Modeling Experiments Research Articles

FLIGHTED: Inferring Fitness Landscapes from Noisy High-Throughput Experimental Data.

Machine learning (ML) for protein design requires large protein fitness datasets generated by high-throughput experiments for training, fine-tuning, and benchmarking models. However, most models do not account for experimental noise inherent in these datasets, harming model performance and changing model rankings in benchmarking studies. Here we develop FLIGHTED, a Bayesian method of accounting for uncertainty by generating probabilistic fitness landscapes from noisy high-throughput experiments. We demonstrate how FLIGHTED can improve model performance on two categories of experiments: single-step selection assays, such as phage display and SELEX, and a novel high-throughput assay called DHARMA that ties activity to base editing. We then compare the performance of standard machine-learning models on fitness landscapes generated with and without FLIGHTED. Accounting for noise significantly improves model performance, especially of CNN architectures, and changes relative rankings on numerous common benchmarks. Based on our new benchmarking with FLIGHTED, data size, not model scale, currently appears to be limiting the performance of protein fitness models, and the choice of top model architecture matters more than the protein language model embedding. Collectively, our results indicate that FLIGHTED can be applied to any high-throughput assay and any machine learning model, making it straightforward for protein designers to account for experimental noise when modeling protein fitness.

Energy Landscape Reveals the Underlying Mechanism of Cancer-Adipose Conversion in Gene Network Models.

Cancer is a systemic heterogeneous disease involving complex molecular networks. Tumor formation involves an epithelial-mesenchymal transition (EMT), which promotes both metastasis and plasticity of cancer cells. Recent experiments have proposed that cancer cells can be transformed into adipocytes via a combination of drugs. However, the underlying mechanisms for how these drugs work, from a molecular network perspective, remain elusive. To reveal the mechanism of cancer-adipose conversion (CAC), this study adopts a systems biology approach by combing mathematical modeling and molecular experiments, based on underlying molecular regulatory networks. Four types of attractors are identified, corresponding to epithelial (E), mesenchymal (M), adipose (A) and partial/intermediate EMT (P) cell states on the CAC landscape. Landscape and transition path results illustrate that intermediate states play critical roles in the cancer to adipose transition. Through a landscape control approach, two new therapeutic strategies for drug combinations are identified, that promote CAC. These predictions are verified by molecular experiments in different cell lines. The combined computational and experimental approach provides a powerful tool to explore molecular mechanisms for cell fate transitions in cancer networks. The results reveal underlying mechanisms of intermediate cell states that govern the CAC, and identified new potential drug combinations to induce cancer adipogenesis.

Biological and dust aerosols as sources of ice-nucleating particles in the eastern Mediterranean: source apportionment, atmospheric processing and parameterization

Abstract. Aerosol–cloud interactions in mixed-phase clouds (MPCs) are one of the most uncertain drivers of the hydrological cycle and climate change. A synergy of in situ, remote-sensing and modelling experiments were used to determine the source of ice-nucleating particles (INPs) for MPCs at Mount Helmos in the eastern Mediterranean. The influences of boundary layer turbulence, vertical aerosol distributions and meteorological conditions were also examined. When the observation site is in the free troposphere (FT), approximately 1 in ×106 aerosol particles serve as INPs around −25 °C. The INP abundance spans 3 orders of magnitude and increases in the following order: marine aerosols; continental aerosols; and, finally, dust plumes. Biological particles are important INPs observed in continental and marine aerosols, whereas they play a secondary, although important, role during Saharan dust events. Air masses in the planetary boundary layer (PBL) show both enriched INP concentrations and a higher proportion of INPs to total aerosol particles, compared with cases in the FT. The presence of precipitation/clouds enriches INPs in the FT but decreases INPs in the PBL. Additionally, new INP parameterizations are developed that incorporate the ratio of fluorescent-to-nonfluorescent or coarse-to-fine particles and predict >90 % of the observed INPs within an uncertainty range of a factor of 10; these new parameterizations exhibit better performance than current widely used parameterizations and allow ice formation in models to respond to variations in dust and biological particles. The improved parameterizations can help MPC formation simulations in regions with various INP sources or different regions with prevailing INP sources.

