Large-scale comparison of machine learning methods for profiling prediction of kinase inhibitors.

Conventional machine learning (ML) and deep learning (DL) play a key role in the selectivity prediction of kinase inhibitors. A number of models based on available datasets can be used to predict the kinase profile of compounds, but there is still controversy about the advantages and disadvantages of ML and DL for such tasks. In this study, we constructed a comprehensive benchmark dataset of kinase inhibitors, involving in 141,086 unique compounds and 216,823 well-defined bioassay data points for 354 kinases. We then systematically compared the performance of 12 ML and DL methods on the kinase profiling prediction task. Extensive experimental results reveal that (1) Descriptor-based ML models generally slightly outperform fingerprint-based ML models in terms of predictive performance. RF as an ensemble learning approach displays the overall best predictive performance. (2) Single-task graph-based DL models are generally inferior to conventional descriptor- and fingerprint-based ML models, however, the corresponding multi-task models generally improves the average accuracy of kinase profile prediction. For example, the multi-task FP-GNN model outperforms the conventional descriptor- and fingerprint-based ML models with an average AUC of 0.807. (3) Fusion models based on voting and stacking methods can further improve the performance of the kinase profiling prediction task, specifically, RF::AtomPairs + FP2 + RDKitDes fusion model performs best with the highest average AUC value of 0.825 on the test sets. These findings provide useful information for guiding choices of the ML and DL methods for the kinase profiling prediction tasks. Finally, an online platform called KIPP ( https://kipp.idruglab.cn ) and python software are developed based on the best models to support the kinase profiling prediction, as well as various kinase inhibitor identification tasks including virtual screening, compound repositioning and target fishing.

BioTreasury: a community-based repository enabling indexing and rating of bioinformatics tools.

The exponential growth of bioinformatics tools in recent years has posed challenges for scientists in selecting the most suitable one for their data analysis assignments. Therefore, to aid scientists in making informed choices, a community-based platform that indexes and rates bioinformatics tools is urgently needed. In this study, we introduce BioTreasury ( http://biotreasury.rjmart.cn ), an integrated community-based repository that provides an interactive platform for users and developers to share their experiences in various bioinformatics tools. BioTreasury offers a comprehensive collection of well-indexed bioinformatics software, tools, and databases, totaling over 10,000 entries. In the past two years, we have continuously improved and maintained BioTreasury, adding several exciting features, including creating structured homepages for every tool and user, a hierarchical category of bioinformatics tools and classifying tools using large language model (LLM). BioTreasury streamlines the tool submission process with intelligent auto-completion. Additionally, BioTreasury provides a wide range of social features, for example, enabling users to participate in interactive discussions, rate tools, build and share tool collections for the public. We believe BioTreasury can be a valuable resource and knowledge-sharing platform for the biomedical community. It empowers researchers to effectively discover and evaluate bioinformatics tools, fostering collaboration and advancing bioinformatics research.

A Conceptual Protocol for a Single-Session Solution-Focused Brief Therapy for Medication Adherence Intervention Delivered by General Providers.

New approaches to medication adherence interventions are needed. This manuscript presents a highly structured protocol of a single-session solution-focused brief therapy (SFBT) for medication adherence intervention (SFBT-MAI) delivered by general providers. It conceptually integrates the procedure of tailored interventions, techniques of SFBT, and the four steps of Qitang Lin' conceptualization of single-session SFBT. With specific techniques and examples to reduce operational difficulties, the SFBT-MAI includes two parts. The first part focuses on selecting non-adherent patients and clarifying their barriers to medication adherence. The second part focuses on individualized interventions with four steps: closing, hoping, empowering, and changing and acting. It is hoped that this work will improve the effectiveness of medication adherence interventions for patients with coronary heart disease and to promote the use of brief psychological interventions in clinical practice.

Enhanced Flame Retardancy of Styrene-Acrylic Emulsion Based Damping Composites Based on an APP/EG Flame-Retardant System.

