomics in clinical applications.

Humans↗

PubMed1991

A predictive and feedback control algorithm maintains a constant glucose concentration in fed-batch fermentations.

A combined predictive and feedback control algorithm based on measurements of the concentration of glucose on-line has been developed to control fed-batch fermentations of Escherichia coli. The predictive control algorithm was based on the on-line calculation of glucose demand by the culture and plotting a linear regression to the next datum point to obtain a predicted glucose demand. This provided a predictive "coarse" control for the glucose-based nutrient feed. A direct feedback control using a proportional controller, based on glucose measurements every 2 min, fine-tuned the feed rate. These combined control schemes were used to maintain glucose concentrations in fed-batch fermentations as tight as 0.49 +/- 0.04 g/liter during growth of E. coli to high cell densities.

Acetates↗

PubMed2025

Effectidor II: a pan-genomic AI-based algorithm for the prediction of type III secretion system effectors.

MOTIVATION: Type III secretion systems are used by many Gram-negative bacteria to inject type 3 effectors (T3Es) directly into eukaryotic cells, promoting disease or provoking immune response. Because of these opposing evolutionary forces, T3E repertoires often vary within taxonomic groups. Identifying the full effector gene repertoire in genomes of related individuals is crucial for determining core and specialized effectors, understanding the disease dynamics, and developing appropriate management strategies against pathogens. It can also help uncover novel T3Es that have recently emerged in a population. Our previously published Effectidor web server successfully addressed the challenge of identifying T3Es in a single bacterial genome. Here, we enriched the web server with various novel capabilities, including the identification of T3Es from multiple genome sequences simultaneously. RESULTS: We present Effectidor II, a web server that relies on machine learning to predict T3E-encoding genes within bacterial pan-genomes. We demonstrate the benefit of learning based on features extracted from the entire sequences comprising the pan-genome and report a novel T3E discovered by it in Xanthomonas euroxanthea. AVAILABILITY AND IMPLEMENTATION: Effectidor II is available at: https://effectidor.tau.ac.il and the source code is available at: https://github.com/naamawagner/Effectidor. A stand-alone version of Effectidor II is available at: https://github.com/naamawagner/Effectidor/tree/StandAlone. The source code for the standalone version and the data used in this work are also provided in https://doi.org/10.5281/zenodo.15081636.

Type III Secretion Systems↗

PubMed1988

Selection of class I MHC-restricted peptides with the strip-of-helix hydrophobicity algorithm.

A strip-of-helix hydrophobicity algorithm to predict class II MHC-restricted peptides, on the basis of their structural similarity to an amphipathic, alpha-helix in Ii, also predicted peptides which were presented to cytotoxic T-cells by class I MHC molecules. This algorithm ranked peptides according to mean Kyte-Doolittle hydrophobicity values of amino acids at positions n, n + 4, n + 7, n + 11, n + 14 and n + 18 in a sequence which when coiled as a putative alpha-helix, had the indicated residues in an axial strip along one side of the helix. Sequences selected for highly scoring, hydrophobic strips were required to have at least 1 of the 4 adjacent strips scoring more negatively than -1 in the strip-of-helix hydrophobicity index and the entire sequence could contain no prolines. This algorithm predicted the class I MHC-restricted, T-cell-presented peptides in sequences of 4 proteins from which some class I MHC-restricted, T-cell-presented sequences had been experimentally determined. Since both class I and class II MHC-restricted peptides could be identified with this algorithm, one can propose that: (1) foreign peptide-binding sites (desetopes) of the class I and class II MHC molecules are structurally similar; and (2) any one T-cell-presented peptide can be presented by some specific allele of both a class I and a class II MHC antigen.

Algorithms↗

PubMed1989

Structure and conformational analysis of lipid-associating peptides of apolipoprotein B-100 produced by trypsinolysis.

