41,866 research outputs found
AI3SD Video: The Application of Machine Learning in Molecular Spectroscopy Study
No full textOptical-spectroscopy provides powerful toolkits to decipher molecular structures and their configuration evolutions. However, the theoretical analysis of spectroscopic signals and connecting them with structural detail is a challenging task. Moreover, the intrinsic complexity of spectroscopic signals of molecular systems makes it difficult to correlate spectral characteristics with the underlying molecular structure and dynamics. Herein, we have developed data-driven machine learning (ML) protocols that can predict infrared (IR), ultraviolet/visible (UV/Vis) and Raman spectra of molecule systems with 3 to 5 orders of magnitude reduced computation cost compared to direct quantum chemistry calculations. A convolutional neural network (CNN) model was trained and tested on a dataset consisting 87993 spectra computed from protein peptide segments with α-helical, β-sheet, and other typical secondary structures. The secondary structure classification accuracy reached near 100% and over 98.7% on spectra sets of new segments extracted from the same and homologous proteins, respectively. Importantly, we demonstrate the ML protocol to realize cost-effective relations between spectra, structure, and chemical properties, i.e. spectra determination/prediction from structural information, and configuration or chemical properties determination/recognition from spectroscopic signals.1. S. Ye, K. Zhong, J.X. Zhang, W. Hu, J. Hirst, G.Z. Zhang, S. Mukamel, J. Jiang*, A Machine Learning Protocol for Predicting Protein Infrared Spectra, J. Am. Chem. Soc. 142 (2020) 19071-19077.2. X.J. Wang, S. Ye, W. Hu, E. Sharman, R. Liu, Y. Liu, Y. Luo, J. Jiang*, Electric Dipole Descriptor for Machine Learning Prediction of Catalyst Surface-Molecular Adsorbate Interactions, J. Am. Chem. Soc. 142 (2020) 7737-7743.3. S. Ye, W. Hu, X. Li, J.X. Zhang, K. Zhong, G.Z. Zhang, Y. Luo, S. Mukamel*, J. Jiang*, A Neural Network Protocol for Electronic excitations of N-Methylacetamide, Proc Natl Acad Sci USA. 116 (2019) 11612-11617.4. W. Hu, S. Ye, Y.J Zhang, T.D. Li, G.Z. Zhang, Y. Luo, S. Mukamel, J. Jiang*, Machine Learning Protocol for Surface-Enhanced Raman Spectroscopy, J. Phys. Chem. Lett. 10 (2019) 6026-6031
Resume: Yang Jun 杨君
No full textThis is Yang Jun\u27s resume from the 1995 New Asian Art Show. His brief educational background and exhibition experience up until 1995 are on the left side of this paper, while one of his collage of photos is on the right. (Zhuocheng Jiang \u2726)https://digital.kenyon.edu/zhoudocs/1361/thumbnail.jp
Editing flagellin derivatives for exploration of potent radioprotective agents
No full textExploration of medical radiation countermeasures (MRCs) has great implications in protection of mammals from radiation damages. While flagellin has been recently reported to show radioprotective effects, the relationships between flagellin structure and radioprotective activity are rarely explored. Herein, we deliberately edited the amino acid sequence of flagellin in its binding domain with toll-like receptor 5 (TLR5) for exploration of potent flagellin derivatives (Fds). An in vitro screening paradigm was developed to examine the radioprotective effects of six engineered Fds. Notably, mutation of 103 threonine on flagellin into asparagine resulted in a potent MRC candidate (Fd-T103N) displaying 1.28-fold increment of interactions with TLR5. Fd-T103N was able to further activate NF-κB pathway, induce immune protective cytokine (e.g. G-CSF) release, and significantly ameliorate γ-irradiation induced cell death. The protection effects of Fd-T103N were further validated in mice exposed to 10 Gray γ-irradiations. Compared to parent flagellin, Fd-T103N treatment showed higher G-CSF release in mouse blood, lower intestine damages, and 13% increments of mouse survival rates. In short, the established predictive paradigm could greatly reduce the labor-, time- and animal-costs in exploration of MRC candidates. Fd-T103N is a promising candidate of investigational new drug for radioprotection
Signaling through CD44 affects cell cycle progression and c-Jun expression in acute myeloid leukemia cells
