Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Representation Machine Learning”

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 91 records · Page 5Linked to original sources
PubMed2025

Adapting systems biology to address the complexity of human disease in the single-cell era.

Systems biology aims to achieve holistic insights into the molecular workings of cellular systems through iterative loops of measurement, analysis and perturbation. This framework has had remarkable success in unicellular model organisms, and recent experimental and computational advances - from single-cell and spatial profiling to CRISPR genome editing and machine learning - have raised the exciting possibility of leveraging such strategies to prevent, diagnose and treat human diseases. However, adapting systems-inspired approaches to dissect human disease complexity is challenging, given that discrepancies between the biological features of human tissues and the experimental models typically used to probe function (which we term 'translational distance') can confound insight. Here we review how samples, measurements and analyses can be contextualized within overall multiscale human disease processes to mitigate data and representation gaps. We then examine ways to bridge the translational distance between systems-inspired human discovery loops and model system validation loops to empower precision interventions in the era of single-cell genomics.

Humans↗
PubMedDate not supplied

Cerebellar circuitry as a neuronal machine.

Shortly after John Eccles completed his studies of synaptic inhibition in the spinal cord, for which he was awarded the 1963 Nobel Prize in physiology/medicine, he opened another chapter of neuroscience with his work on the cerebellum. From 1963 to 1967, Eccles and his colleagues in Canberra successfully dissected the complex neuronal circuitry in the cerebellar cortex. In the 1967 monograph, "The Cerebellum as a Neuronal Machine", he, in collaboration with Masao Ito and Janos Szentágothai, presented blue-print-like wiring diagrams of the cerebellar neuronal circuitry. These stimulated worldwide discussions and experimentation on the potential operational mechanisms of the circuitry and spurred theoreticians to develop relevant network models of the machinelike function of the cerebellum. In following decades, the neuronal machine concept of the cerebellum was strengthened by additional knowledge of the modular organization of its structure and memory mechanism, the latter in the form of synaptic plasticity, in particular, long-term depression. Moreover, several types of motor control were established as model systems representing learning mechanisms of the cerebellum. More recently, both the quantitative preciseness of cerebellar analyses and overall knowledge about the cerebellum have advanced considerably at the cellular and molecular levels of analysis. Cerebellar circuitry now includes Lugaro cells and unipolar brush cells as additional unique elements. Other new revelations include the operation of the complex glomerulus structure, intricate signal transduction for synaptic plasticity, silent synapses, irregularity of spike discharges, temporal fidelity of synaptic activation, rhythm generators, a Golgi cell clock circuit, and sensory or motor representation by mossy fibers and climbing fibers. Furthermore, it has become evident that the cerebellum has cognitive functions, and probably also emotion, as well as better-known motor and autonomic functions. Further cerebellar research is required for full understanding of the cerebellum as a broad learning machine for neural control of these functions.

Action Potentials↗
PubMed2025

Privacy-hardened and hallucination-resistant synthetic data generation with logic-solvers.

MOTIVATION: Machine-generated or synthetic data is a valuable resource for training artificial intelligence algorithms, evaluating rare workflows, and sharing data under stricter data legislations. However, current statistical and deep learning methods struggle with large data volumes, are prone to hallucinating scenarios incompatible with reality, and seldom quantify privacy meaningfully. RESULTS: Here, we introduce Genomator, a logic solving approach (SAT solving), which efficiently produces private and realistic representations of the original data. We demonstrate the method on genomic data, which arguably is the most complex and private information. We benchmark Genomator against state-of-the-art methodologies (Markov generation, Wasserstein Generative Adversarial Network and Conditional Restricted Boltzmann Machines), demonstrating a 40%-530% accuracy improvement and 57%-172% higher privacy. Genomator is also 3-100 times more efficient, making it the only tested method that scales to whole genomes. We show the universal trade-off between privacy and accuracy, and use Genomator's tuning capability to cater to all applications along the spectrum, from provable private representations of sensitive cohorts, to datasets with indistinguishable pharmacogenomic profiles. Demonstrating the production-scale generation of tuneable synthetic genomes hold great potential for balancing underrepresented populations in medical research and advancing global data exchange. AVAILABILITY AND IMPLEMENTATION: Genomator is available at https://github.com/csiro/genomator.

Algorithms↗
PubMed1994

A shape-based machine learning tool for drug design.

