Variational Autoencoding Tissue Response to Microenvironment Perturbation.
Ontology highlight
ABSTRACT:
SUBMITTER: Schau GF
PROVIDER: S-EPMC6677277 | biostudies-literature | 2019 Feb
REPOSITORIES: biostudies-literature
Publications
Variational Autoencoding Tissue Response to Microenvironment Perturbation.
Schau Geoffrey F GF Thibault Guillaume G Dane Mark A MA Gray Joe W JW Heiser Laura M LM Chang Young Hwan YH
Proceedings of SPIE--the International Society for Optical Engineering 20190201
This work applies deep variational autoencoder learning architecture to study multi-cellular growth characteristics of human mammary epithelial cells in response to diverse microenvironment perturbations. Our approach introduces a novel method of visualizing learned feature spaces of trained variational autoencoding models that enables visualization of principal features in two dimensions. We find that unsupervised learned features more closely associate with expert annotation of cell colony org ...[more]
Similar Datasets
Project description:Based on Kleinert's variational perturbation (KP) theory [Path Integrals in Quantum Mechanics, Statistics, Polymer Physics, and Financial Markets, 3rd ed. (World Scientific, Singapore, 2004)], we present an analytic path-integral approach for computing the effective centroid potential. The approach enables the KP theory to be applied to any realistic systems beyond the first-order perturbation (i.e., the original Feynman-Kleinert [Phys. Rev. A 34, 5080 (1986)] variational method). Accurate values are obtained for several systems in which exact quantum results are known. Furthermore, the computed kinetic isotope effects for a series of proton transfer reactions, in which the potential energy surfaces are evaluated by density-functional theory, are in good accordance with experiments. We hope that our method could be used by non-path-integral experts or experimentalists as a "black box" for any given system.
| S-EPMC3780612 | biostudies-other
Project description:BackgroundMultiple studies have shown the utility of transcriptome-wide RNA-seq profiles as features for machine learning-based prediction of response to chemotherapy in cancer. While tumor transcriptome profiles are publicly available for thousands of tumors for many cancer types, a relatively modest number of tumor profiles are clinically annotated for response to chemotherapy. The paucity of labeled examples and the high dimension of the feature data limit performance for predicting therapeutic response using fully-supervised classification methods. Recently, multiple studies have established the utility of a deep neural network approach, the variational autoencoder (VAE), for generating meaningful latent features from original data. Here, we report the first study of a semi-supervised approach using VAE-encoded tumor transcriptome features and regularized gradient boosted decision trees (XGBoost) to predict chemotherapy drug response for five cancer types: colon, pancreatic, bladder, breast, and sarcoma.ResultsWe found: (1) VAE-encoding of the tumor transcriptome preserves the cancer type identity of the tumor, suggesting preservation of biologically relevant information; and (2) as a feature-set for supervised classification to predict response-to-chemotherapy, the unsupervised VAE encoding of the tumor's gene expression profile leads to better area under the receiver operating characteristic curve and area under the precision-recall curve classification performance than the original gene expression profile or the PCA principal components or the ICA components of the gene expression profile, in four out of five cancer types that we tested.ConclusionsGiven high-dimensional "omics" data, the VAE is a powerful tool for obtaining a nonlinear low-dimensional embedding; it yields features that retain biological patterns that distinguish between different types of cancer and that enable more accurate tumor transcriptome-based prediction of response to chemotherapy than would be possible using the original data or their principal components.
| S-EPMC8456615 | biostudies-literature
Project description:Cell-to-cell communication in skin participates to the maintenance of homeostatic responses to foreign substances. Certain strains of Staphylococcus (S) aureus are vicious pathogens that cause deleterious effects in host cells and tissues. Both secreted toxins and structural components of S. aureus trigger an immune response, though how S. aureus stimulates host immune responses is poorly understood. We explored here how keratinocytes and fibroblasts initiate the first steps of an immune response by activating dendritic cells (DCs) through recognition of structural components of S. aureus. We treated monocyte-derived Langerhans cells (moLCs) and monocyte-derived DCs (moDCs) with conditioned media from keratinocytes (K-CM) and fibroblasts (F-CM) treated with heat-killed S. aureus (HKSA) respectively, or directly with HKSA. Immune and inflammatory responses from keratinocytes, fibroblasts, moLCs and moDCs were assessed by analysis of cell surface markers and cytokine production using flow cytometry, real-time PCR and ELISA assays. K-CM and F-CM increased the expression of CD86 and HLA-DR on moLCs and moDCs, in association with a specific cytokine profile. K-CM upregulated TNFA, IL-1B and GM-CSF mRNA expression in moLCs, while F-CM upregulated IL-12 and downregulated TNFA and TGFB mRNA expression in moDCs. Additionally, F-CM attenuated the induction of an inflammatory profile in monocytes. The recognition of structural components from S. aureus by cutaneous microenvironment induces the activation and the expression of specific cytokines from LCs and DCs.
