SEARCH · PubMed Health

Results for “Network inference”

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 523 records · Page 29Linked to original sources

PubMed2000

Mechanics of living cells measured by laser tracking microrheology.

To establish laser-tracking microrheology (LTM) as a new technique for quantifying cytoskeletal mechanics, we measure viscoelastic moduli with wide bandwidth (5 decades) within living cells. With the first subcellular measurements of viscoelastic phase angles, LTM provides estimates of solid versus liquid behavior at different frequencies. In LTM, the viscoelastic shear moduli are inferred from the Brownian motion of particles embedded in the cytoskeletal network. Custom laser optoelectronics provide sub-nanometer and near-microsecond resolution of particle trajectories. The kidney epithelial cell line, COS7, has numerous spherical lipid-storage granules that are ideal probes for noninvasive LTM. Although most granules are percolating through perinuclear spaces, a subset of perinuclear granules is embedded in dense viscoelastic cytoplasm. Over all time scales embedded particles exhibit subdiffusive behavior and are not merely tethered by molecular motors. At low frequencies, lamellar regions (820 +/- 520 dyne/cm(2)) are more rigid than viscoelastic perinuclear regions (330 +/- 250 dyne/cm(2), p < 0.0001), but spectra converge at high frequencies. Although the actin-disrupting agent, latrunculin A, softens and liquefies lamellae, physiological levels of F-actin, alone (11 +/- 1.2 dyne/cm(2)) are approximately 70-fold softer than lamellae. Therefore, F-actin is necessary for lamellae mechanics, but not sufficient. Furthermore, in time-lapse of apparently quiescent cells, individual lamellar granules can show approximately 4-fold changes in moduli that last >10 s. Over a broad range of frequencies (0.1-30, 000 rad/s), LTM provides a unique ability to noninvasively quantify dynamic, local changes in cell viscoelasticity.

Actins↗

PubMed2005

Bayesian inference and attentional modulation in the visual cortex.

The responses of neurons in cortical areas V2 and V4 can be significantly modulated by attention to particular locations within an input image. We show that such effects emerge naturally when perception is viewed as a probabilistic inference process governed by Bayesian principles and implemented in hierarchical cortical networks. The proposed model can explain a rich variety of attention-related responses in cortical area V4 including multiplicative modulation of tuning curves, restoration of neural responses in the presence of distracting stimuli, and influence of attention on neighboring unattended locations. Our results suggest a new interpretation of attention as a cortical mechanism for reducing perceptual uncertainty by combining top-down task-relevant information with bottom-up sensory inputs in a probabilistic manner.

Animals↗

PubMed2026

Decoding age-stratified clinical and molecular heterogeneity in male breast cancer through multiomic profiling.

OBJECTIVE: Age-associated molecular heterogeneity is well described in female breast cancer but remains insufficiently characterized in male breast cancer (MBC). We profiled age-stratified clinical and molecular differences between younger (≤55 years) male breast cancer (YMBC) and older (>55 years) male breast cancer (OMBC). METHODS: We retrospectively analyzed 347 patients with MBC diagnosed at Fudan University Shanghai Cancer Center by integrating clinicopathological data, RNA sequencing, and whole-exome sequencing (WES). Survival, differential expression, and mutational signature analyses were performed. Tumor microenvironment features were inferred using xCell and ESTIMATE, and weighted gene co-expression network analysis (WGCNA) was conducted to identify age-associated co-expression modules. Candidate therapeutics were prioritized using the Genomics of Drug Sensitivity in Cancer (GDSC) resource and evaluated using patient-derived organoids (PDOs). RESULTS: Compared with OMBC, YMBC more frequently had human epidermal growth factor receptor 2 (HER2)-positive status (14.91% vs. 4.02%) and triple-negative tumors (4.92% vs. 1.78%), and had worse 5-year recurrence-free survival (hazard ratio=2.19, P=0.018). Transcriptomic analyses indicated enrichment of neural-related programs and reduced immune-related signaling in YMBC, and xCell/ESTIMATE supported lower immune infiltration. Consistently, WGCNA identified age-associated modules linking neural-related programs with reduced immune infiltration. Immunohistochemistry supported increased perineural invasion and lower CD8+ T cell infiltration in YMBC. GDSC-guided prioritization with PDO testing nominated sepantronium bromide (YM155) as a candidate vulnerability in YMBC. WES showed a higher NBPF10 mutation frequency in YMBC (54.5% vs. 14.3%, P<0.05). CONCLUSIONS: Integrated multi-omics profiling revealed age-stratified clinical and molecular heterogeneity in MBC. YMBC patients demonstrated inferior recurrence-free survival, neural signaling enrichment, an immune-cold microenvironment, and enriched NBPF10 mutations. These findings support age as a meaningful stratification variable in MBC risk assessment and treatment planning, and highlight the need for caution when considering treatment de-escalation in younger patients, while nominating YM155 as a candidate agent for prospective evaluation.