Building predictive models for iterative drug design in the absence of a known target protein structure is an important challenge. We present a novel technique, Compass, that removes a major obstacle to accurate prediction by automatically selecting conformations and alignments of molecules without the benefit of a characterized active site. The technique combines explicit representation of molecular shape with neural network learning methods to produce highly predictive models, even across chemically distinct classes of molecules. We apply the method to predicting human perception of musk odor and show how the resulting models can provide graphical guidance for chemical modifications.

Algorithms↗
PubMed2006

Predicting the effect of missense mutations on protein function: analysis with Bayesian networks.

BACKGROUND: A number of methods that use both protein structural and evolutionary information are available to predict the functional consequences of missense mutations. However, many of these methods break down if either one of the two types of data are missing. Furthermore, there is a lack of rigorous assessment of how important the different factors are to prediction. RESULTS: Here we use Bayesian networks to predict whether or not a missense mutation will affect the function of the protein. Bayesian networks provide a concise representation for inferring models from data, and are known to generalise well to new data. More importantly, they can handle the noisy, incomplete and uncertain nature of biological data. Our Bayesian network achieved comparable performance with previous machine learning methods. The predictive performance of learned model structures was no better than a naïve Bayes classifier. However, analysis of the posterior distribution of model structures allows biologically meaningful interpretation of relationships between the input variables. CONCLUSION: The ability of the Bayesian network to make predictions when only structural or evolutionary data was observed allowed us to conclude that structural information is a significantly better predictor of the functional consequences of a missense mutation than evolutionary information, for the dataset used. Analysis of the posterior distribution of model structures revealed that the top three strongest connections with the class node all involved structural nodes. With this in mind, we derived a simplified Bayesian network that used just these three structural descriptors, with comparable performance to that of an all node network.

Algorithms↗
PubMed2006

Gene functional annotation by statistical analysis of biomedical articles.

BACKGROUND: Functional annotation of genes is an important task in biology since it facilitates the characterization of genes relationships and the understanding of biochemical pathways. The various gene functions can be described by standardized and structured vocabularies, called bio-ontologies. The assignment of bio-ontology terms to genes is carried out by means of applying certain methods to datasets extracted from biomedical articles. These methods originate from data mining and machine learning and include maximum entropy or support vector machines (SVM). PURPOSE: The aim of this paper is to propose an alternative to the existing methods for functionally annotating genes. The methodology involves building of classification models, validation and graphical representations of the results and reduction of the dimensions of the dataset. METHODS: Classification models are constructed by Linear discriminant analysis (LDA). The validation of the models is based on statistical analysis and interpretation of the results involving techniques like hold-out samples, test datasets and metrics like confusion matrix, accuracy, recall, precision and F-measure. Graphical representations, such as boxplots, Andrew's curves and scatterplots of the variables resulting from the classification models are also used for validating and interpreting the results. RESULTS: The proposed methodology was applied to a dataset extracted from biomedical articles for 12 Gene Ontology terms. The validation of the LDA models and the comparison with the SVM show that LDA (mean F-measure 75.4%) outperforms the SVM (mean F-measure 68.7%) for the specific data. CONCLUSION: The application of certain statistical methods can be beneficial for functional gene annotation from biomedical articles. Apart from the good performance the results can be interpreted and give insight of the bio-text data structure.

Abstracting and Indexing↗
PubMed2006

Learning beyond finite memory in recurrent networks of spiking neurons.

We investigate possibilities of inducing temporal structures without fading memory in recurrent networks of spiking neurons strictly operating in the pulse-coding regime. We extend the existing gradient-based algorithm for training feedforward spiking neuron networks, SpikeProp (Bohte, Kok, & La Poutré, 2002), to recurrent network topologies, so that temporal dependencies in the input stream are taken into account. It is shown that temporal structures with unbounded input memory specified by simple Moore machines (MM) can be induced by recurrent spiking neuron networks (RSNN). The networks are able to discover pulse-coded representations of abstract information processing states coding potentially unbounded histories of processed inputs. We show that it is often possible to extract from trained RSNN the target MM by grouping together similar spike trains appearing in the recurrent layer. Even when the target MM was not perfectly induced in a RSNN, the extraction procedure was able to reveal weaknesses of the induced mechanism and the extent to which the target machine had been learned.

Action Potentials↗
PubMed2006

Classification of fMRI independent components using IC-fingerprints and support vector machine classifiers.