MSC Exosomes Containing Valproic Acid Promote Wound Healing by Modulating Inflammation and Angiogenesis.

Chronic wounds that are difficult to heal pose a major challenge for clinicians and researchers. Currently, common treatment methods focus on isolating the wound from the outside world, relying on the tissue at the wound site to grow and heal unaided. Umbilical cord mesenchymal stem cell (MSC) exosomes can promote wound healing by enhancing new blood vessel growth at the wound site. Valproic acid (VPA) reduces the inflammatory response and acts on macrophages to accelerate wound closure. In this study, VPA was loaded into umbilical cord MSC exosomes to form a drug carrier exosome (VPA-EXO) with the aim of investigating the effect of VPA-EXO on wound healing. This study first isolated and obtained umbilical cord MSC exosomes, then added VPA to the exosomes and explored the ability of VPA-EXO to promote the proliferation and migration of human skin fibroblasts (HSFs) and human umbilical vein endothelial cells (HUVECs), as well as the ability to promote the angiogenesis of HUVECs, by using scratch, Transwell, and angiogenesis assays. An in vitro cell model was established and treated with VPA-EXO, and the expression levels of inflammation and pro-angiogenesis-related proteins and genes were examined using Western blot and qRT-PCR. The therapeutic effect of VPA-EXO on promoting wound healing in a whole skin wound model was investigated using image analysis of the wound site, H&E staining, and immunohistochemical staining experiments in a mouse wound model. The in vitro model showed that VPA-EXO effectively promoted the proliferation and migration of human skin fibroblast cells and human umbilical vein endothelial cells; significantly inhibited the expression of MMP-9, IL-1β, IL-8, TNF-α, and PG-E2; and promoted the expression of vascular endothelial growth factors. In the mouse wound model, VPA-EXO reduced inflammation at the wound site, accelerated wound healing, and significantly increased the collagen content of tissue at the wound site. As a complex with dual efficacy in simultaneously promoting tissue regeneration and inhibiting inflammation, VPA-EXO has potential applications in tissue wound healing and vascular regeneration. In future studies, we will further investigate the mechanism of action and application scenarios of drug-loaded exosome complexes in different types of wound healing and vascular regeneration.

Structural and Biochemical Characterization of Aminoglycoside Nucleotidyltransferase(6)-Ib From Campylobacter fetus subsp. fetus.

Aminoglycoside antibiotics have played a critical role in the treatment of both Gram-negative and Gram-positive bacterial infections. However, antibiotic resistance has severely compromised the efficacy of aminoglycosides. A leading cause of aminoglycoside resistance is mediated by bacterial enzymes that inactivate these drugs via chemical modification. Aminoglycoside nucleotidyltransferase-6 (ANT(6)) enzymes inactivate streptomycin by transferring an adenyl group from ATP to position 6 on the antibiotic. Despite the clinical significance of this activity, ANT(6) enzymes remain relatively uncharacterized. Here, we report the first high resolution x-ray crystallographic structure of ANT(6)-Ib from Campylobacter fetus subsp. fetus bound with streptomycin. Structural modeling and gel filtration chromatography experiments suggest that the enzyme exists as a dimer in which both subunits contribute to the active site. Moreover, superposition of the ANT(6)-Ib structure with the structurally related enzyme lincosamide nucleotidyltransferase B (LinB) permitted the identification of a putative nucleotide binding site. These data also suggest that residues D44 and D46 coordinate essential divalent metal ions and D102 functions as the catalytic base.

Mechanistic study of leukopenia treatment by Qijiao shengbai Capsule via the Bcl2/Bax/CASAPSE3 pathway.