Developing flame-retarded styrene-acrylic emulsion (SAE) based damping composites is a challenging task because of their very high flammability. A promising approach is the synergistic combination of expandable graphite (EG) and ammonium polyphosphate (APP). In this study, the surface modification of APP was modified by commercial titanate coupling agent ndz-201 through ball milling, and the SAE-based composite material was prepared with SAE and different ratios of modified ammonium polyphosphate (MAPP) and EG. The surface of MAPP was successfully chemically modified by NDZ-201 through scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FTIR), X-ray diffraction analysis (XRD), Energy Dispersion Spectroscopy (EDS), and contact angle. The effects of different ratios of MAPP and EG on the dynamic and static mechanical properties and flame retardancy of composite materials were explored. The results showed that when MAPP:EG = 1:4, the limiting oxygen index (LOI) of the composite material was 52.5%, and the vertical burning test (UL-94) was at the V0 level. Its LOI increased by 141.9% compared to the composite materials without flame retardant. The optimized formulation of MAPP and EG in SAE-based damping composite materials showed a significant synergistic effect on the flame retardancy of the composite material.

DeepCancerMap: A versatile deep learning platform for target- and cell-based anticancer drug discovery.

Discovering new anticancer drugs has been widely concerned and remains an open challenge. Target- and phenotypic-based experimental screening represent two mainstream anticancer drug discovery methods, which suffer from time-consuming, labor-intensive, and high experimental costs. In this study, we collected 485,900 compounds involving in 3,919,974 bioactivity records against 426 anticancer targets and 346 cancer cell lines from academic literature, as well as 60 tumor cell lines from NCI-60 panel. A total of 832 classification models (426 target- and 406 cell-based predictive models) were then constructed to predict the inhibitory activity of compounds against targets and tumor cell lines using FP-GNN deep learning method. Compared to the classical machine learning and deep learning methods, the FP-GNN models achieve considerable overall predictive performance, with the highest AUC values of 0.91, 0.88, 0.91 for the test sets of targets, academia-sourced and NCI-60 cancer cell lines, respectively. A user-friendly webserver called DeepCancerMap and its local version were developed based on these high-quality models, enabling users to perform anticancer drug discovery-related tasks including large-scale virtual screening, profiling prediction of anticancer agents, target fishing, and drug repositioning. We anticipate this platform to accelerate the discovery of anticancer drugs in the field. DeepCancerMap is freely available at https://deepcancermap.idruglab.cn.

A multi-task FP-GNN framework enables accurate prediction of selective PARP inhibitors.

PARP (poly ADP-ribose polymerase) family is a crucial DNA repair enzyme that responds to DNA damage, regulates apoptosis, and maintains genome stability; therefore, PARP inhibitors represent a promising therapeutic strategy for the treatment of various human diseases including COVID-19. In this study, a multi-task FP-GNN (Fingerprint and Graph Neural Networks) deep learning framework was proposed to predict the inhibitory activity of molecules against four PARP isoforms (PARP-1, PARP-2, PARP-5A, and PARP-5B). Compared with baseline predictive models based on four conventional machine learning methods such as RF, SVM, XGBoost, and LR as well as six deep learning algorithms such as DNN, Attentive FP, MPNN, GAT, GCN, and D-MPNN, the evaluation results indicate that the multi-task FP-GNN method achieves the best performance with the highest average BA, F1, and AUC values of 0.753 ± 0.033, 0.910 ± 0.045, and 0.888 ± 0.016 for the test set. In addition, Y-scrambling testing successfully verified that the model was not results of chance correlation. More importantly, the interpretability of the multi-task FP-GNN model enabled the identification of key structural fragments associated with the inhibition of each PARP isoform. To facilitate the use of the multi-task FP-GNN model in the field, an online webserver called PARPi-Predict and its local version software were created to predict whether compounds bear potential inhibitory activity against PARPs, thereby contributing to design and discover better selective PARP inhibitors.

FP-GNN: a versatile deep learning architecture for enhanced molecular property prediction.