No full textWe present here the first evidence linking CD44 signaling to c-Jun expression and cell cycle progression in myeloid cell line models. CD44 ligation with the anti-CD44 monoclonal antibodies have been shown to induce differentiation and inhibit the proliferation of human acute myeloid leukemia (AML) cells, and c-Jun is involved in the regulation of these processes. The effects of anti-CD44 monoclonal antibody A3D8, on myeloid cells were associated with specific disruption of cell cycle events and induction of G0/G1 arrest. Induction of G0/G1 arrest was accompanied by an increase in the expression of p21, attenuation of pRb phosphorylation and associated with decreased CDK2 and CDK4 kinase activities. We observed that A3D8 treatment of AML patient blasts and HL60/U937 cells led to the downregulation of c-Jun expression at mRNA and protein level. Transient transfection studies showed the inhibition of c-jun promoter activity by A3D8, involving both AP-1 sites. Furthermore, A3D8 treatment caused a decrease in JNK protein expression and a decrease in the level of phosphorylated c-Jun. Ectopic overexpression of c-Jun in HL60 cells was able to induce proliferation and prevent the anti-proliferative effects of A3D8. Targeting of G1 regulatory proteins and the resulting induction of G1 arrest by A3D8 may provide new insights into anti-proliferative and differentiation therapy of AML
Nano-enabled Photosynthesis in Tumor to Activate Lipid Peroxidation for Overcome of Cancer Resistances
No full textApoptosis dysregulation is an important mechanism responsible for the intrinsic and acquired resistances of melanoma, which necessitates the exploration of oncological treatments to activate non-apoptotic cell deaths. Herein, we engineered a lipid peroxidation reaction in hyperoxia tumor microenvironment for induction of ferroptosis and overcome of melanoma resistances. The hyperoxia microenvironment was deliberately constructed by implants of photosynthetic microcapsules (PMCs) consisting of cyanobacteria and upconversion nanoparticles for photosynthesis in chloroplast driven by external near infrared photons. Combination of PMCs and X rays evoked a lipid peroxidation reaction and ferroptosis in melanoma cells and xenografts. Consequently, the intrinsic and acquired resistances in melanoma could be overcome by the engineered reaction, which further contributed to the amelioration of tumor metastases and improvement of survival rates in melanoma-bearing mice. Overall, our findings provide opportunities to overcome melanoma resistances by engineered biochemical reactions and will inspire the exploration of oncological treatments
Synthetic Vectors for Activating the Driving Axis of Ferroptosis
No full textFerroptosis is a promising strategy for cancer therapy, with numerous inhibitors of its braking axes under investigation as potential drugs. However, few studies have explored the potential of activating the driving axes to induce ferroptosis. Herein, phosphatidylcholine peroxide decorating liposomes (LIPPCPO) were synthesized to induce ferroptosis by targeting divalent metal transporter 1 (DMT1). LIPPCPO was found to boost lysosomal Fe2+ efflux by inducing cysteinylation of lysosomal DMT1, resulting in glutathione peroxidase 4 (GPX4) suppression, glutathione depletion and ferroptosis in breast cancer cells and xenografts. Importantly, LIPPCPO induced ferroptotic cell death was independent of acquired resistance to radiation, chemotherapy, or targeted agents in 11 cancer cell lines. Furthermore, a strong synergistic ferroptosis effect was observed between LIPPCPO and an FDA-approved drug, artesunate. The formula of LIPPCPO encapsulating artesunate significantly inhibited tumor growth and metastasis and improved the survival rate of breast cancer-bearing mice. These findings provide a distinct strategy for inducing ferroptosis and highlight the potential of LIPPCPO as a vector to synergize the therapeutic effects of conventional ferroptosis inducers
Statistical/climatic models to predict and project extreme precipitation events dominated by large-scale atmospheric circulation over the central-eastern China
No full textGlobal warming has posed non-negligible effects on regional extreme precipitation changes and increased the uncertainties when meteorologists predict such extremes. More importantly, floods, landslides, and waterlogging caused by extreme precipitation have had catastrophic societal impacts and led to steep economic damages across the world, in particular over central-eastern China (CEC), where heavy precipitation due to the Meiyu-front and typhoon activities often causes flood disaster. There is mounting evidence that the anomaly atmospheric circulation systems and water vapor transport have a dominant role in triggering and maintaining the processes of regional extreme precipitation. Both understanding and accurately predicting extreme precipitation events based on these anomalous signals are hot issues in the field of hydrological research.