Leukopenia can be caused by chemotherapy, which suppresses bone marrow function and can impact the effectiveness of cancer treatment. Qijiao Shengbai Capsule (QJSB) is commonly used to treat leukopenia, but the specific bioactive components and mechanisms of action are not well understood. This study aimed to analyze the active ingredients of QJSB and its potential targets for treating leukopenia using network pharmacology and molecular docking. Through a combination of serum pharmacochemistry, multi-omics, network pharmacology, and validation experiments in a murine leukopenia model, the researchers sought to understand how QJSB improves leukopenia. The study identified 16 key components of QJSB that act in vivo to increase the number of white blood cells in leukopenic mice. Multi-omics analysis and network pharmacology revealed that the PI3K-Akt and MAPK signaling pathways are important in the treatment of leukopenia with QJSB. Five specific targets (JUN, FOS, BCl-2, CASPAS-3) were identified as key targets. Validation experiments confirmed that QJSB regulates genes related to cell growth and inhibits apoptosis, suggesting that apoptosis may play a crucial role in leukopenia development and that QJSB may improve immune function by regulating apoptotic proteins and increasing CD4+ T cell count in leukopenic mice.

Климатические изменения температуры воздуха западной части российской Арк­тики в 1940—2099 гг. по данным ERA5 и моделям CMIP6

The study deals with the climatic changes in surface air temperature (SAT) in the western part of the Russian Arctic (60—80° N, 30—90° E). To analyze the changes in SAT that occurred over the period 1940—2023, we use data from the ERA5 reanalysis and the results of the Historical model experiment CMIP6. Future changes in SAT until the end of the 21st century we consider based on the results of SSP experiments of CMIP6 models. An increase in the average SAT of the studied region by 2—4°C is shown from approximately the mid-1970s to 2023. Moreover, this increase in SAT is most noticeable in the White and Kara Seas, as well as in the north and east of the Barents Sea. CMIP6 models based on different greenhouse gas emission scenarios produce markedly different forecasts for the increase in SAT for the region under study until the end of the 21st century. Thus, depending on the scenario, the average increase in SAT in the region under study by the end of the 21st century can range from 2—4°C to 6—10°, with stronger growth of SAT in the north of the region under study. At the same time, with little dependence on the future scenario, CMIP6 models predict that in the next 30 years the average SAT of the western part of the Russian Arctic will increase by approximately 2—3°C, and in the north of the region under study its increase may be more than 3°C, and in south — about 2°C. Thus, the difference in average SAT between the north and south of the region under study will decrease throughout the 21st century under any of the SSP scenarios considered.

Construction and Performance Study of a Dual-Network Hydrogel Dressing Mimicking Skin Pore Drainage for Photothermal Exudate Removal and On-Demand Dissolution.

In recent years, negative pressure wound dressings have garnered widespread attentions. However, it is challenging to drain the accumulated fluid under negative pressures for hydrogel dressings. To address this issue, this study prepared a chemical/physical duel-network PEG-CMCS/AG/MXene hydrogel composed by chemical disulfide crosslinked network of four-arm polyethylene glycol/carboxymethyl chitosan (4-Arm-PEG-SH/CMCS), and the physical network of hydrogen bond of agar (AG). Under near-infrared light (NIR) irradiation, the PEG-CMCS/AG/MXene hydrogel undergoes photothermal heating due to integrate of MXene, which destructs the hydrogen bond network and allows the removal of exudate through a mechanism mimicking the sweat gland-like effect of skin pores. The photothermal heating effect also enables the antimicrobial activity to prevent wound infections. The excellent electrical conductivity of PEG-CMCS/AG/MXene can promote cell proliferation under the external electrical stimulation (ES) in vitro. The animal experiments of full-thickness skin defect model further demonstrate its ability to accelerate wound healing. The conversion between thioester and thiol achieved with L-cysteine methyl ester hydrochloride (L-CME) can provides the on-demand dissolution of the dressing in situ. This study holds promises to provide a novel solution to the issue of fluid accumulations under hydrogel dressings and offers new approaches to alleviating or avoiding the significant secondary injuries caused by frequent dressing changes.