Accurate prediction of molecular properties, such as physicochemical and bioactive properties, as well as ADME/T (absorption, distribution, metabolism, excretion and toxicity) properties, remains a fundamental challenge for molecular design, especially for drug design and discovery. In this study, we advanced a novel deep learning architecture, termed FP-GNN (fingerprints and graph neural networks), which combined and simultaneously learned information from molecular graphs and fingerprints for molecular property prediction. To evaluate the FP-GNN model, we conducted experiments on 13 public datasets, an unbiased LIT-PCBA dataset and 14 phenotypic screening datasets for breast cell lines. Extensive evaluation results showed that compared to advanced deep learning and conventional machine learning algorithms, the FP-GNN algorithm achieved state-of-the-art performance on these datasets. In addition, we analyzed the influence of different molecular fingerprints, and the effects of molecular graphs and molecular fingerprints on the performance of the FP-GNN model. Analysis of the anti-noise ability and interpretation ability also indicated that FP-GNN was competitive in real-world situations. Collectively, FP-GNN algorithm can assist chemists, biologists and pharmacists in predicting and discovering better molecules with desired functions or properties.

Boosted Charge Transfer in Twinborn α-(Mn2O3-MnO2) Heterostructures: Toward High-Rate and Ultralong-Life Zinc-Ion Batteries.

Aqueous ZIBs are one of the most promising next-generation rechargeable batteries because of the high capacity, high hydrogen evolution overpotential, and chemically stable reversible plating/stripping of the zinc electrode in the mild aqueous electrolyte. However, there are limited cathode materials that can store Zn2+ reversibly with superior cycling and rate capability. Herein, hierarchically porous nanorods composed of twinborn α-(Mn2O3-MnO2) heterostructures are proposed as a robust cathode for Zn storage. Thanks to the hierarchically porous nanorod morphology and the abundant interface of the heterostructures involving a built-in electric field, the as-obtained twinborn α-(Mn2O3-MnO2) electrode delivers a high capacity of 170 mA h g-1 for 2000 cycles at 500 mA g-1 and shows an excellent rate capability of up to 1.5 A g-1 with a capacity of 124 mA h g-1. The inspiring results achieved exhibit the enormous potential of the high-performance heterostructure cathode for fast and stable ZIBs.

Enhancement of lignocellulosic degradation in high-moisture alfalfa via anaerobic bioprocess of engineered Lactococcus lactis with the function of secreting cellulase.

BACKGROUND: Butyric fermentation and a substantial loss of dry matter (DM) often occur in alfalfa silage during the rainy season, which is not conducive to subsequent biofuel production. Currently, there have been negative effects on the combination of cellulases and lactic acid bacteria (LAB) on processing high-moisture alfalfa silage; however, transgenically engineered LAB strains that secrete cellulase have been proposed as an alternative approach to avoid the above problem. The objective of the present study was to construct engineered Lactococcus lactis strains with high-efficiency secretory-expressing cellulase genes from Trichoderma reesei and to investigate the effects of the combination of transgenically engineered L. lactis strains HT1/pMG36e-usp45-bgl1, HT1/pMG36e-usp45-cbh2, and HT1/pMG36e-usp45-egl3 (HT2) on fermentation quality, structural carbohydrate degradability and nonstructural carbohydrate fermentation kinetics of high-moisture alfalfa silage treated without additive as a negative control (Control), or/and with cellulase (EN), wild-type L. lactis subsp. lactis MG1363 (HT1) and the combination of HT1 and EN (HT1 + EN) as positive additive controls. RESULTS: Engineered L. lactis strains were successfully constructed and efficiently secreted endoglucanase (1118 mU/mL), cellobiohydrolase (222 mU/mL), and ß-glucosidase (131 mU/mL) and had high filter paper activity (236 mU/mL). Ensiling experiments verified that HT2 obtained the highest fermentation quality score (83.6) and most efficiently processed high-moisture alfalfa silage, demonstrated by a low pH (4.49) and ammonia-N content (106 g/kg nitrogen) and a high lactic acid content (67.1 g/kg DM) and without butyric acid. Change curves of structural carbohydrates revealed that HT2 degraded more lignocelluloses, demonstrated by the lowest contents of neutral detergent fibre, acid detergent fibre, cellulose and hemicellulose after ensiling for 60 days. Kinetic analysis showed that the most residual water-soluble carbohydrates, glucose, fructose and xylose generated by lignocellulose degradation were produced by HT2, followed by HT1 + EN. The HT2-treated silages had the highest DM recovery, had the fewest Clostridia spores, emitted a fragrance and were not sticky. CONCLUSION: HT2 improved the conversion of lignocellulose to sugars and processed high-moisture alfalfa silage efficiently. This is a novel strategy that can be used to enhance lignocellulosic degradation in high-moisture alfalfa via a bioprocess with transgenically engineered L. lactis strains, which could enhance the development of alfalfa as a biomass feedstock and promote second-generation biofuel development in the rainy season.