We present a general method for the classification of independent components (ICs) extracted from functional MRI (fMRI) data sets. The method consists of two steps. In the first step, each fMRI-IC is associated with an IC-fingerprint, i.e., a representation of the component in a multidimensional space of parameters. These parameters are post hoc estimates of global properties of the ICs and are largely independent of a specific experimental design and stimulus timing. In the second step a machine learning algorithm automatically separates the IC-fingerprints into six general classes after preliminary training performed on a small subset of expert-labeled components. We illustrate this approach in a multisubject fMRI study employing visual structure-from-motion stimuli encoding faces and control random shapes. We show that: (1) IC-fingerprints are a valuable tool for the inspection, characterization and selection of fMRI-ICs and (2) automatic classifications of fMRI-ICs in new subjects present a high correspondence with those obtained by expert visual inspection of the components. Importantly, our classification procedure highlights several neurophysiologically interesting processes. The most intriguing of which is reflected, with high intra- and inter-subject reproducibility, in one IC exhibiting a transiently task-related activation in the 'face' region of the primary sensorimotor cortex. This suggests that in addition to or as part of the mirror system, somatotopic regions of the sensorimotor cortex are involved in disambiguating the perception of a moving body part. Finally, we show that the same classification algorithm can be successfully applied, without re-training, to fMRI collected using acquisition parameters, stimulation modality and timing considerably different from those used for training.

Algorithms↗
PubMed1988

Knowledge-based educational systems.

In knowledge-based educational systems, the key concept is that information and procedures are represented in the same data structure. These structures can search for each other in flexible and, consequently, very robust ways. At the Air Force Human Resources Laboratory (AFHRL), our researchers are building computer environments that know what they know, know how people can best use them, and know how to draw inferences about their state--self-referential electronic tutors. In September 1986, artificial intelligence researchers participated in AFHRL's Research Planning Forum for Intelligent Tutorial Systems (ITS). This essay reviews the state of the philosophy, art, and science of artificial intelligence (AI) approaches to education. Then it summarizes the research issues which were presented, discussed, and better defined in this Forum--namely the nature and representation of 1) expertise modules, 2) student diagnostic modules, 3) adaptive instructional and curriculum modules, 4) instructional environments, and 5) man-machine interfaces. Advances in artificial intelligence, cognitive science, and instructional discourse have provided a means for investigating human learning, for representing an individual's own "knowledge processing." Research and development in knowledge-based educational systems seems promising, not only for helping people learn how to perform complex tasks, but also for explicitly expressing how people learn to learn. Therefore, would it not be wise to establish a scientific legacy for the development of effective knowledge-based tutorial systems which is informed by the best studies of mind and meaning, language and thought, purpose and paradox?

Artificial Intelligence↗
PubMed1993

Constructive induction and protein tertiary structure prediction.

To date, the only methods that have been used successfully to predict protein structures have been based on identifying homologous proteins whose structures are known. However, such methods are limited by the fact that some proteins have similar structure but no significant sequence homology. We consider two ways of applying machine learning to facilitate protein structure prediction. We argue that a straightforward approach will not be able to improve the accuracy of classification achieved by clustering by alignment scores alone. In contrast, we present a novel constructive induction approach that learns better representations of amino acid sequences in terms of physical and chemical properties. Our learning method combines knowledge and search to shift the representation of sequences so that semantic similarity is more easily recognized by syntactic matching. Our approach promises not only to find new structural relationships among protein sequences, but also expands our understanding of the roles knowledge can play in learning via experience in this challenging domain.

Artificial Intelligence↗
PubMed2025

Efficient Detection and Characterization of Targets of Natural Selection Using Transfer Learning.

Natural selection leaves detectable patterns of altered spatial diversity within genomes, and identifying affected regions is crucial for understanding species evolution. Recently, machine learning approaches applied to raw population genomic data have been developed to uncover these adaptive signatures. Convolutional neural networks (CNNs) are particularly effective for this task, as they handle large data arrays while maintaining element correlations. However, shallow CNNs may miss complex patterns due to their limited capacity, while deep CNNs can capture these patterns but require extensive data and computational power. Transfer learning addresses these challenges by utilizing a deep CNN pretrained on a large dataset as a feature extraction tool for downstream classification and evolutionary parameter prediction. This approach reduces extensive training data generation requirements and computational needs while maintaining high performance. In this study, we developed TrIdent, a tool that uses transfer learning to enhance detection of adaptive genomic regions from image representations of multilocus variation. We evaluated TrIdent across various genetic, demographic, and adaptive settings, in addition to unphased data and other confounding factors. TrIdent demonstrated improved detection of adaptive regions compared to recent methods using similar data representations. We further explored model interpretability through class activation maps and adapted TrIdent to infer selection parameters for identified adaptive candidates. Using whole-genome haplotype data from European and African populations, TrIdent effectively recapitulated known sweep candidates and identified novel cancer, and other disease-associated genes as potential sweeps.