| S-EPMC6706075 | biostudies-literature
Project description:Recent advances in spatially resolved transcriptomics (SRT) have brought ever-increasing opportunities to characterize expression landscape in the context of tissue spatiality. Nevertheless, there still exist multiple challenges to accurately detect spatial functional regions in tissue. Here, we present a novel contrastive learning framework, SPAtially Contrastive variational AutoEncoder (SpaCAE), which contrasts transcriptomic signals of each spot and its spatial neighbors to achieve fine-grained tissue structures detection. By employing a graph embedding variational autoencoder and incorporating a deep contrastive strategy, SpaCAE achieves a balance between spatial local information and global information of expression, enabling effective learning of representations with spatial constraints. Particularly, SpaCAE provides a graph deconvolutional decoder to address the smoothing effect of local spatial structure on expression's self-supervised learning, an aspect often overlooked by current graph neural networks. We demonstrated that SpaCAE could achieve effective performance on SRT data generated from multiple technologies for spatial domains identification and data denoising, making it a remarkable tool to obtain novel insights from SRT studies.
| S-EPMC10849194 | biostudies-literature
Project description:In this paper, we describe an automated integration-free path-integral (AIF-PI) method, based on Kleinert's variational perturbation (KP) theory, to treat internuclear quantum-statistical effects in molecular systems. We have developed an analytical method to obtain the centroid potential as a function of the variational parameter in the KP theory, which avoids numerical difficulties in path-integral Monte Carlo or molecular dynamics simulations, especially at the limit of zero-temperature. Consequently, the variational calculations using the KP theory can be efficiently carried out beyond the first order, i.e., the Giachetti-Tognetti-Feynman-Kleinert variational approach, for realistic chemical applications. By making use of the approximation of independent instantaneous normal modes (INM), the AIF-PI method can readily be applied to many-body systems. Previously, we have shown that in the INM approximation, the AIF-PI method is accurate for computing the quantum partition function of a water molecule (3 degrees of freedom) and the quantum correction factor for the collinear H(3) reaction rate (2 degrees of freedom). In this work, the accuracy and properties of the KP theory are further investigated by using the first three order perturbations on an asymmetric double-well potential, the bond vibrations of H(2), HF, and HCl represented by the Morse potential, and a proton-transfer barrier modeled by the Eckart potential. The zero-point energy, quantum partition function, and tunneling factor for these systems have been determined and are found to be in excellent agreement with the exact quantum results. Using our new analytical results at the zero-temperature limit, we show that the minimum value of the computed centroid potential in the KP theory is in excellent agreement with the ground state energy (zero-point energy) and the position of the centroid potential minimum is the expectation value of particle position in wave mechanics. The fast convergent property of the KP theory is further examined in comparison with results from the traditional Rayleigh-Ritz variational approach and Rayleigh-Schrödinger perturbation theory in wave mechanics. The present method can be used for thermodynamic and quantum dynamic calculations, including to systematically determine the exact value of zero-point energy and to study kinetic isotope effects for chemical reactions in solution and in enzymes.
| S-EPMC2676110 | biostudies-literature
Project description:During the mitotic cell cycle, microtubule depolymerization leads to a cell cycle arrest in metaphase, due to activation of the spindle checkpoint. Here, we show that under microtubule-destabilizing conditions, such as low temperature or the presence of the spindle-depolymerizing drug benomyl, meiotic budding yeast cells arrest in G(1) or G(2), instead of metaphase. Cells arrest in G(1) if microtubule perturbation occurs as they enter the meiotic cell cycle and in G(2) if cells are already undergoing premeiotic S phase. Concomitantly, cells down-regulate genes required for cell cycle progression, meiotic differentiation, and spore formation in a highly coordinated manner. Decreased expression of these genes is likely to be responsible for halting both cell cycle progression and meiotic development. Our results point towards the existence of a novel surveillance mechanism of microtubule integrity that may be particularly important during specialized cell cycles when coordination of cell cycle progression with a developmental program is necessary.