Visible detection of copper ions using a fluorescent probe based on red carbon dots and zirconium metal-organic frameworks.

A highly selective and sensitive ratiometric fluorescent probe for copper ions (Cu2+) in aqueous solution based on a blue zirconium metal-organic framework [UiO-66-(COOH)2] with red bandgap fluorescent carbon quantum dots (R-BF-CQDs) is developed. UiO-66-(COOH)2 displays broad ligand-centered (LC) emission and R-BF-CQDs shows red emission. Interestingly, the remaining -NH2 at the surface of the prepared R-BF-CQDs can coordinate Cu2+ efficiently and produce a strong visible absorption, which overlaps the emission of UiO-66-(COOH)2. There is also a strong coordination capacity of the Cu2+ ions with R-BF-CQDs and UiO-66-(COOH)2, which allows the energy donor and the energy receptor to be closer together. Both of these two conditions provide the possibility of spectral energy transfer. In practice, the LC emission of UiO-66-(COOH)2 becomes weaker with increasing concentration of Cu2+, whereas the red fluorescence of R-BF-CQDs increases slightly.

Eu(III)-functionalized In-MOF (In(OH)bpydc) as fluorescent probe for highly selectively sensing organic small molecules and anions especially for CHCl3 and MnO4.

A highly fluorescent metal-organic framework (In-MOF, In(OH)bpydc (bpydc=biphenyl-5,5'-dicarboxylate)) imparting additional luminescent Eu3+ ions has been synthesized for sensing organic molecules and anions, especially chloroform and MnO4-, both of them have a significant spectral response. This Eu3+-incorporated hybrid (In-MOF-Eu) has been fully characterized, which reveals large surface area, great pore size, favorable thermal stability, excellent luminescence as well as good fluorescence stability. On the one hand, the dual-emitting MOF hybrid exhibits a distinctly different response to the chloroform, one of which is almost unchanged, while the other intrinsic broad emission has a remarkable enhancement, leading to an obvious change in color; on the other hand, it also shows a high selectivity for detection of MnO4- because of the largest quenching effect on the luminescent emission, which results in a dark emission. This work represents a new approach for the fabrication of highly selective and sensitive probe for detection of chloroform and MnO4-, as it underlines the opportunity of MOFs with prominent optical properties to explore the existing target substance in the environment.

A dual-emission probe to detect moisture and water in organic solvents based on green-Tb3+ post-coordinated metal-organic frameworks with red carbon dots.

Using p-phenylenediamine as a precursor, p-carbon dots (p-CDs) with strong red-light emission were encapsulated into a metal-organic framework (MOF) followed by introduction of green light-emitting Tb3+ to form a two-color light-emitting hybrid (Tb3+@p-CDs/MOF). The as-prepared fluorescent-functionalized MOF not only maintained the excellent optical properties of p-CDs and Tb3+ to give strong emission, but also had good chemical and physical properties. The chosen p-CDs were aggregated readily in water, which led to only very weak photoluminescence, whereas the opposite effect was noted in the organic solvents ethanol, dimethylformamide and cyclopropane. Therefore, the as-prepared hybrid showed different color light emission in water or organic solvents, and acted as a ratiometric and colorimetric fluorescent probe to detect water content in organic solvents. Moreover, this hybrid also served as a ratiometric luminescent sensor for detection of relative humidity (RH): the ratio of light intensity at 545 nm to that at 605 nm increased linearly with increasing RH from 33.0% to 85.1% in the atmosphere.