Operation parameters study on the performance of PEMFC based orthogonal test method

By selecting five key operational parameters—inlet temperature, operating temperature, anode humidity, cathode humidity, and operating pressure—and using PEMFC power density, membrane temperature difference, and membrane water content as performance indicators, simulations were conducted using 3D CFD modeling and orthogonal experiments to identify the most significant factors influencing PEMFC performance and determine the optimal operational parameter combination. The results show that working pressure has the most significant impact on PEMFC performance. Pressure variations cause significant changes in diffusion and heat conduction within the PEMFC. When the working pressure increases from 1.0 MPa to 3.0 MPa, the power density increases by approximately 11.2 %, but the temperature difference within the membrane increases by about 30.5 %. The difference between operating temperature and inlet temperature directly leads to changes in the temperature difference within the membrane. Moreover, changes in operating temperature greatly affect the water content within the proton exchange membrane, which in turn impacts the PEMFC's power density. When the operating temperature increases from 60 °C to 100 °C, the water content within the membrane decreases by approximately 50.5 %, and the power density decreases by 60.4 %. By optimizing the combination of operating parameters using a comprehensive scoring method, the temperature difference within the membrane can be controlled at 27.78 °C while ensuring the power density is increased to 0.92 W/cm2.

Land surface processes response to warming and wetting trend in Northwest China

Abstract The climate in Northwest China (NWC) has undergone a warming and wetting trend (WWT) since the 1980s, which has attracted considerable attention from the scientific and policy communities. However, the majority of previous studies have focused on overall effects of WWT, and very few have examined how land surface system responds to climate warming or wetting trend, respectively. For this purpose, this study uses the Community Land Model (CLM5) driven by the Chinese Meteorological Forcing Dataset (CMFD) to conduct four modelling experiments: a control experiment (CTRL) and three sensitivity experiments, in which the annual trend of air temperature (NonWarm), precipitation (NonWet), and both (NonWWT) are removed from the CMFD from 1979 to 2018. Compared to CTRL, the land hydrological variables (i.e., soil moisture, runoff and evapotranspiration) show a visible reduction in magnitude, interannual variability, as well as annual trend in NonWet, while they are enhanced in NonWarm. In both NonWarm and NonWet, the magnitude and trend of both net radiation and sensible heat fluxes increase, with a more pronounced change in NonWWT. Further analysis indicates that the land surface processes are more sensitive to wetting trend than to warming trend. Among all land surface hydrological variables and energy variables, runoff and snow cover fraction are the most susceptible to climate change. Overall, the effects of climate change in Ta and Pr on surface hydrological variables are non-linearly offsetting, while the effects on surface energy budgets are non-linearly superimposed. Compared to warming trend, wetting trend plays a larger impact in the variability of land surface processes in NWC.

Construction of a sequence activated fluorescence probe for simultaneous detection of γ-glutamyl transpeptidase and peroxynitrite in acute kidney injury

Acute kidney injury (AKI) can result in a sudden decline in kidney function and, if not promptly diagnosed and treated, can lead to a high mortality rate. Therefore, there is a critical need for the development of a non-invasive and dependable early diagnostic method for AKI to prevent its progression and deterioration. To address the risk of misdiagnosis or overlooked diagnosis due to reliance on a single biomarker, we developed a novel molecular fluorescent probe (HX-GP) to simultaneously detect and image two biomarkers, γ-Glutamyl transpeptidase (γ-GGT) and Peroxynitrite (ONOO−), in the AKI process. HX-GP can specifically detect γ-GGT in the red fluorescence channel (λem = 613 nm) and ONOO− in the green fluorescence channel (λem = 518 nm). HX-GP demonstrated high sensitivity, selectivity, and rapid response, showing excellent biocompatibility and detection performance. In addition, HX-GP was successful in imaging experiments in a cell model of cisplatin-induced AKI, a result that highlights its potential application value in early diagnosis of AKI.

Bayesian Conavigation: Dynamic Designing of the Material Digital Twins via Active Learning.

