CanCellVar: A database for single-cell variants map in human cancer.

Numerous variants, including both single-nucleotide variants (SNVs) in DNA and A>G RNA edits in mRNA as essential drivers of cellular proliferation and tumorigenesis, are commonly associated with cancer progression and growth. Thus, mining and summarizing single-cell variants will provide a refined and higher-resolution view of cancer and further contribute to precision medicine. Here, we established a database, CanCellVar, which aims to provide and visualize the comprehensive atlas of single-cell variants in tumor microenvironment. The current CanCellVar identified ∼3 million variants (∼1.4 million SNVs and ∼1.4 million A>G RNA edits) involved in 2,754,531 cells of 5 major cell types across 37 cancer types. CanCellVar provides the basic annotation information as well as cellular and molecular function properties of variants. In addition, the clinical relevance of variants can be obtained including tumor grade, treatment, metastasis, and others. Several flexible tools were also developed to aid retrieval and to analyze cell-cell interactions, gene expression, cell-development trajectories, regulation, and molecular structure affected by variants. Collectively, CanCellVar will serve as a valuable resource for investigating the functions and characteristics of single-cell variations and their roles in human tumor evolution and treatment.

Hyperbolic graph embedding of MEG brain networks to study brain alterations in individuals with subjective cognitive decline.

An expansive area of research focuses on discerning patterns of alterations in functional brain networks from the early stages of Alzheimer's disease, even at the subjective cognitive decline (SCD) stage. Here, we developed a novel hyperbolic MEG brain network embedding framework for transforming high-dimensional complex MEG brain networks into lower-dimensional hyperbolic representations. Using this model, we computed hyperbolic embeddings of the MEG brain networks of two distinct participant groups: individuals with SCD and healthy controls. We demonstrated that these embeddings preserve both local and global geometric information, presenting reduced distortion compared to rival models, even when brain networks are mapped into low-dimensional spaces. In addition, our findings showed that the hyperbolic embeddings encompass unique SCD-related information that improves the discriminatory power above and beyond that of connectivity features alone. Notably, we introduced a unique metric-the radius of the node embeddings-which effectively proxies the hierarchical organization of the brain. Using this metric, we identified subtle hierarchy organizational differences between the two participant groups, suggesting increased hierarchy in the dorsal attention, frontoparietal, and ventral attention subnetworks among the SCD group. Last, we assessed the correlation between these hierarchical variations and cognitive assessment scores, revealing associations with diminished performance across multiple cognitive evaluations in the SCD group. Overall, this study presents the first evaluation of hyperbolic embeddings of MEG brain networks, offering novel insights into brain organization, cognitive decline, and potential diagnostic avenues of Alzheimer's disease.

A framework of computer vision-enhanced microfluidic approach for automated assessment of the transient sickling kinetics in sickle red blood cells.

The occurrence of vaso-occlusive crisis greatly depends on the competition between the sickling delay time and the transit time of individual sickle cells, i.e., red blood cells (RBCs) from sickle cell disease (SCD) patients, while they are traversing the circulatory system. Many drugs for treating SCD work by inhibiting the polymerization of sickle hemoglobin (HbS), effectively delaying the sickling process in sickle cells (SS RBCs). Most previous studies on screening anti-sickling drugs, such as voxelotor, rely on in vitro testing of sickling characteristics, often conducted under prolonged deoxygenation for up to 1 hour. However, since the microcirculation of RBCs typically takes less than 1 minute, the results of these studies may be less accurate and less relevant for in vitro-in vivo correlation. In our current study, we introduce a computer vision-enhanced microfluidic framework designed to automatically capture the transient sickling kinetics of SS RBCs within a 1-min timeframe. Our study has successfully detected differences in the transient sickling kinetics between vehicle control and voxelotor-treated SS RBCs. This approach has the potential for broader applications in screening anti-sickling therapies.