In this thesis, the self-organizing map (SOM) and event synchronization were used to cluster the large-scale atmospheric circulation reflected by geopotential height at 500 hPa and to quantify the level of synchronization between the identified circulation patterns with extreme precipitation over CEC. With the understanding of which patterns were associated with extreme precipitation events, and corresponding water vapor transport fields, a hybrid deep learning model of multilayer perceptron and convolutional neural networks (MLP-CNN) was proposed to achieve the binary predictions of extreme precipitation. The inputs to MLP-CNN were the anomalous fields of GP at 500 hPa and vertically integrated water vapor transport (IVT). Compared with the original MLP, CNN, and two other machine learning models (random forest and support vector machine), MLP-CNN showed the best performance. Additionally, since the coarse spatial resolution of global circulation models and its large biases in extremes precipitation estimations, a new precipitation downscaling framework that combination of ensemble-learning and nonhomogeneous hidden Markov model (Ensemble-NHMM) was developed, to improve the reliabilities of GCMs in historical simulations and future projection. The performances of downscaled precipitation from reanalysis and GCM datasets were validated against the gauge observations and also compared with the results of traditional NHMM. Finally, the Ensemble-NHMM downscaling model was applied to future scenario data of GCM. On the projections of change trends in precipitation over CEC in the early-, medium- and late- 21st centuries under different emission scenarios, the possible causes were discussed in term of both thermodynamic and dynamic factors. Main results are enumerated as follows.
(1) The large-scale atmospheric circulation patterns and associated water vapor transport fields synchronized with extreme precipitation events over CEC were quantitatively identified, as well as the contribution of circulation pattern changes to extreme precipitation changes and their teleconnection with the interdecadal modes of the ocean. Firstly, based on the nonparametric Pettitt test, it was found that 23% of rain gauges had significant abrupt changes in the annual extreme precipitation from 1960 to 2015. The average change point in the annual extreme precipitation frequency and amount occurred near 1989. Complex network analysis showed that the rain gauges highly synchronized on extreme precipitation events can be clustered into four clusters based on modularity information. Secondly, the dominant circulation patterns over CEC were robustly identified based on the SOM. From the period 1960–1989 to 1990–2015, the categories of identified circulation patterns generally remain almost unchanged. Among these, the circulation patterns characterized by obvious positive anomalies of 500 hPa geopotential height over the Eastern Eurasia continent and negative values over the surrounding oceans are highly synchronized with extreme precipitation events. An obvious water vapor channel originating from the northern Indian Ocean driven by the southwesterly airflow was observed for the representative circulation patterns (synchronized with extreme precipitation). Finally, the circulation pattern changes produced an increase in extreme precipitation frequency from 1960–1989 to 1990–2015. Empirical mode decomposition of the annual frequency variation signals in the representative circulation pattern showed that the 2–4 yr oscillation in the annual frequency was closely related to the phase of El Niño and Southern Oscillation (ENSO); while the 20–25 yr and 42–50 yr periodic oscillations were responses to the Pacific Decadal Oscillation and the Atlantic Multidecadal Oscillation.
(2) A regional extreme precipitation prediction model was constructed. Two deep learning models-MLP and CNN were linearly stacked and used two atmospheric variables associated with extreme precipitation, that is, geopotential height at 500 hPa and IVT. The hybrid model can learn both the local-scale information with MLP and large-scale circulation information with CNN. Validation results showed that the MLP-CNN model can predict extreme or non-extreme precipitation days with an overall accuracy of 86%. The MLP-CNN also showed excellent seasonal transferability with an 81% accuracy on the testing set from different seasons of the training set. MLP-CNN significantly outperformed over other machine learning models, including MLP, CNN, random forest, and support vector machine. Additionally, the MLP-CNN can be used to produce precursor signals by 1 to 2 days, though the accuracy drops quickly as the number of precursor days increases.