Selection, Genetic↗
PubMed2026

Identification and Classification of Expressed Orphan Genes, Spurious Orphan Genes, and Conserved Genes in the Human Gut Microbiome.

Orphan genes (OGs)-genes lacking detectable homologs outside a species-are widespread in microbial genomes and are thought to contribute to their adaptation and molecular innovation. However, not all predicted OGs may represent novel functional coding sequences. False positive OGs, also called spurious OGs, can arise from gene prediction errors. We reason that OGs lacking detectable expression are more likely to be spurious. To test this, we combined large-scale metatranscriptomic profiling of the human gut microbiome with machine learning to distinguish expressed OGs from spurious ones and compare them with conserved genes (CGs) found in multiple species. Using nearly 5,000 metatranscriptome libraries, we identified ∼218,000 OGs supported by expression evidence, while ∼330,000 predicted OGs lacked detectable expression and were classified as spurious. We extracted 154 features for sequence, structural, and evolutionary properties for each gene and trained XGBoost classifiers while accounting for genomic representation. The models achieved an area under the receiver operating characteristic curve (AUC) of 0.82 in distinguishing expressed OGs from spurious OGs and an AUC of 0.93 in distinguishing expressed OGs from CGs. Interpretation based on SHAP (SHapley Additive exPlanations) revealed clear biological signals. Particularly, expressed orphans were present in more genomes than spurious ones, and expressed OGs were shorter than CGs. This work improves OG discovery and suggests that expressed OGs differ systematically from CGs and spurious OGs in sequence composition, structural constraints, and evolutionary signals.

Humans↗
PubMed2025

Integration of single cell multiomics data by deep transfer hypergraph neural network.

Multi-omics characterization of individual cells offers remarkable potential for analyzing the dynamics and relationships of gene regulatory states across millions of cells. How to integrate multimodal data is an open problem, existing integration methods struggle with accuracy and modality-specific biological variation retention. In this paper, we present scHyper (scalable, interpretable machine learning for single cell integration), a low-code and data-efficient deep transfer model designed for integrating paired and unpaired single-cell multimodal data. We benchmark scHyper against datasets from different multimodal data. ScHyper learns a low-dimensional representation and aligns the covariance matrices of the measured modalities, achieving high accuracy even with large scale atlas-level datasets with low memory and computational time across different cell lines, shedding light on regulatory relationships between different types of omics. Altogether, we show that scHyper is a versatile and robust tool for cell-type label transfer and integration from multimodal single-cell datasets.

Single-Cell Analysis↗
PubMed2006

Bioinformatics and its impact on clinical research methods. Findings from the Section on Bioinformatics.

OBJECTIVES: To summarize current excellent research in the field of bioinformatics. METHOD: Synopsis of the articles selected for the IMIA Yearbook 2006. RESULTS: Current research in the field of bioinformatics clearly shows ongoing unification of experimental findings and clinical outcomes. Microarray data, gene sequences and clinical data are more and more perceived as different but related facets of one entity. Significant work is done in the area of text and data mining in order to bring together patient data and biochemical phenomena by means of ontologies. A strong trend in the clinical field is performance of exhaustive studies on DNA material derived from patients that suffer from diseases that are already known to be inherited. Examination of appropriate methods covers data and text mining, ontologies as well as machine learning and classification. CONCLUSIONS: The best paper selection of articles on bioinformatics shows examples of excellent research on methods used for studying inherited diseases and their underlying genetic dispositions. Clinical studies, inclusion of experimental findings like microarray data, and of knowledge representation formats all lead to a better understanding the linkage between gene sequences, biological functions and clinical findings in the form of healthy state or physiological disorders.

Awards and Prizes↗
PubMed2002

Conserved codon composition of ribosomal protein coding genes in Escherichia coli, Mycobacterium tuberculosis and Saccharomyces cerevisiae: lessons from supervised machine learning in functional genomics.