Interaction of widowhood, gender, and age in predicting loneliness among older adults in China.

BACKGROUND: Loneliness is a significant issue for the elderly, and widowhood is considered a major risk factor. However, research on the intersectional effects of gender, age, and widowhood on loneliness is limited, especially within the Chinese cultural context. METHODS: Using six waves (2002-2018) of national longitudinal data from the Chinese Longitudinal Healthy Longevity Survey (N = 22,777), this study employed multilevel mixed-effects ordered logistic regression to analyze the impact of widowhood on loneliness. Moderating roles of gender and age were examined through interaction effects. RESULTS: Widowhood significantly increased loneliness across genders and age groups, but this effect diminished with age. Widowed men experienced greater loneliness than women, but this difference converged by age 90. The buffering effect of age on the widowhood-loneliness link was less pronounced among older women. CONCLUSION: The study unravels the complexity of how gender, age, and widowhood interact to shape loneliness in later life. Targeted interventions considering these intersections are needed to alleviate loneliness among Chinese widowed elderly.

TransformerG2G: Adaptive time-stepping for learning temporal graph embeddings using transformers.

Dynamic graph embedding has emerged as a very effective technique for addressing diverse temporal graph analytic tasks (i.e., link prediction, node classification, recommender systems, anomaly detection, and graph generation) in various applications. Such temporal graphs exhibit heterogeneous transient dynamics, varying time intervals, and highly evolving node features throughout their evolution. Hence, incorporating long-range dependencies from the historical graph context plays a crucial role in accurately learning their temporal dynamics. In this paper, we develop a graph embedding model with uncertainty quantification, TransformerG2G, by exploiting the advanced transformer encoder to first learn intermediate node representations from its current state (t) and previous context (over timestamps [t-1,t-l], l is the length of context). Moreover, we employ two projection layers to generate lower-dimensional multivariate Gaussian distributions as each node's latent embedding at timestamp t. We consider diverse benchmarks with varying levels of "novelty" as measured by the TEA (Temporal Edge Appearance) plots. Our experiments demonstrate that the proposed TransformerG2G model outperforms conventional multi-step methods and our prior work (DynG2G) in terms of both link prediction accuracy and computational efficiency, especially for high degree of novelty. Furthermore, the learned time-dependent attention weights across multiple graph snapshots reveal the development of an automatic adaptive time stepping enabled by the transformer. Importantly, by examining the attention weights, we can uncover temporal dependencies, identify influential elements, and gain insights into the complex interactions within the graph structure. For example, we identified a strong correlation between attention weights and node degree at the various stages of the graph topology evolution.

Hyperbolic graph embedding of MEG brain networks to study brain alterations in individuals with subjective cognitive decline.

An expansive area of research focuses on discerning patterns of alterations in functional brain networks from the early stages of Alzheimer's disease, even at the subjective cognitive decline (SCD) stage. Here, we developed a novel hyperbolic MEG brain network embedding framework for transforming high-dimensional complex MEG brain networks into lower-dimensional hyperbolic representations. Using this model, we computed hyperbolic embeddings of the MEG brain networks of two distinct participant groups: individuals with SCD and healthy controls. We demonstrated that these embeddings preserve both local and global geometric information, presenting reduced distortion compared to rival models, even when brain networks are mapped into low-dimensional spaces. In addition, our findings showed that the hyperbolic embeddings encompass unique SCD-related information that improves the discriminatory power above and beyond that of connectivity features alone. Notably, we introduced a unique metric-the radius of the node embeddings-which effectively proxies the hierarchical organization of the brain. Using this metric, we identified subtle hierarchy organizational differences between the two participant groups, suggesting increased hierarchy in the dorsal attention, frontoparietal, and ventral attention subnetworks among the SCD group. Last, we assessed the correlation between these hierarchical variations and cognitive assessment scores, revealing associations with diminished performance across multiple cognitive evaluations in the SCD group. Overall, this study presents the first evaluation of hyperbolic embeddings of MEG brain networks, offering novel insights into brain organization, cognitive decline, and potential diagnostic avenues of Alzheimer's disease.