Male breast cancer↗

PubMed1987

Integrating management and expertise in a computerised system for hypertensive patients.

An expert system has been integrated to the data management system of the ARTEMIS programme for hypertensive patients. The patient database, which has been used since 1975, contains the medical records of about 20,000 patients. Information is interactively entered by physicians, nurses and secretaries on video display units. The computerised medical record has replaced the traditional handwritten medical record. The database management system is used to produce different summary reports (inpatient and outpatient care) and personalized recall letters which are mailed to the patients before their appointments. Suggestions provided by the expert system include additional information to be obtained (complementary patient interrogation, biological or radiological investigations, etc.), possible causes of hypertension, and medical prescriptions. The information base allows the description of both static knowledge (in the form of a semantic network) and dynamic knowledge (in the form of production rules). The inference system sequentially uses a combination of forward and backward chaining and performs both exact and approximate reasoning. The diagnostic performance of the expert system was evaluated in 100 cases of hypertension (50 of essential hypertension and 50 of secondary hypertension. Concordance between the diagnosis proposed by the expert system and the one proposed by the specialist was achieved in 92% of secondary hypertension cases and 88% of essential hypertension cases. It is suggested that the integration of data and knowledge management might enhance the overall acceptance by medical staff of a computerised system, and facilitate the validation of a knowledge base.

Database Management Systems↗

PubMed1999

INTERACT: an object oriented protein-protein interaction database.

MOTIVATION: Protein-protein interactions provide vital information concerning the function of proteins, complexes and networks. Currently there is no widely accepted repository of this interaction information. Our aim is to provide a single database with the necessary architecture to fully store, query and analyse interaction data. RESULTS: An object oriented database has been created which provides scientists with a resource for examining existing protein-protein interactions and inferring possible interactions from the data stored. It also provides a basis for examining networks of interacting proteins, via analysis of the data stored. The database contains over a thousand interactions. CONTACT: k.eilbeck@stud.man.ac.uk

Computational Biology↗

PubMed2003

Inference of transcriptional regulation relationships from gene expression data.

MOTIVATION: In order to find gene regulatory networks from microarray data, it is important to first find direct regulatory relationships between pairs of genes. RESULTS: We propose a new method for finding potential regulatory relationships between pairs of genes from microarray time series data and apply it to expression data for cell-cycle related genes in yeast. We compare our algorithm, dubbed the event method, with the earlier correlation method and the edge detection method by Filkov et al. When tested on known transcriptional regulation genes, all three methods are able to find similar numbers of true positives. The results indicate that our algorithm is able to identify true positive pairs that are different from those found by the two other methods. We also compare the correlation and the event methods using synthetic data and find that typically, the event method obtains better results. AVALIABILITY: software is available upon request.

Algorithms↗

PubMed2006

Artificial neural networks can learn to estimate extinction rates from molecular phylogenies.

Molecular phylogenies typically consist of only extant species, yet they allow inference of past rates of extinction, because recently originated species are less likely to be extinct than ancient species. Despite the simple structure of the assumed underlying speciation-extinction process, parametric functions to estimate extinction rates from phylogenies turned out to be complex and often difficult to derive. Moreover, these parametric functions are specific to a particular process (e.g. complete species level phylogeny with constant birth and death rates) and a particular type of data (e.g. times between bifurcations). Here, it is shown that artificial neural networks can substitute for parametric estimation functions once they have been sufficiently trained on simulated data. This technique can in principle be used for different processes and data types, and because it circumvents the time-consuming and difficult task of deriving parametric estimation functions, it may greatly extend the possibilities to make macro-evolutionary inferences from molecular phylogenies. This novel approach is explained, applied to estimate speciation and extinction rates from a molecular phylogeny of the reef fish genus Naso (Acanturidae), and its performance is compared to that of maximum likelihood estimation.

Animals↗

PubMed2006

Protein ranking by semi-supervised network propagation.