Apolipoprotein B-100 (apo B-100) contains putative lipid-associating regions that are, in part, responsible for its overall structure in human plasma low-density lipoproteins. Some of these regions have been identified by reassembly of the total tryptic peptides of apo B-100 with bovine brain sphingomyelin, 1-palmitoyl-2-oleoyl phosphatidylcholine (POPC) and dimyristoylphosphatidylcholine (DPMC). Although more than 500 tryptic peptides are predicted from the known number of arginines and lysines in apo B-100, significant amounts of only 13 peptides spontaneously associate with all three phospholipids. These peptides share some structural characteristics, as predicted by several algorithms, that distinguish them from the water-soluble apolipoproteins. Most apolipoproteins associate with lipids via amphipathic helices and are highly helical in native and reassembled lipoproteins. Analysis of all apo B-100 lipophilic peptides by circular dichroism and by use of a predictive algorithm reveals no evidence of amphipathic helices. Although the predictive algorithm suggested that the lipophilic peptides of apo B-100 contain the sequence determinants for beta-sheet, no spectroscopic evidence for this structure was found. We conclude that the lipophilic regions of apo B-100 liberated by trypsinolysis are highly hydrophobic, although their secondary structures do not fit any simple model.

Amino Acid Sequence↗

PubMed1989

Prediction of protein helices with a derivative of the strip-of-helix hydrophobicity algorithm.

The strip-of-helix hydrophobicity algorithm was devised to identify protein sequences which, when coiled as alpha or 3(10) helices, had one axial, hydrophobic strip and otherwise variably hydrophilic residues. The strip-of-helix hydrophobicity algorithm also ranked such sequences according to an index, the mean hydrophobicity of amino acids in the axial strip. This algorithm well predicted T cell-presented fragments of antigenic proteins. A derivative of this algorithm (the structural helices algorithm (SHA] was tested for the prediction of helices in crystallographically defined proteins. For the SHA, eight amino acid sequences, 2 cycles plus one amino acid in an alpha helix, with strip-of-helix hydrophobicity indices greater than 2.5, were selected with overlapping segments joined. These selections were terminated according to simple "capping rules," which took into account the roles of N-terminal Asn or Pro and C-terminal Gly in the stability of helices. In analyses of 35 crystallographically defined proteins with known alpha and 3(10) helices, the predictions with the SHA overlapped (had overlap indices x greater than or equal to 0.5) with 34% of known helices, touched (had overlap indices 0.5 greater than x greater than 0) or overlapped with 66% of known helices, or were neighboring (came within 6 residues) or touched or overlapped with 82% of known helices. At each level of judging the quality of prediction, the SHA was usually less sensitive (correct predictions/total number of known helices) and more efficient (correct predictions/total number of predictions) than the Chou-Fasman and Garnier-Robson methods. It was simpler in design and calculation. The chemical mechanisms underlying these algorithms appear to apply both to protein folding and to selection of T cell-presented antigenic sequences.

Algorithms↗

PubMed1985

[Prediction of secondary structures of nucleic acids: algorithmic and physical aspects].

Prediction of secondary structures in nucleic acids requires both an adequate physical model and powerful calculation algorithms. In our approach, we cut the molecules in sections of which the contributions to the global energy are context-dependent but roughly additive. The structure of minimum energy is obtained by a tree search under constraints of binary incompatibilities. Our algorithm of the "incompatibility islets" is shown to be more powerful than the "bit parallel forward checking" algorithm, well known in Artificial Intelligence. Recurrent algorithms, proposed by other authors are even more rapid, but often miss the correct structures, for they demand a strict additivity of the energetic contributions, physically unjustified. New strategies, required to deal with molecules of more than 200 nucleotides are discussed. Our physical model has been improved by considering the special case of internal loops beginning with a G-A opposition. A bonus of 1.5 kcal. is attributed to such a feature, at each side of an internal loop. To illustrate our programs, we give the computed schemes for the 3' termini of the small subunit ribosomal RNA.

Base Sequence↗

PubMed1987

Computer aids to mammographic diagnosis.