Scientific advancement is universally based on the dynamic interplay between theoretical insights, modeling, and experimental discoveries. However, this feedback loop is often slow, including delayed community interactions and the gradual integration of experimental data into theoretical frameworks. This challenge is particularly exacerbated in domains dealing with high-dimensional object spaces, such as molecules and complex microstructures. Hence, the integration of theory within automated and autonomous experimental setups, or theory in the loop-automated experiment, is emerging as a crucial objective for accelerating scientific research. The critical aspect is to use not only theory but also on-the-fly theory updates during the experiment. Here, we introduce a method for integrating theory into the loop through Bayesian conavigation of theoretical model space and experimentation. Our approach leverages the concurrent development of surrogate models for both simulation and experimental domains at the rates determined by latencies and costs of experiments and computation, alongside the adjustment of control parameters within theoretical models to minimize epistemic uncertainty over the experimental object spaces. This methodology facilitates the creation of digital twins of material structures, encompassing both the surrogate model of behavior that includes the correlative part and the theoretical model itself. While being demonstrated here within the context of functional responses in ferroelectric materials, our approach holds promise for broader applications, such as the exploration of optical properties in nanoclusters, microstructure-dependent properties in complex materials, and properties of molecular systems.

Experimentation of AI Models Towards the Prediction of Medium-Risk Emergency Department Cases Disposition Outcome.

The overwhelming volume of patients in emergency departments (EDs) is a significant problem that hinders the delivery of high quality healthcare. Despite their great value, triage protocols are challenging to put into practice. This paper examines the utility of prediction models as a tool for clinical decision support, with a focus on medium-severity patients as defined by the ESI algorithm. 689 cases of medium-risk patients were gathered from the AHEPA hospital, evaluated, and their data fed three classifiers: XGBoost (XGB), Random Forest (RF) and Logistic Regression (LR), with the prediction goal being the outcome of their visit, i.e. admission or discharge. Essential features for the prediction task were determined using feature importance and distribution analysis. Despite having many missing values or high sparsity datasets, several symptoms and metrics are recommended as crucial for outcome prediction. When fed the patients' vital signs, XGB achieved an accuracy score of 91.30%. Several chief complaints were also proven beneficial. Prediction models can, in general, not only lessen the drawbacks of triage implementation, but also enhance its delivery.

Computational Modeling of Drug Response Identifies Mutant-Specific Constraints for Dosing panRAF and MEK Inhibitors in Melanoma.

This study explores the potential of pre-clinical in vitro cell line response data and computational modeling in identifying the optimal dosage requirements of pan-RAF (Belvarafenib) and MEK (Cobimetinib) inhibitors in melanoma treatment. Our research is motivated by the critical role of drug combinations in enhancing anti-cancer responses and the need to close the knowledge gap around selecting effective dosing strategies to maximize their potential. In a drug combination screen of 43 melanoma cell lines, we identified specific dosage landscapes of panRAF and MEK inhibitors for NRAS vs. BRAF mutant melanomas. Both experienced benefits, but with a notably more synergistic and narrow dosage range for NRAS mutant melanoma (mean Bliss score of 0.27 in NRAS vs. 0.1 in BRAF mutants). Computational modeling and follow-up molecular experiments attributed the difference to a mechanism of adaptive resistance by negative feedback. We validated the in vivo translatability of in vitro dose-response maps by predicting tumor growth in xenografts with high accuracy in capturing cytostatic and cytotoxic responses. We analyzed the pharmacokinetic and tumor growth data from Phase 1 clinical trials of Belvarafenib with Cobimetinib to show that the synergy requirement imposes stricter precision dose constraints in NRAS mutant melanoma patients. Leveraging pre-clinical data and computational modeling, our approach proposes dosage strategies that can optimize synergy in drug combinations, while also bringing forth the real-world challenges of staying within a precise dose range. Overall, this work presents a framework to aid dose selection in drug combinations.

Bearing Dynamics Modeling Based on the Virtual State-Space and Hammerstein-Wiener Model.

This study investigates a novel approach for assessing the health status of rotating machinery transmission systems by analyzing the dynamic degradation of bearings. The proposed method generates multi-dimensional data by creating virtual states and constructs a multi-dimensional model using virtual state-space in conjunction with mechanism model analysis. Innovatively, the Hammerstein-Wiener (HW) modeling technique from control theory is applied to identify these dynamic multi-dimensional models. The modeling experiments are performed, focusing on the model's input and output types, the selection of nonlinear module estimators, the configuration of linear module transfer functions, and condition transfer. Dynamic degradation response signals are generated, and the method is validated using four widely recognized databases consisting of accurate measurement signals collected by vibration sensors. Experimental results demonstrated that the model achieved a modeling accuracy of 99% for multiple bearings under various conditions. The effectiveness of this dynamic modeling method is further confirmed through comparative experimental data and signal images. This approach offers a novel reference for evaluating the health status of transmission systems.