BACKGROUND: Biologists regularly search DNA or protein databases for sequences that share an evolutionary or functional relationship with a given query sequence. Traditional search methods, such as BLAST and PSI-BLAST, focus on detecting statistically significant pairwise sequence alignments and often miss more subtle sequence similarity. Recent work in the machine learning community has shown that exploiting the global structure of the network defined by these pairwise similarities can help detect more remote relationships than a purely local measure. METHODS: We review RankProp, a ranking algorithm that exploits the global network structure of similarity relationships among proteins in a database by performing a diffusion operation on a protein similarity network with weighted edges. The original RankProp algorithm is unsupervised. Here, we describe a semi-supervised version of the algorithm that uses labeled examples. Three possible ways of incorporating label information are considered: (i) as a validation set for model selection, (ii) to learn a new network, by choosing which transfer function to use for a given query, and (iii) to estimate edge weights, which measure the probability of inferring structural similarity. RESULTS: Benchmarked on a human-curated database of protein structures, the original RankProp algorithm provides significant improvement over local network search algorithms such as PSI-BLAST. Furthermore, we show here that labeled data can be used to learn a network without any need for estimating parameters of the transfer function, and that diffusion on this learned network produces better results than the original RankProp algorithm with a fixed network. CONCLUSION: In order to gain maximal information from a network, labeled and unlabeled data should be used to extract both local and global structure.

Algorithms↗

PubMedDate not supplied

Acetylcholine in cortical inference.

Acetylcholine (ACh) plays an important role in a wide variety of cognitive tasks, such as perception, selective attention, associative learning, and memory. Extensive experimental and theoretical work in tasks involving learning and memory has suggested that ACh reports on unfamiliarity and controls plasticity and effective network connectivity. Based on these computational and implementational insights, we develop a theory of cholinergic modulation in perceptual inference. We propose that ACh levels reflect the uncertainty associated with top-down information, and have the effect of modulating the interaction between top-down and bottom-up processing in determining the appropriate neural representations for inputs. We illustrate our proposal by means of an hierarchical hidden Markov model, showing that cholinergic modulation of contextual information leads to appropriate perceptual inference.

Acetylcholine↗

PubMed2000

Approximate maximum entropy joint feature inference consistent with arbitrary lower-order probability constraints: application to statistical classification

We propose a new learning method for discrete space statistical classifiers. Similar to Chow and Liu (1968) and Cheeseman (1983), we cast classification/inference within the more general framework of estimating the joint probability mass function (p.m.f.) for the (feature vector, class label) pair. Cheeseman's proposal to build the maximum entropy (ME) joint p.m.f. consistent with general lower-order probability constraints is in principle powerful, allowing general dependencies between features. However, enormous learning complexity has severely limited the use of this approach. Alternative models such as Bayesian networks (BNs) require explicit determination of conditional independencies. These may be difficult to assess given limited data. Here we propose an approximate ME method, which, like previous methods, incorporates general constraints while retaining quite tractable learning. The new method restricts joint p.m.f. support during learning to a small subset of the full feature space. Classification gains are realized over dependence trees, tree-augmented naive Bayes networks, BNs trained by the Kutato algorithm, and multilayer perceptrons. Extensions to more general inference problems are indicated. We also propose a novel exact inference method when there are several missing features.

Journal Article↗

PubMed2005

Expression of cell cycle markers and human papillomavirus infection in oral squamous cell carcinoma: use of fuzzy neural networks.

Our aim was to evaluate in oral squamous cell carcinoma (OSCC) the relationship between some cell cycle markers and HPV infection, conditionally to age, gender and certain habits of patients, and to assess the ability of fuzzy neural networks (FNNs) in building up an adequate predictive model based on logic inference rules. Eighteen cases of OSCC were examined by immunohistochemistry for MIB-1, PCNA and survivin expression; presence of HPV DNA was investigated in exfoliated oral mucosa cells by nested PCR (nPCR, MY09-MY11/GP5-GP6), and HPV genotype was determined by direct DNA sequencing. Data were analyzed by traditional statistics (TS) and FNNs. HPV DNA was found in 9/18 OSCCs (50.0 %) without any significant higher risk of HPV infection with respect to the sociodemographic variables considered (p > 0.2), apart from tobacco smoking, reported in 44.4% of OSCC HPV-positive vs. 100% HPV-negative subjects (p = 0.029). Regarding cell cycle markers, TS and FNN revealed that survivin was expressed significantly more in HPV-negative than in HPV-positive OSCC [root mean-square error (RMSE) = 5.89 x 10(-6), % predicted 100.0]; furthermore, smoking played a protective role for survivin expression in HPV-positive cases (OR = 0.019, 95%CI 0.001-0.723, RMSE = 0.20, % of prevision 94.4). FNN, although on a small sample size, allowed us to confirm data by TS and to hypothesize a different cell cycle pattern for HPV-positive vs. HPV-negative OSCC. In the latter cases, the relevance of apoptotic vs. proliferative markers suggested that they may be related to the different supposed outcome of HPV-negative OSCC and that HPV may have a protective role in the expression level of survivin, especially in tobacco smokers.