(3) The GCM seriously underestimated extreme precipitation over CEC but showed convincing results for reproducing large-scale atmospheric circulation patterns. The accuracies of 10 GCMs in extreme precipitation and large-scale atmospheric circulation simulations were evaluated. First, five indices were selected to measure the characteristics of extreme precipitation and the performances of GCMs were compared to the gauge-based daily precipitation analysis dataset over the Chinese mainland. The results showed that except for FGOALS-g3, most GCMs can reproduce the spatial distribution characteristics of the average precipitation from 1960 to 2015. However, all GCMs failed to accurately estimate the extreme precipitation with large underestimation (relative bias exceeds 85%). In addition, using the circulation patterns identified by the fifth-generation reanalysis data (ERA5) as benchmarks, GCMs can reproduce most CP types for the periods 1960–1989 and 1990–2015. In terms of the spatial similarity of the identified CPs, MPI-ESM1-2-HR was superior.
(4) To improve the reliabilities of precipitation simulations and future projections from GCMs, a new statistical downscaling framework was proposed. This framework comprises two models, ensemble learning and NHMM. First, the extreme gradient boosting (XGBoost) and random forest (RF) were selected as the basic- and meta- classifiers for constructing the ensemble learning model. Based on the top 50 principal components of GP at 500 hPa and IVT, this model was trained to predict the occurrence probabilities for the different levels of daily precipitation (no rain, very light, light, moderate, and heavy precipitation) aggregated by multi-sites. Confusion matrix results showed that the ensemble learning model had sufficient accuracy (>88%) in classifying no rain or rain days and (>83%) predicting moderate precipitation events. Subsequently, precipitation downscaling was done using the probability sequences of daily precipitation as large-scale predictors to NHMM. Statistical metrics showed that the Ensemble-NHMM downscaled results matched best to the gauge observations in precipitation variabilities and extreme precipitation simulations, compared with the result from the one that directly used circulation variables as predictors. Finally, the downscaling model also performed well in the historical simulations of MPI-ESM1-2-HR, which reproduced the change trends of annual precipitation and the means of total extreme precipitation index.
(5) Three climate scenarios with different Shared Socioeconomic Pathways and Representative Concentration Pathways (SSPs) were selected to project the future precipitation change trends. The Ensemble-NHMM downscaling model was applied to the scenario data from MPI-ESM1-2-HR. Projection results showed that the CEC would receive more precipitation in the future by ~30% through the 2075–2100 period. Compared to the recent 26-year epoch (1990–2015), the frequency and magnitude of extreme precipitation would increase by 21.9–48.1% and 12.3–38.3% respectively under the worst emission scenario (SSP585). In particular, the south CEC region is projected to receive more extreme precipitation than the north. Investigations of thermodynamic and dynamic factors showed that climate warming would increase the probability of stronger water vapor convergence over CEC. More wet weather states due to the enhanced water vapor transport, as well as the increased favoring large-scale atmospheric circulation and the strengthen pressure gradient would be the factors for the increased precipitation
c-Jun reprograms Schwann cells of injured nerves to generate a repair cell essential for regeneration.
No full textThe radical response of peripheral nerves to injury (Wallerian degeneration) is the cornerstone of nerve repair. We show that activation of the transcription factor c-Jun in Schwann cells is a global regulator of Wallerian degeneration. c-Jun governs major aspects of the injury response, determines the expression of trophic factors, adhesion molecules, the formation of regeneration tracks and myelin clearance and controls the distinctive regenerative potential of peripheral nerves. A key function of c-Jun is the activation of a repair program in Schwann cells and the creation of a cell specialized to support regeneration. We show that absence of c-Jun results in the formation of a dysfunctional repair cell, striking failure of functional recovery, and neuronal death. We conclude that a single glial transcription factor is essential for restoration of damaged nerves, acting to control the transdifferentiation of myelin and Remak Schwann cells to dedicated repair cells in damaged tissue
- …