A global database for modeling tumor-immune cell communication.

Communications between tumor cells and surrounding immune cells help shape the tumor immunity continuum. Recent breakthroughs in high-throughput technologies as well as computational algorithms had reported many important tumor-immune cell (TIC) communications, which were scattered in thousands of published studies and impeded systematical characterization of the TIC communications across cancer. Here, a comprehensive database, TICCom, was developed to model TIC communications, containing 739 experimentally-validated or manually-curated interactions collected from more than 3,000 literatures as well as 4,537,709 predicted interactions inferred via six computational algorithms by reanalyzing 32 scRNA-seq datasets and bulk RNA-seq data across 25 cancer types. The communications between tumor cells and 14 types of immune cells were characterized, and the involved ligand-receptor interactions were further integrated. 14190 human and 3650 mouse integrated ligand-receptor interactions with supplemented corresponding function information were also stored in the TICCom database. Our database would serve as a valuable resource for investigating TIC communications.

Heteroatom Introduction to Reconstruct Interfacial Chemical Structures for High-Reliability CFRTP/A6061-T6 Hybrid Structures.

The application of carbon-fiber-reinforced thermoplastic (CFRTP)/metal hybrid structures is a vital step for realizing the lightweight design concepts in aerospace. However, the CFRTP/metal hybrid structures are usually not reliable enough in practical applications due to the high differences in chemical and physical properties between these two materials. The current work provides a bottom-up strategy of introducing heteroatoms into CFRTP/metal interfaces to reconstruct the interfacial chemical structures and thus manufacture high-reliability hybrid structures. Based on the principle of utmost using reaction sites at metal surfaces, the heteroatoms of oxygen and hydrogen are specially designed and introduced to the CFRTP/A6061-T6 (6061) interfaces by simple and green plasma polymerization. The introduced oxygen and hydrogen heteroatoms react with the aluminum and oxygen of the oxidation film at 6061 surfaces to produce great interfacial Al-O covalencies and hydrogen bonds. The reconstructing interfacial chemical structures strengthen the joint strength of CFRTP/6061 hybrid structures from 8.82 to 23.97 MPa. Our heteroatom introduction strategy is expected to get a fresh insight into the interfacial design concept and has several important implications for the future application of high-reliability CFRTP/metal hybrid structures.

Microstructure-driven electrical conductivity optimization in additively manufactured microscale copper interconnects.

As the microelectronics field pushes to increase device density through downscaling component dimensions, various novel micro- and nano-scale additive manufacturing technologies have emerged to expand the small scale design space. These techniques offer unprecedented freedom in designing 3D circuitry but have not yet delivered device-grade materials. To highlight the complex role of processing on the quality and microstructure of AM metals, we report the electrical properties of micrometer-scale copper interconnects fabricated by Fluid Force Microscopy (FluidFM) and Electrohydrodynamic-Redox Printing (EHD-RP). Using a thin film-based 4-terminal testing chip developed for the scope of this study, the electrical resistance of as-printed metals is directly related to print strategies and the specific morphological and microstructural features. Notably, the chip requires direct synthesis of conductive structures on an insulating substrate, which is shown for the first time in the case of FluidFM. Finally, we demonstrate the unique ability of EHD-RP to tune the materials resistivity by one order of magnitude solely through printing voltage. Through its novel electrical characterization approach, this study offers unique insight into the electrical properties of micro- and submicrometer-sized copper interconnects and steps towards a deeper understanding of micro AM metal properties for advanced electronics applications.

Injectable chitosan/xyloglucan composite hydrogel with mechanical adaptivity and endogenous bioactivity for skin repair.