Adult↗

PubMedDate not supplied

Outlier detection in scatterometer data: neural network approaches.

Satellite-borne scatterometers are used to measure backscattered micro-wave radiation from the ocean surface. This data may be used to infer surface wind vectors where no direct measurements exist. Inherent in this data are outliers owing to aberrations on the water surface and measurement errors within the equipment. We present two techniques for identifying outliers using neural networks; the outliers may then be removed to improve models derived from the data. Firstly the generative topographic mapping (GTM) is used to create a probability density model; data with low probability under the model may be classed as outliers. In the second part of the paper, a sensor model with input-dependent noise is used and outliers are identified based on their probability under this model.GTM was successfully modified to incorporate prior knowledge of the shape of the observation manifold; however, GTM could not learn the double skinned nature of the observation manifold. To learn this double skinned manifold necessitated the use of a sensor model which imposes strong constraints on the mapping. The results using GTM with a fixed noise level suggested the noise level may vary as a function of wind speed. This was confirmed by experiments using a sensor model with input-dependent noise, where the variation in noise is most sensitive to the wind speed input. Both models successfully identified gross outliers with the largest differences between models occurring at low wind speeds.

Light↗

PubMed2006

Inferring functional pathways from multi-perturbation data.

BACKGROUND: Recently, a conceptually new approach for analyzing gene networks, the Functional Influence Network (FIN) was presented. The FIN approach uses the measured performance of a given cellular function under different multi-perturbations, to identify the main functional pathways and interactions underlying its processing. Here we present and study an iterative, extended version of FIN, the Functional Influence Network Extractor (FINE), which is specifically geared towards the accurate analysis of sparse cellular systems. We employ it to study a conceptually fundamental question of practical importance--how well should we know the system studied (such that we can predict its performance) so that we can understand its workings (i.e., chart its underlying functional network)? RESULTS AND CONCLUSIONS: The performance of FINE is studied in both simulated and biological sparse systems. It successfully obtains an accurate and compact description of the underlying functional network even with limited data, and outperforms FIN. We show that prior estimates of a system's functional complexity are instrumental in determining how much predictive knowledge is required to accurately chart its underlying functional network. AVAILABILITY: The FINE software is available for download at http://www.cns.tau.ac.il/resc.html.

Algorithms↗

PubMed1997

Efficient temporal probabilistic reasoning via context-sensitive model construction.

We present a language for representing context-sensitive temporal probabilistic knowledge. Context constraints allow inference to be focused on only the relevant portions of the probabilistic knowledge. We provide a declarative semantics for our language. We present a sound and complete algorithm for computing posterior probabilities of temporal queries, as well as an efficient implementation of the algorithm. Throughout we illustrate the approach with the problem of reasoning about the effects of medications and interventions on the state of a patient in cardiac arrest. We empirically evaluate the efficiency of our system by comparing its inference times on problems in this domain with those of standard Bayesian network representations of the problems.

Algorithms↗

PubMed1991

Distributed changes in rat brain DNA synthesis with long-term habituation and potentiation of the perforant path-granule cell synapse.

The involvement of brain deoxyribonucleic acid (DNA) synthesis in adaptive neural events was studied in the adult rat during long-term habituation (LTH) or potentiation (LTP) of the perforant path-granule cell synapse. Male Long-Evans rats were given 50 muCi [3H]thymidine intraventricularly under urethane anesthesia. Soon thereafter, field excitatory postsynaptic potential (EPSP) slope and population spike were monitored from the right dentate gyrus before and at various times (5, 10, 15, 60 min) following the delivery to the ipsilateral perforant bundle of a low frequency (LFS: 1.0 Hz, 160 s) or a high-frequency train (HFS: 400 Hz, 200 ms), repeated once after 5 min. Unstimulated implanted rats served as controls. DNA synthesis was evaluated by the incorporation of the radioactive precursor into DNA of several brain areas at the end of a 1 h incorporation period. In CA1, LTH and LTP increased DNA synthesis by 30% on the stimulated side. In the entorhinal cortex, LTH but not LTP increased DNA synthesis (by 30%) on the stimulated side. Conversely, in the frontal cortex, LTP but not LTH increased DNA synthesis (by 100%) on both sides. Long-lasting changes in synaptic efficacy covaried non-linearly with DNA synthesis in mono- and polysynaptically stimulated hippocampal regions, and in functionally associated neocortical areas. The co-variations of population spike amplitude were positive for LTH and negative for LTP in the dentate gyrus and frontal cortex of both sides, and in CA3/CA1 of the stimulated side, indicating higher DNA synthesis at lower values of LTH and LTP, and viceversa. Further, regional cross-correlation analyses revealed a high degree of synchronization among brain sites, following low- or high-frequency train pulses, indicating that (i) extra-target sites participate on the stimulated and on the contralateral side, and (ii) small distributed changes take place across the sampled neural networks. A modulatory role of information flow on brain DNA synthesis is inferred to take place in a diffuse, distributed manner.