The improvement of mammographic specificity was investigated by means of identifying specific radiological features. Data are presented on the first 500 patients studied who had previously undergone mammography followed by biopsy. The presence of specific mammographic features on each radiograph, first determined by retrospective examination, was entered into a computer database. Subsequent discriminant function analysis demonstrated the importance of a small number of features whose presence could be used in an algorithm to predict diagnostic outcome. Using this algorithm, this feature-identification approach correctly identified 87.6% of benign and 79% of malignant cases. Specificity was improved to 88% as compared with the original radiological diagnosis of 49%. It is argued that this approach is very promising and a computer-assisted diagnosis based on these findings is described.

Algorithms↗

PubMed1987

Prediction of immunodominant helper T cell antigenic sites from the primary sequence.

We have used a data base of 23 known immunodominant helper T cell antigenic sites located on 12 proteins to systematically develop an optimized algorithm for predicting T cell antigenic sites. The algorithm is based on the amphipathic helix model in which antigenic sites are postulated to be helices with one face predominantly polar and the opposite face predominantly apolar. Such amphipathic structures can form when the polarity of residues along the sequence varies with a more or less regular period. Hence they can be identified by methods (so called power spectrum procedures) that detect periodic variations in properties of a sequence. The choice of power spectrum procedure, hydrophobicity scale, and model parameters are examined. An algorithm is tested by comparing the predicted amphipathic segments with the locations of the known T cell sites, counting the number of matches, and calculating the probability of getting this number by chance alone. The optimum algorithm, which predicts the largest number of sites with the lowest chance probability, uses the Fauchere-Pliska hydrophobicity scale and a least squares fit of a sinusoid as its power spectrum procedure. By applying this algorithm, 18 of the 23 known sites are identified (75% sensitivity) with a high degree of significance (p less than 0.001). The success of the algorithm supports the hypothesis that stable amphipathic helices are fundamentally important in determining immunodominance. This approach may be of practical value in designing synthetic vaccines aimed at T cell immunity.

Algorithms↗

PubMed1992

Echocardiographic algorithms for admission and predischarge prediction of mortality in acute myocardial infarction.

To develop improved prognostic algorithms for routine bedside use in acute myocardial infarction (AMI), the prognostic value concerning 2- and 12-month mortality of an early (within 72 hours after AMI) resting echocardiogram was defined in 201 consecutive patients. The relation between (1) the clinical variables (age, sex, prior and repeat AMI, arrhythmias, cardiac arrest, early [less than 72 hours after AMI] and late heart failure, early and maximal in-hospital Killip class, and maximal creatine kinase-MB isoenzyme), (2) early myocardial performance by echocardiography, and (3) mortality was characterized by Kaplan-Meier survival curves and receiver-operating characteristic curves based on Cox regression model. Only age and clinical heart failure in terms of the maximal in-hospital Killip class had independent predictive value of death (p less than 0.05) when an early echocardiographic estimate of left ventricular ejection fraction (LVEF) was included in the multivariate statistical models. The following 2 optimized algorithms for admission and predischarge calculation of risk of mortality at 2 and 12 months were developed based on the Cox model, using combinations of age, maximal Killip class and early echocardiographic LVEF: mortality at 2 months = 1 - exp - [0.051 x exp [0.044 x (age -60) - (0.117 x (LVEF - 40)]]; and mortality at 1 year = 1 - exp - [0.101 x exp [0.408 x (maxKillip - 1) - (0.061 x (LVEF - 40)]]. Discriminative power for prediction of mortality of the predischarge algorithm in an independent population of 195 patients 5 days after AMI compared favorably with that obtained in the original population, confirming the validity of the proposed method of prognostication.

Adult↗

PubMed1991

Protein topology prediction through constraint-based search and the evaluation of topological folding rules.