Genomics projects have resulted in a flood of sequence data. Functional annotation currently relies almost exclusively on inter-species sequence comparison and is restricted in cases of limited data from related species and widely divergent sequences with no known homologs. Here, we demonstrate that codon composition, a fusion of codon usage bias and amino acid composition signals, can accurately discriminate, in the absence of sequence homology information, cytoplasmic ribosomal protein genes from all other genes of known function in Saccharomyces cerevisiae, Escherichia coli and Mycobacterium tuberculosis using an implementation of support vector machines, SVM(light). Analysis of these codon composition signals is instructive in determining features that confer individuality to ribosomal protein genes. Each of the sets of positively charged, negatively charged and small hydrophobic residues, as well as codon bias, contribute to their distinctive codon composition profile. The representation of all these signals is sensitively detected, combined and augmented by the SVMs to perform an accurate classification. Of special mention is an obvious outlier, yeast gene RPL22B, highly homologous to RPL22A but employing very different codon usage, perhaps indicating a non-ribosomal function. Finally, we propose that codon composition be used in combination with other attributes in gene/protein classification by supervised machine learning algorithms.

Algorithms↗
PubMed2004

Tools for the body (schema).

What happens in our brain when we use a tool to reach for a distant object? Recent neurophysiological, psychological and neuropsychological research suggests that this extended motor capability is followed by changes in specific neural networks that hold an updated map of body shape and posture (the putative "Body Schema" of classical neurology). These changes are compatible with the notion of the inclusion of tools in the "Body Schema", as if our own effector (e.g. the hand) were elongated to the tip of the tool. In this review we present empirical support for this intriguing idea from both single-neuron recordings in the monkey brain and behavioural performance of normal and brain-damaged humans. These relatively simple neural and behavioural aspects of tool-use shed light on more complex evolutionary and cognitive aspects of body representation and multisensory space coding for action.

Animals↗
PubMed2001

A method for automated temporal knowledge acquisition applied to sleep-related breathing disorders.

This paper presents a method for the discovery of temporal patterns in multivariate time series and their conversion into a linguistic knowledge representation applied to sleep-related breathing disorders. The main idea lies in introducing several abstraction levels that allow a step-wise identification of temporal patterns. Self-organizing neural networks are used to discover elementary patterns in the time series. Machine learning (ML) algorithms use the results of the neural networks to automatically generate a rule-based description. At the next levels, temporal grammatical rules are inferred. This method covers one of the main "bottlenecks" in the design of knowledge-based systems, namely, the knowledge acquisition problem. An evaluation of the rules lead to an overall sensitivity of 0.762, and a specificity of 0.758.

Artificial Intelligence↗
PubMed2003

Learning optimized features for hierarchical models of invariant object recognition.

There is an ongoing debate over the capabilities of hierarchical neural feedforward architectures for performing real-world invariant object recognition. Although a variety of hierarchical models exists, appropriate supervised and unsupervised learning methods are still an issue of intense research. We propose a feedforward model for recognition that shares components like weight sharing, pooling stages, and competitive nonlinearities with earlier approaches but focuses on new methods for learning optimal feature-detecting cells in intermediate stages of the hierarchical network. We show that principles of sparse coding, which were previously mostly applied to the initial feature detection stages, can also be employed to obtain optimized intermediate complex features. We suggest a new approach to optimize the learning of sparse features under the constraints of a weight-sharing or convolutional architecture that uses pooling operations to achieve gradual invariance in the feature hierarchy. The approach explicitly enforces symmetry constraints like translation invariance on the feature set. This leads to a dimension reduction in the search space of optimal features and allows determining more efficiently the basis representatives, which achieve a sparse decomposition of the input. We analyze the quality of the learned feature representation by investigating the recognition performance of the resulting hierarchical network on object and face databases. We show that a hierarchy with features learned on a single object data set can also be applied to face recognition without parameter changes and is competitive with other recent machine learning recognition approaches. To investigate the effect of the interplay between sparse coding and processing nonlinearities, we also consider alternative feedforward pooling nonlinearities such as presynaptic maximum selection and sum-of-squares integration. The comparison shows that a combination of strong competitive nonlinearities with sparse coding offers the best recognition performance in the difficult scenario of segmentation-free recognition in cluttered surround. We demonstrate that for both learning and recognition, a precise segmentation of the objects is not necessary.

Learning↗