Animals↗

PubMed1981

Externalization of an endogenous chicken muscle lectin with in vivo development.

Chicken-lactose-lectin-I (CLL-I), an endogenous lectin that is developmentally regulated in embryonic muscle, was localized by immunohistochemical techniques in tissue samples taken at various stages of in vivo development and in primary muscle cultures. Lectin, which was diffusely distributed in myoblasts, became localized in myotubes in a distribution similar to that of the sarcoplasmic reticulum and T tubules. Later in development, lectin was predominantly extracellular. This sequence suggests that externalization may have occurred by migration in the T tubules, which are continuous with the extracellular space, although alternative explanations are possible. Only traces of lectin were found in the adult. These studies did not reveal the function of CLL-I in muscle development. However, we infer that it acts by organizing complementary glycoconjugates in the intracellular tubular network, on the muscle surface, and/or in extracellular materials.

Animals↗

PubMed2025

CaLMPhosKAN: prediction of general phosphorylation sites in proteins via fusion of codon aware embeddings with amino acid aware embeddings and wavelet-based Kolmogorov-Arnold network.

MOTIVATION: The mapping from codon to amino acid is surjective due to codon degeneracy, suggesting that codon space might harbor higher information content. Embeddings from the codon language model have recently demonstrated success in various protein downstream tasks. However, predictive models for residue-level tasks such as phosphorylation sites, arguably the most studied Post-Translational Modification (PTM), and PTM sites prediction in general, have predominantly relied on representations in amino acid space. RESULTS: We introduce a novel approach for predicting phosphorylation sites by utilizing codon-level information through embeddings from the codon adaptation language model (CaLM), trained on protein-coding DNA sequences. Protein sequences are first reverse-translated into reliable coding sequences by mapping UniProt sequences to their corresponding NCBI reference sequences and extracting the exact coding sequences from their GenBank format using a dynamic programming-based global pairwise alignment. The resulting coding sequences are encoded using the CaLM encoder to generate codon-aware embeddings, which are subsequently integrated with amino acid-aware embeddings obtained from a protein language model, through an early fusion strategy. Next, a window-level representation of the site of interest, retaining the full sequence context, is constructed from the fused embeddings. A ConvBiGRU network extracts feature maps that capture spatiotemporal correlations between proximal residues within the window. This is followed by a prediction head based on a Kolmogorov-Arnold network (KAN) using the derivative of gaussian wavelet transform to generate the inference for the site. The overall model, dubbed CaLMPhosKAN, performs better than the existing approaches across multiple datasets. AVAILABILITY AND IMPLEMENTATION: CaLMPhosKAN is publicly available at https://github.com/KCLabMTU/CaLMPhosKAN.

Codon↗

PubMed1999

A large-scale distributed network for covert spatial attention: further anatomical delineation based on stringent behavioural and cognitive controls.

Functional MRI was used to examine cerebral activations in 12 subjects while they performed a spatial attention task. This study applied more stringent behavioural and cognitive controls than previously used for similar experiments: (i) subjects were included only if they showed evidence of attentional shifts while performing the task in the magnet; (ii) the experimental task and baseline condition were designed to eliminate the contributions of motor output, visual fixation, inhibition of eye movements, working memory and the conditional (no-go) component of responding. Activations were seen in all three hypothesized cortical epicentres forming a network for spatial attention: the lateral premotor cortex (frontal eye fields), the posterior parietal cortex and the cingulate cortex. Subcortical activations were seen in the basal ganglia and the thalamus. Although the task required attention to be equally shifted to the left and to the right, eight of 10 subjects showed a greater area of activation in the right parietal cortex, consistent with the specialization of the right hemisphere for spatial attention. Other areas of significant activation included the posterior temporo-occipital cortex and the anterior insula. The temporo-occipital activation was within a region broadly defined as MT+ (where MT is the middle temporal area) which contains the human equivalent of area MT in the macaque monkey. This temporo-occipital area appears to constitute a major component of the functional network activated by this spatial attention task. Its activation may reflect the 'inferred' shift of the attentional focus across the visual scene.

Adult↗