An algorithm for predicting protein alpha/beta-sheet topologies from secondary structure and topological folding rules (constraints) has been developed and implemented in Prolog. This algorithm (CBS1) is based on constraint satisfaction and employs forward pruned breadth-first search and rotational invariance. CBS1 showed a 37-fold increase in efficiency over an exhaustive generate and test algorithm giving the same solution for a typical sheet of five strands whose topology was predicted from secondary structure with four topological folding constraints. Prolog specifications of a range of putative protein folding rules were then used to (i) replicate published protein topology predictions and (ii) validate these rules against known protein structures of nucleotide-binding domains. This demonstrated that (i) manual techniques for topology prediction can lead to non-exhaustive search and (ii) most of these protein folding principles were violated by specific proteins. Various extensions to the algorithm are discussed.

Algorithms↗

PubMed2026

How not to be seen: predicting unseen enzyme functions using contrastive learning.

MOTIVATION: Predicting enzyme function from its sequence is still an unsolved problem in the life sciences. Moreover, with the explosion of annotated genome data, we are inundated with potential enzymatic sequences that have not yet been biochemically characterized. While it is not possible to assign a not-yet-existing label to such a sequence, there is high value in placing the sequence as accurately as possible in known function space. Doing so can help provide more accurate falsifiable hypotheses for experimentalists wishing to characterize enzymes from specific functional families. RESULTS: Here we present a contrastive learning algorithm for predicting enzyme function from sequence. Our method, EnzPlacer, predicts the third, second, and first EC numbers for a protein whose fourth EC number is not in the training corpus. This novel prediction mechanism accurately places a protein sequence within a narrowed-down functional context, even if the precise function remains unknown. AVAILABILITY AND IMPLEMENTATION: EnzPlacer and data is available at https://github.com/drxiangma/EnzPlacer under a GPL3 license.

Enzymes↗

PubMed2026

Investigating cross-organism prediction of prokaryotic essential proteins using unsupervised language model and ensemble strategy.

Cross-organism prediction of essential proteins is a critical task for drug discovery and microbial engineering, yet the generalizability of existing machine learning models across diverse species remains a significant challenge. In this study, we propose DeepPEP, a large language model-based framework designed to reliably transfer essential protein annotations between distantly related organisms. Utilizing 66 curated prokaryotic datasets, we systematically evaluated DeepPEP's cross-organism performance under various conditions. Initial pairwise predictions revealed a correlation between performance and evolutionary distance; however, further investigation demonstrated that integrating training data from multiple organisms yields superior predictive power. In a benchmark scenario designed to simulate real-world applications, DeepPEP outperformed the state-of-the-art tool Geptop 2.0, showcasing a robust ability to identify species-specific essential proteins. Finally, a case study on novel genomes confirmed the model's practical effectiveness. Our results suggest that DeepPEP is a powerful strategy for prokaryotic essential protein prediction, and the rigorous evaluation framework established in this study provides a new benchmark for the field.

Large Language Models↗

PubMed2026

Clinical Variable-Based Machine Learning for Predicting Early mCRPC Using Exclusively Clinical Variables: Development and Multicenter External Validation.