My Fuzzers Won’t Build: An Empirical Study of Fuzzing Build Failures

Fuzzing is an automated software testing technique used to find software vulnerabilities that works by sending large amounts of inputs to a software system to trigger bad behaviors. In recent years, the open-source software ecosystem has seen a significant increase in the adoption of fuzzing to avoid spreading vulnerabilities throughout the ecosystem. While fuzzing can uncover vulnerabilities, there is currently a lack of knowledge regarding the challenges of conducting fuzzing activities over time. Specifically, fuzzers are very complex tools to set up and build before they can be used. We set out to empirically find out how challenging is build maintenance in the context of fuzzing. We mine over 1.2 million build logs from Google's OSS-Fuzz service to investigate fuzzing build failures. We first conduct a quantitative analysis to quantify the prevalence of fuzzing build failures. We then manually investigate 677 failing fuzzing builds logs and establish a taxonomy of 25 root causes of build failures. We finally train a machine learning model to recognize common failure patterns in failing build logs. Our taxonomy can serve as a reference for practitioners conducting fuzzing build maintenance. Our modeling experiment shows the potential of using automation to simplify the process of fuzzing.

SIRPα engagement regulates ILC2 effector function and alleviates airway hyperreactivity via modulating energy metabolism

Group-2 innate lymphoid cells (ILC2) are part of a growing family of innate lymphocytes known for their crucial role in both the development and exacerbation of allergic asthma. The activation and function of ILC2s are regulated by various activating and inhibitory molecules, with their balance determining the severity of allergic responses. In this study, we aim to elucidate the critical role of the suppressor molecule signal regulatory protein alpha (SIRPα), which interacts with CD47, in controlling ILC2-mediated airway hyperreactivity (AHR). Our data indicate that activated ILC2s upregulate the expression of SIRPα, and the interaction between SIRPα and CD47 effectively suppresses both ILC2 proliferation and effector function. To evaluate the function of SIRPα in ILC2-mediated AHR, we combined multiple approaches including genetically modified mouse models and adoptive transfer experiments in murine models of allergen-induced AHR. Our findings suggest that the absence of SIRPα leads to the overactivation of ILC2s. Conversely, engagement of SIRPα with CD47 reduces ILC2 cytokine production and effectively regulates ILC2-dependent AHR. Furthermore, the SIRPα-CD47 axis modulates mitochondrial metabolism through the JAK/STAT and ERK/MAPK signaling pathways, thereby regulating NF-κB activity and the production of type 2 cytokines. Additionally, our studies have revealed that SIRPα is inducible and expressed on human ILC2s, and administration of human CD47-Fc effectively suppresses the effector function and cytokine production. Moreover, administering human CD47-Fc to humanized ILC2 mice effectively alleviates AHR and lung inflammation. These findings highlight the promising therapeutic potential of targeting the SIRPα-CD47 axis in the treatment of ILC2-dependent allergic asthma.

Dual-Functional Implant Based on Gellan-Xanthan Hydrogel with Diopside, BMP-2 and Lysostaphin for Bone Defect Repair and Control of Staphylococcal Infection.

A new dual-functional implant based on gellan-xanthan hydrogel with calcium-magnesium silicate ceramic diopside and recombinant lysostaphin and bone morphogenetic protein 2(BMP-2)-ray is developed. In this composite, BMP-2 is immobilized on microparticles of diopside while lysostaphin is mixed directly into the hydrogel, providing sustained release of BMP-2 to allow gradual bone formation and rapid release of lysostaphin to eliminate infection immediately after implantation. Introduction of diopside of up to 3% (w/v) has a negligible effect on the mechanical properties of the hydrogel but provides a high sorption capacity for BMP-2. The hydrogels show good biocompatibility and antibacterial activity. Lysostaphin released from the implants over a 3h period efficiently kills planktonic cells and completely destroys 24h pre-formed biofilms of Staphylococcus aureus. Furthermore, in vivo experiments in a mouse model of critically-sized cranial defects infected with S. aureus show a complete lack of osteogenesis when implants contain only BMP-2, whereas, in the presence of lysostaphin, complete closure of the defect with newly formed mineralized bone tissue is observed. Thus, the new implantable gellan-xanthan hydrogel with diopside and recombinant lysostaphin and BMP-2 shows both osteogenic and antibacterial properties and represents a promising material for the treatment and/or prevention of osteomyelitis after bone trauma.