| S-EPMC1140642 | biostudies-literature
Project description:In the current study, we introduce an integrative machine learning strategy for the autonomous molecular design of protein kinase inhibitors using variational autoencoders and a novel cluster-based perturbation approach for exploration of the chemical latent space. The proposed strategy combines autoencoder-based embedding of small molecules with a cluster-based perturbation approach for efficient navigation of the latent space and a feature-based kinase inhibition likelihood classifier that guides optimization of the molecular properties and targeted molecular design. In the proposed generative approach, molecules sharing similar structures tend to cluster in the latent space, and interpolating between two molecules in the latent space enables smooth changes in the molecular structures and properties. The results demonstrated that the proposed strategy can efficiently explore the latent space of small molecules and kinase inhibitors along interpretable directions to guide the generation of novel family-specific kinase molecules that display a significant scaffold diversity and optimal biochemical properties. Through assessment of the latent-based and chemical feature-based binary and multiclass classifiers, we developed a robust probabilistic evaluator of kinase inhibition likelihood that is specifically tailored to guide the molecular design of novel SRC kinase molecules. The generated molecules originating from LCK and ABL1 kinase inhibitors yielded ~40% of novel and valid SRC kinase compounds with high kinase inhibition likelihood probability values (p > 0.75) and high similarity (Tanimoto coefficient > 0.6) to the known SRC inhibitors. By combining the molecular perturbation design with the kinase inhibition likelihood analysis and similarity assessments, we showed that the proposed molecular design strategy can produce novel valid molecules and transform known inhibitors of different kinase families into potential chemical probes of the SRC kinase with excellent physicochemical profiles and high similarity to the known SRC kinase drugs. The results of our study suggest that task-specific manipulation of a biased latent space may be an important direction for more effective task-oriented and target-specific autonomous chemical design models.
| S-EPMC9569663 | biostudies-literature
Project description:Intense interval exercise has proven to be as effective as traditional endurance exercise in improving maximal oxygen uptake. Shared by these two exercise regimes is an acute reduction in plasma volume, which is a suggested stimulus behind exercise-induced increases in blood volume and maximal oxygen uptake. This study aimed to link exercise-induced metabolic perturbation with volume shifts into skeletal muscle tissue. Ten healthy subjects (mean age 33 ± 8 years, 5 males and 5 females) performed three 30 s all-out sprints on a cycle ergometer. Upon cessation of exercise magnetic resonance imaging, 31 Phosphorus magnetic resonance spectroscopy and blood samples were used to measure changes in muscle volume, intramuscular energy metabolites and plasma volume. Compared to pre-exercise, muscle volume increased from 1147.1 ± 35.6 ml to 1283.3 ± 11.0 ml 8 min post-exercise. At 30 min post-exercise, muscle volume was still higher than pre-exercise (1147.1 ± 35.6 vs. 1222.2 ± 6.8 ml). Plasma volume decreased by 16 ± 3% immediately post-exercise and recovered back to - 5 ± 6% after 30 min. Principal component analysis of exercise performance, muscle and plasma volume changes as well as changes in intramuscular energy metabolites showed generally strong correlations between metabolic and physiological variables. The strongest predictor for the volume shifts of muscle and plasma was the magnitude of glucose-6-phosphate accumulation post-exercise. Interval training leads to large metabolic and hemodynamic perturbations with accumulation of glucose-6-phosphate as a possible key event in the fluid flux between the vascular compartment and muscle tissue.
| S-EPMC8077120 | biostudies-literature
Project description:Conformational sampling of biomolecules using molecular dynamics simulations often produces a large amount of high dimensional data that makes it difficult to interpret using conventional analysis techniques. Dimensionality reduction methods are thus required to extract useful and relevant information. Here, we devise a machine learning method, Gaussian mixture variational autoencoder (GMVAE), that can simultaneously perform dimensionality reduction and clustering of biomolecular conformations in an unsupervised way. We show that GMVAE can learn a reduced representation of the free energy landscape of protein folding with highly separated clusters that correspond to the metastable states during folding. Since GMVAE uses a mixture of Gaussians as its prior, it can directly acknowledge the multi-basin nature of the protein folding free energy landscape. To make the model end-to-end differentiable, we use a Gumbel-softmax distribution. We test the model on three long-timescale protein folding trajectories and show that GMVAE embedding resembles the folding funnel with folded states down the funnel and unfolded states outside the funnel path. Additionally, we show that the latent space of GMVAE can be used for kinetic analysis and Markov state models built on this embedding produce folding and unfolding timescales that are in close agreement with other rigorous dynamical embeddings such as time independent component analysis.
| S-EPMC8605902 | biostudies-literature
Project description:Correct tissue architecture is essential for normal physiology, yet there have been few attempts to recreate tissues using micro-patterning. We have used polymer brush micro-engineering to generate a stratified micro-epidermis with fewer than 10 human keratinocytes. Epidermal stem cells are captured on 100 ?m diameter circular collagen-coated disks. Within 24 h they assemble a stratified micro-tissue, in which differentiated cells have a central suprabasal location. For rings with a non-adhesive centre of up to 40 ?m diameter, cell-cell and cell-matrix adhesive interactions together result in correct micro-epidermis assembly. Assembly requires actin polymerization, adherens junctions and desmosomes, but not myosin II-mediated contractility nor coordinated cell movement. Squamous cell carcinoma cells on micro-patterned rings exhibit disturbed architecture that correlates with the characteristics of the original tumours. The micro-epidermis we have generated provides a new platform for screening drugs that modulate tissue assembly, quantifying tissue stratification and investigating the properties of tumour cells.
| S-EPMC3437971 | biostudies-literature