BACKGROUND AND OBJECTIVE: Metastatic hormone-sensitive prostate cancer (mHSPC) exhibits heterogeneous progression patterns, with early progression to metastatic castration-resistant prostate cancer (mCRPC) within 12 months indicating aggressive tumor biology and poor prognosis. Current risk stratification tools (CHAARTED, LATITUDE) offer limited individualized prediction. Machine learning approaches are increasingly applied to predict prostate cancer progression, but most models show modest performance (AUC 0.68-0.72), limited external validation, or require genomic variables unavailable in routine practice. This study aimed to develop and externally validate a novel RINH algorithm for predicting early mCRPC progression (≤ 12 months) using exclusively clinical variables, positioning it as a superior alternative to conventional ML classifiers. METHODS: This multicenter study enrolled 412 patients with de novo mHSPC from seven Spanish academic centers using mixed retrospective-prospective data collection. Twenty clinical variables were recorded, including demographics, PSA, ISUP grade, metastatic localization, CHAARTED/LATITUDE classifications, and treatment modalities. Following RINH-based outlier exclusion (55 patients), 357 patients (29 with early progression, 8.1%) were used to train six ML algorithms: RINH, Logistic Regression, Linear Discriminant, Support Vector Machine, Random Forest, and Subspace Discriminant. A two-tiered validation strategy integrated stratified fivefold cross-validation across all centers and formal external validation using center 1 (n = 121, 19 events) for training and centers 2-7 (n = 207, 10 events) for independent testing. Performance metrics included AUC, sensitivity, specificity, accuracy, and F1-score. KEY FINDINGS AND LIMITATIONS: Artificial intelligence and machine learning (ML) are transforming oncology, promising personalized risk stratification beyond traditional clinical criteria. In metastatic hormone-sensitive prostate cancer (mHSPC), early progression to castration resistance (mCRPC) within 12 months signals aggressive biology and poor prognosis, yet current tools (CHAARTED, LATITUDE) offer limited individualized prediction. Multiple ML models have been proposed with variable success: most achieve modest performance (AUC 0.68-0.72), lack robust external validation, or rely on genomic variables inaccessible in routine practice. We propose a novel approach using the Rivality Index Neighborhood (RINH) algorithm, demonstrating superior predictive capacity in an initial multicenter validation with exclusively clinical variables. This study provides rigorous multicenter external validation, advancing toward implementable precision oncology tools. CONCLUSIONS AND CLINICAL IMPLICATIONS: The RINH algorithm achieves superior predictive performance for early mCRPC progression using exclusively clinical variables, representing a significant advance toward implementable risk stratification. However, low reliability scores in external validation underscore that excellent performance metrics alone do not guarantee stability. Before clinical deployment, validation in substantially larger cohorts with higher progression events is essential. If validated, this model could enable personalized, risk-adapted therapeutic strategies, refining patient selection for treatment intensification or de-escalation.

Humans↗

PubMed1987

Computer-assisted predictions of signal peptidase processing sites.

Computer programs are presented which incorporate 2 different algorithms for predicting the site of signal peptide cleavage for eukaryotic preproteins. These programs can be used to identify sites of signal peptidase cleavage of putative preproteins to facilitate the design and interpretation of signal peptide mutagenesis experiments, and to engineer artificial prepolypeptides.

Amino Acids↗

PubMed2026

Development and Validation of a Predictive Model for Identification of Cognitive Impairment Risk in Older Adults with Subjective Cognitive Decline：A Longitudinal Study.

BACKGROUND: Subjective cognitive decline (SCD) is a transitional state between objective cognitive impairment and cognitively intact mental status, providing a critical window for implementing preventive interventions to delay objective cognitive decline. AIMS: We aimed to develop a predictive model for SCD progression in older adults with mild cognitive impairment (MCI). This model will facilitate the identification of risk factors and establishment of targeted interventions for community-based SCD management. METHODS: Data from the China Health and Retirement Longitudinal Study (CHARLS) was utilized in this study, extracting 18 indicators. Potential predictors selected through univariate Cox regression and LASSO regression analyses were sequentially incorporated into a multivariable Cox regression model. A nomogram was constructed to establish a predictive model. Model validation encompassed Area Under Curve (AUC) metrics for discriminative capacity, complemented by quantitative assessments using calibration curve analysis for precision verification and decision curve analysis (DCA) for clinical utility evaluation. RESULTS: A total of 1099 older adults with SCD were included in the final analysis, of whom 114 (10.3%) developed MCI. Multivariable Cox regression identified residence, marital status, educational level, social participation, gait speed, and baseline cognitive function. The model demonstrated time-dependent AUC values of 0.885, 0.830, 0.839, and 0.836 in the training set when evaluating discriminative capacity at 2-, 4-, 7-, and 9-year, respectively. The predictive model showed excellent predictive ability according to AUC, calibration curve, and DCA. CONCLUSIONS: A predictive model was created to estimate the risk of developing MCI in older individuals with SCD, offering clinician-actionable intervention benchmarks for preventive care.