GeoExplainer: Interpreting Graph Convolutional Networks with geometric masking

For Graph Convolutional Network (GCN), the inherent geometric topology and relationships within spatio-temporal graphs are critical to factors affect its predictive performance. Understanding the decision-making process of GCN remains a challenging task. Existing perturbation-based explanation methods are primarily tailored for Graph Neural Network (GNN) models and focus on how the information of nodes or subgraphs in static graphs impacts GNN predictions. Due to the geometric properties of spatio-temporal graphs, screw theory serves as crucial mathematical tools for describing the geometric relationships and patterns of motion within the geometric space of spatio-temporal graphs. Therefore,we propose GeoExplainer, which utilizes screw theory to geometric masking for spatio-temporal graphs with strong geometric correlations, generating explanations for GCN. Our method perturbs spatio-temporal graphs using geometric masking, and ultimately generating explanations in the form of geometric masks serve as importance mappings for model prediction results. Firstly, we design a screw motion vector model with geometric constraints to describe the spatio-temporal graph in screw system. Then, we design a geometric screw perturbation to generate smooth geometric masks. Finally, we propose a geometric masking approach to optimize and update the geometric mixture mask. We utilize pointwise mutual information (PMI) to maximize the correlation between the model predictions after geometric perturbation and the original predictions, obtaining the optimal geometric mask as an explanation for GCN. The interpretable experiments of ST-GCN, HD-GCN, GCN and GIN models on four datasets with geometric relationships demonstrate that GeoExplainer outperforms eight competitive interpretation methods in quantitative and visualization analysis.

Synthesis of biochar-doped MgO catalyst for highly efficient conversion of glucose into formic acid at low temperatures

Formic acid (FA) is not only a significant platform chemical but also a promising candidate for hydrogen storage. Although homogeneous catalysts are commonly employed for the conversion of carbohydrate biomass into FA, which often come with substantial costs and inadequate reusability. It is noteworthy that there has been an increasing focus on the development of heterogeneous catalysts in this area, which aim to address these limitations and offer advantages such as easier separation and better reusability. Herein, an eco-friendly and efficient method for the catalytic conversion of glucose to FA by utilizing a biochar-MgO (BC-MgO) catalyst, which is prepared through pyrolysis of coconut husk biochar impregnated with magnesium chloride (MgCl2), is proposed. The FA yield at 74.6% is attained from glucose at 60 °C within only 3 h, utilizing a mere 0.1 g of catalyst and hydrogen peroxide (H2O2) amount of 100%. In conjunction with random forest (RF) algorithms and box-behnken experimental designs, further optimization resulted in an increased yield of 79.1%, with a selectivity of 82.7%, in only 3.8 h at 67 °C, employing 0.1 g of catalyst amount and 125% H2O2 amounts. Notably, the BC-MgO catalyst retains its catalytic efficiency over multiple cycles, only requiring a simple straightforward regeneration procedure. The mechanism underlying the glucose conversion to FA reaction system is demystified via kinetic modeling, time-gradient studies, and model compounds experiments. Life cycle assessment (LCA) indicates that when used for glucose conversion to FA, BC-MgO catalysts generate lower CO2 emissions than traditional homogeneous catalysts such as LiOH. Concurrently, an operation cost analysis indicates that operation cost of FA production from glucose is 4.1 $/kg with the BC-MgO catalyst, which is lower than LiOH catalyst (7.5 $/kg), evidencing the dual advantage of economic viability and an improved environmental impact inherent to this novel process.