Lanthanides post-functionalized indium metal-organic frameworks (MOFs) for luminescence tuning, polymer film preparation and near-UV white LED assembly.

A class of hybrid materials based on indium 2,2'-bipyridine-5,5'-dicarboxylate metal-organic frameworks, In(OH)bpydc, was synthesized by postsynthetic introduction with lanthanide ions (Eu3+, Tb3+ and Sm3+). The structure, thermal stability, morphology and more detailed information about these materials were characterized by XRD, DSC, BET, FTIR, SEM and so forth. The further study of luminescent properties in detail showed that these compounds possess characteristic emission, and the In-MOF-Eu maintains different colors of light from blue-green to red under different excitation wavelengths (excited at 400 nm to 320 nm), which includes the near-white light region (the color coordinates are X = 0.34, Y = 0.36). It is a remarkable fact that the trend of ligand-central emission is opposite to that of the characteristic emission of Eu3+. Moreover, a kind of thin film and assembled white light near-UV LED based on the optically lanthanide-functionalized MOFs was prepared in order to extend their potential applications; both of them lead to desirable white light (X = 0.34, Y = 0.36; X = 0.35, Y = 0.37). In addition, the matrix does not affect the white luminescence.

Feasibility and safety of fiber optic micro-imaging in canine peripheral airways.

PURPOSE: To assess the feasibility and safety of imaging canine peripheral airways (<1 mm) with an experimental micro-imaging fiber optic bronchoscope. METHODS: Twenty healthy dogs were scoped with a micro-imaging fiber optic bronchoscope (0.8 mm outer diameter). Images at various levels of the bronchioles, mucosal color, and tracheal secretions were recorded. The apparatus was stopped once it was difficult to insert. CT imaging was performed simultaneously to monitor progression. The safety of the device was evaluated by monitoring heart rate (HR), respiratory rate (RR), mean artery pressure (MAP), peripheral oxygen saturation (SpO2) and arterial blood gases (partial pressure of arterial carbon-dioxide, PaCO2, partial pressure of arterial oxygen, PaO2, and blood pH). RESULTS: (1) According to the CT scan, the micro-imaging fiber was able to access the peripheral airways (<1 mm) in canines. (2) There was no significant change in the values of HR, MAP, pH and PaCO2 during the procedure (P>0.05). Comparing pre-manipulation and post-manipulation values, SpO2 (F = 13.06, P<0.05) and PaO2 (F = 3.01, P = 0.01) were decreased, whereas RR (F = 3.85, P<0.05) was elevated during the manipulation. (3) Self-limited bleeding was observed in one dog; severe bleeding or other complications did not occur. CONCLUSION: Although the new apparatus had little effect on SpO2, PaO2 and RR, it can probe into small peripheral airways (<1 mm), which may provide a new platform for the early diagnosis of bronchiolar diseases.

Laboratory test of a visual sputum suctioning system.

BACKGROUND: Conventional sputum suctioning is a routine clinical practice, but complications may arise from the blind manipulation of the catheter. Recently, a visual sputum suctioning system (VSSS) was developed, and we tested this new system in a laboratory setting. METHODS: We used coagulant concentrations of 1.5% and 3.0% to simulate mucus and sputum. Conventional single-lumen and triple-lumen catheters were inserted separately into a beaker for sputum suctioning (15 s, 200 mm Hg). A micro-imaging fiber was integrated into the triple-lumen catheter to create the VSSS. The single-lumen catheter and the VSSS were inserted separately into the mouth cavity, the nasal cavity, the tracheostomy tube, and the endotracheal tube of a human analog model for further comparisons. RESULTS: As the suction channel of the triple-lumen catheter was reduced by 46.8%, the amount of simulant it suctioned was significantly less than that suctioned by the single-lumen catheter. However, under real-time guidance, the VSSS suctioned more simulant than the conventional single-lumen catheter in the human analog model. CONCLUSION: Sputum suctioning with the VSSS was feasible. Because of its real-time imaging guidance, the efficiency of the VSSS procedure was greater than that of the conventional single-lumen catheter. Therefore, this system may provide a new platform for sputum suctioning.