Humans↗

PubMed2026

Machine learning for population-level risk prediction of future cholangiocarcinoma.

BACKGROUND: The poor prognosis of cholangiocarcinoma (CCA) is largely driven by rapid, asymptomatic disease progression, which usually results in a late diagnosis in the absence of established screening strategies. An early, cost-effective, and universally applicable risk assessment strategy would therefore be valuable. METHODS: We developed machine learning (ML) models on prospective, multimodal data from 487,495 UK Biobank (UKB) participants, of whom 649 developed CCA during follow-up. Data from England (80%) were utilised for ML development via five-fold cross-validation, and then all models were tested on withheld data from Scotland, Wales, and Newcastle (20%). Iterative ablation studies reduced inputs from >150 features across demographic data, lifestyle, health records, blood parameters, genomics, and metabolomics to models built on five and ten routinely available clinical parameters. These were externally validated in the Penn Medicine Biobank (PMBB; n = 2638; 28 CCA), All of Us Research Program (AOU; n = 330,433; 362 CCA), Japan Medical Data Centre Claims Database (JMDC; n = 8,425,522; 723 CCA) and TriNetX (n = 728,886; 1592 CCA). FINDINGS: We show that ML models integrating biliary-disease associated health records and Gamma glutamyltransferase can stratify risk of future CCA. Evaluation on the UKB test set as well as three independent cohorts revealed robust performance and generalisability across ethnicities. We achieved AUROCs of 0.71 [95% CI: 0.703-0.711], 0.77 [95% CI: 0.764-0.778 ], 0.796 [95% CI: 0.795-0.798] and 0.8 [95% CI: 0.794-0.805] for UKB, PMBB, AOU, and JMDC respectively, with respective AUPRCs of 0.014 [95% CI: 0.009-0.018], 0.042 [95% CI: 0.037-0.048], 0.038 [95% CI: 0.033-0.042] and 0.001 [95% CI: 0.001-0.001]. In AOU, application of the Youden J-optimised threshold yielded a number needed to screen of 79. Separate models for intra- and extrahepatic CCA did not improve performance. In line with the pathophysiology, performance declined for longer intervals between assessment and event. A group-level analysis in the TriNetX cohort revealed hazard ratios of up to 82.5 [95% CI: 26.4-257.96]. We provide extensive interpretability results and release all source codes used to develop the presented models. INTERPRETATION: We provide a comprehensive framework for early CCA risk stratification in the general population, identifying key predictors, and demonstrating the potential of data-driven models in personalised screening for hepatobiliary cancer. FUNDING: German Cancer Aid (grant #70115730), Junior Principal Investigator Fellowship programme of RWTH Aachen Excellence strategy.

Humans↗

PubMed2026

AI-enabled viral genomics: from virus discovery to host prediction and emerging variant forecasting.

The rapid expansion of metagenomic sequencing has generated vast repositories of viral sequence data that far outpace our capacity to interpret them using conventional approaches. Highly divergent sequences, sparse functional annotation, and taxonomically uneven sampling present fundamental challenges for reference-dependent methods, which lose sensitivity precisely for novel and understudied viruses with high public health relevance. Artificial intelligence (AI) provides a new avenue to address these challenges by enabling predictive inference from viral genomes and proteins while reducing dependence on sequence similarity. In this Review, we discuss representative advances in AI for virus discovery, taxonomic classification and functional annotation, prediction of host range and zoonotic potential, and efforts toward forecasting emerging variants. These advances are transforming viral genomics from a largely descriptive discipline into one with increasing predictive capability. We also critically assess the major challenges that constrain current approaches, including the availability of high-quality and representative datasets, rigorous model evaluation, biological interpretability and responsible governance for increasingly capable AI models.

Artificial Intelligence↗