Delayed or chronic wound healing is one of severe clinical issues. Developing scaffold materials capable of supporting cells and inducing tissue regeneration remains a challenge. Here, a polysaccharide-based hydrogel is constructed for promoting full-thickness skin wound healing in mouse model. The engineering hydrogel consists of a dynamic crosslinking network formed by the Schiff base reaction between aldehyde-containing xyloglucan and methacrylated chitosan. Its reversible gel-sol-gel transition upon shearing force is highly beneficial to completely cover and fill irregular wound shape. The second covalent cross-linking network achieved by photo-initiated polymerization offers a feasible way to tune the mechanical property of hydrogel after injection, with an ideal mechanical adaptivity for clinical application. Remarkably, both in vitro and in vivo evaluations demonstrate that the hydrogel with endogenously bioactive galactoside units can promote cell spheroid formation and accelerate wound healing by expediting re-epithelialization, collagen deposition, angiogenesis as well as the formation of hair follicles.

Dynamics in Deep Classifiers Trained with the Square Loss: Normalization, Low Rank, Neural Collapse, and Generalization Bounds.

We overview several properties-old and new-of training overparameterized deep networks under the square loss. We first consider a model of the dynamics of gradient flow under the square loss in deep homogeneous rectified linear unit networks. We study the convergence to a solution with the absolute minimum ρ, which is the product of the Frobenius norms of each layer weight matrix, when normalization by Lagrange multipliers is used together with weight decay under different forms of gradient descent. A main property of the minimizers that bound their expected error for a specific network architecture is ρ. In particular, we derive novel norm-based bounds for convolutional layers that are orders of magnitude better than classical bounds for dense networks. Next, we prove that quasi-interpolating solutions obtained by stochastic gradient descent in the presence of weight decay have a bias toward low-rank weight matrices, which should improve generalization. The same analysis predicts the existence of an inherent stochastic gradient descent noise for deep networks. In both cases, we verify our predictions experimentally. We then predict neural collapse and its properties without any specific assumption-unlike other published proofs. Our analysis supports the idea that the advantage of deep networks relative to other classifiers is greater for problems that are appropriate for sparse deep architectures such as convolutional neural networks. The reason is that compositionally sparse target functions can be approximated well by "sparse" deep networks without incurring in the curse of dimensionality.

In vitro simulated digestion and fecal fermentation of exopolysaccharides from Lacticaseibacillus paracasei GL1.

The prebiotic properties of two purified fractions (GL1-E1 and GL1-E2) of exopolysaccharides (EPSs) from Lacticaseibacillus paracasei GL1 were investigated through in vitro fermentation of pure and human fecal cultures. The results indicated that the simulated digestion under saliva, gastric, and small intestinal conditions had no effect on GL1-E1 and GL1-E2. Additionally, GL1-E1 and GL1-E2 can be used as substrates for Lactobacillus and Lactococcus growth. It was also found that both were gradually degraded and utilized by the gut microbiota. As fermentation proceeded, the pH continued to decrease. Additionally, the total short-chain fatty acid (SCFA) production significantly increased, especially the major SCFA of formic, lactic, and acetic acid. Furthermore, GL1-E1 and GL1-E2 could significantly regulate the composition of the gut microbiota, by increasing the relative abundances of Bacteroides and Phascolarctobacterium, and decreasing the pathogenic bacteria Escherichia-Shigella, Klebsiella, and Fusobacterium. These results suggest that GL1-E1 and GL1-E2 have the potential to be developed as a prebiotic.

Determination of Isoflavones in Radix puerariae from Different Origins by Ultra-High Performance Liquid Chromatography Based on Optimal Pretreatment Method.

A new method for simultaneous determination of puerarin, daidzin, daidzein and genistein in Radix puerariae by ultra-high performance liquid chromatography was established. The target analytes were extracted from Radix puerariae by 70% ethylene glycol with the assistance of ultrasonication, purified by the absorption of N-propyl ethylenediamine (PSA), and separated on a Supersil ODS column (4.6 mm × 250 mm × 2.5 µm). Gradient elution in 12 min was performed with the mobile phase 0.1% formic acid(A)-acetonitrile(B). The column temperature was 25 °C and the flow rate was 1 mL/min. The detection wavelength of the four target analytes was 250 nm. The limits of detection (LODs) of puerarin, daidzin, daidzein and genistein were 0.086 mg/L, 0.020 mg/L, 0.027 mg/L and 0.037 mg/L, respectively, and limits of quantitation (LOQs) were 0.29 mg/L, 0.065 mg/L, 0.090 mg/L and 0.12 mg/L, respectively. The recovery of the four substances ranged from 90.5% to 109.6%, and the relative standard deviation (n = 6) was less than 7.7%. With the established methods, puerarin, daidzin, daidzein and genistein in Radix puerariae from 11 origins were determined. The contents of the four compounds varied with the origin and variety. It provides basic data and technical means for quality control and regulation of Radix puerariae.

A two-stage framework for optical coherence tomography angiography image quality improvement.

Introduction: Optical Coherence Tomography Angiography (OCTA) is a new non-invasive imaging modality that gains increasing popularity for the observation of the microvasculatures in the retina and the conjunctiva, assisting clinical diagnosis and treatment planning. However, poor imaging quality, such as stripe artifacts and low contrast, is common in the acquired OCTA and in particular Anterior Segment OCTA (AS-OCTA) due to eye microtremor and poor illumination conditions. These issues lead to incomplete vasculature maps that in turn makes it hard to make accurate interpretation and subsequent diagnosis. Methods: In this work, we propose a two-stage framework that comprises a de-striping stage and a re-enhancing stage, with aims to remove stripe noise and to enhance blood vessel structure from the background. We introduce a new de-striping objective function in a Stripe Removal Net (SR-Net) to suppress the stripe noise in the original image. The vasculatures in acquired AS-OCTA images usually exhibit poor contrast, so we use a Perceptual Structure Generative Adversarial Network (PS-GAN) to enhance the de-striped AS-OCTA image in the re-enhancing stage, which combined cyclic perceptual loss with structure loss to achieve further image quality improvement. Results and discussion: To evaluate the effectiveness of the proposed method, we apply the proposed framework to two synthetic OCTA datasets and a real AS-OCTA dataset. Our results show that the proposed framework yields a promising enhancement performance, which enables both conventional and deep learning-based vessel segmentation methods to produce improved results after enhancement of both retina and AS-OCTA modalities.

A Conceptual Strategy toward High-Reliability Metal-Thermoplastic Hybrid Structures Based on a Covalent-Bonding Mechanism.

Metal-thermoplastic hybrid structures have proven their effectiveness to achieve lightweight design concepts in both primary and secondary structural components of advanced aircraft. However, the drastic differences in physical and chemical properties between metal and thermoplastic make it challenging to fabricate high-reliability hybrid structures. Here, a simple and universal strategy to obtain strong hybrid structures thermoplastics is reported by regulating the bonding behavior at metal/thermoplastic interfaces. To achieve such, we first researched and uncovered the bonding mechanism at metal/thermoplastic interfaces by experimental methods and density functional theory (DFT) calculations. The results suggest that the interfacial covalency, which is formed due to the interfacial reaction between high-electronegativity elements of thermoplastics and metallic elements at metal surfaces, dominates the interfacial bonding interaction of metal-thermoplastic hybrid structures. The differences in electronegativity and atomic size between bonding atoms influence the covalent-bond strength and finally control the interfacial reliability of hybrid structures. Based on our covalent-bonding mechanism, the carboxyl functional group (COOH) is specifically grafted on polyetheretherketone (PEEK) by plasma polymerization to increase the density and strength of interfacial covalency and thus fabricate high-reliability hybrid structures between PEEK and A6061-T6 aluminum alloy. Current work provides an in-depth understanding of the bonding mechanism at metal-thermoplastics interfaces, which opens a fascinating direction toward high-reliability metal-thermoplastic hybrid structures.

Structural and prebiotic activity analysis of the polysaccharide produced by Lactobacillus helveticus SNA12.

Exopolysaccharide (EPS) isolated from Lactobacillus helveticus SNA12 were purified, and one fraction (SNA12-EPS) was obtained. The structure of SNA12-EPS was proposed using Fourier-transform infrared spectroscopy, gas chromatography-mass spectrometry, and nuclear magnetic resonance. Results showed that SNA12-EPS was rich in galactose and glucose with the molar ratios of 2.1:1.0, and SNA12-EPS possessed the repeat units of →3)-ß-D-Glcp-(1→3)-ß-D-Glcp-(1→4)-ß-D-Galp-(1→ with an average molecular weight of 3.81 × 105 Da. The scanning electron microscope results showed that SNA12-EPS had a tight structure with a smooth and uneven surface. Furthermore, the prebiotic potential of SNA12-EPS was performed using in vitro simulated digestion with human faecal fermentation. SNA12-EPS was not digested by digestive juice, and it could markedly regulate the gut microbiota composition by increasing the relative abundances of Parabacteroides and Blautia and decreasing the abundance of pathogenic bacteria of Fusobacterium. Additionally, SNA12-EPS improved the ability of gut microbiota to produce short-chain fatty acids.

AOSLO-net: A Deep Learning-Based Method for Automatic Segmentation of Retinal Microaneurysms From Adaptive Optics Scanning Laser Ophthalmoscopy Images.

Purpose: Accurate segmentation of microaneurysms (MAs) from adaptive optics scanning laser ophthalmoscopy (AOSLO) images is crucial for identifying MA morphologies and assessing the hemodynamics inside the MAs. Herein, we introduce AOSLO-net to perform automatic MA segmentation from AOSLO images of diabetic retinas. Method: AOSLO-net is composed of a deep neural network based on UNet with a pretrained EfficientNet as the encoder. We have designed customized preprocessing and postprocessing policies for AOSLO images, including generation of multichannel images, de-noising, contrast enhancement, ensemble and union of model predictions, to optimize the MA segmentation. AOSLO-net is trained and tested using 87 MAs imaged from 28 eyes of 20 subjects with varying severity of diabetic retinopathy (DR), which is the largest available AOSLO dataset for MA detection. To avoid the overfitting in the model training process, we augment the training data by flipping, rotating, scaling the original image to increase the diversity of data available for model training. Results: The validity of the model is demonstrated by the good agreement between the predictions of AOSLO-net and the MA masks generated by ophthalmologists and skillful trainees on 87 patient-specific MA images. Our results show that AOSLO-net outperforms the state-of-the-art segmentation model (nnUNet) both in accuracy (e.g., intersection over union and Dice scores), as well as computational cost. Conclusions: We demonstrate that AOSLO-net provides high-quality of MA segmentation from AOSLO images that enables correct MA morphological classification. Translational Relevance: As the first attempt to automatically segment retinal MAs from AOSLO images, AOSLO-net could facilitate the pathological study of DR and help ophthalmologists make disease prognoses.

Prebiotic Properties of Exopolysaccharides from Lactobacillus helveticus LZ-R-5 and L. pentosus LZ-R-17 Evaluated by In Vitro Simulated Digestion and Fermentation.

The in vitro digestion and fermentation behaviors of Lactobacillus helveticus LZ-R-5- and L. pentosus LZ-R-17-sourced exopolysaccharides (LHEPS and LPEPS) were investigated by stimulated batch-culture fermentation system. The results illustrated that LHEPS was resistant to simulated saliva and gastrointestinal (GSI) digestion, whereas LPEPS generated a few monosaccharides after digestion without significant influence on its main structure. Additionally, LHEPS and LPEPS could be consumed by the human gut microbiota and presented stronger bifidogenic effect comparing to α-glucan and ß-glucan, as they promote the proliferation of Lactobacillus and Bifidobacterium in cultures and exhibited high values of selectivity index (13.88 and 11.78, respectively). Furthermore, LPEPS achieved higher contents of lactic acid and acetic acid (35.74 mM and 45.91 mM, respectively) than LHEPS (35.20 mM and 44.65 mM, respectively) during fermentation for 48 h, thus also resulting in a larger amount of total SCFAs (110.86 mM). These results have clearly indicated the potential prebiotic property of EPS fractions from L. helveticus LZ-R-5 and L. pentosus LZ-R-17, which could be further developed as new functional food prebiotics to beneficially improve human gut health.

Preparation of ZIF-8 and Its Application in Determination of Pyridoxine and Pyridoxal in Ginkgo Seeds by Ultra-Performance Liquid Chromatography.

A new rapid and accurate method was developed for simultaneous determination of pyridoxine and pyridoxal in ginkgo seeds, using ultra-performance liquid chromatography (UPLC) equipped with a fluorescence detector. Diluted hydrochloric acid solution was used as the extracting solvent. For the pretreatment of extracts, a zeolitic imidazolate framework material (ZIF-8) was prepared and characterized. An ODS-BP column (4.6 mm × 250 mm × 5 µm) was used for separation. The conditions of sample extraction, cleaning and separation were optimized. The linear correlation coefficient (R2) of the analyte was better than 0.9999, indicating good linearity. The limits of detection (LODs) of pyridoxal and pyridoxine were 0.0065 mg/kg and 0.0057 mg/kg, respectively, and limits of quantitation (LOQs) were 0.022 mg/kg and 0.019 mg/kg, respectively. The recovery of the two substances ranged from 86.2% to 110.4%, and the relative standard deviation (n = 6) was less than 7.5%. The method was applied to determine the contents of pyridoxine and pyridoxal in actual ginkgo seed samples with satisfactory results.

DynG2G: An Efficient Stochastic Graph Embedding Method for Temporal Graphs.

Dynamic graph embedding has gained great attention recently due to its capability of learning low-dimensional and meaningful graph representations for complex temporal graphs with high accuracy. However, recent advances mostly focus on learning node embeddings as deterministic "vectors" for static graphs, hence disregarding the key graph temporal dynamics and the evolving uncertainties associated with node embedding in the latent space. In this work, we propose an efficient stochastic dynamic graph embedding method (DynG2G) that applies an inductive feedforward encoder trained with node triplet energy-based ranking loss. Every node per timestamp is encoded as a time-dependent probabilistic multivariate Gaussian distribution in the latent space, and, hence, we are able to quantify the node embedding uncertainty on-the-fly. We have considered eight different benchmarks that represent diversity in size (from 96 nodes to 87 626 and from 13 398 edges to 4 870 863) as well as diversity in dynamics, from slowly changing temporal evolution to rapidly varying multirate dynamics. We demonstrate through extensive experiments based on these eight dynamic graph benchmarks that DynG2G achieves new state-of-the-art performance in capturing the underlying temporal node embeddings. We also demonstrate that DynG2G can simultaneously predict the evolving node embedding uncertainty, which plays a crucial role in quantifying the intrinsic dimensionality of the dynamical system over time. In particular, we obtain a "universal" relation of the optimal embedding dimension, Lo , versus the effective dimensionality of uncertainty, Du , and infer that Lo=Du for all cases. This, in turn, implies that the uncertainty quantification approach we employ in the DynG2G algorithm correctly captures the intrinsic dimensionality of the dynamics of such evolving graphs despite the diverse nature and composition of the graphs at each timestamp. In addition, this L0 - Du correlation provides a clear path to selecting adaptively the optimum embedding size at each timestamp by setting L ≥ Du .