MetaPredictor: in silico prediction of drug metabolites based on deep language models with prompt engineering.

Metabolic processes can transform a drug into metabolites with different properties that may affect its efficacy and safety. Therefore, investigation of the metabolic fate of a drug candidate is of great significance for drug discovery. Computational methods have been developed to predict drug metabolites, but most of them suffer from two main obstacles: the lack of model generalization due to restrictions on metabolic transformation rules or specific enzyme families, and high rate of false-positive predictions. Here, we presented MetaPredictor, a rule-free, end-to-end and prompt-based method to predict possible human metabolites of small molecules including drugs as a sequence translation problem. We innovatively introduced prompt engineering into deep language models to enrich domain knowledge and guide decision-making. The results showed that using prompts that specify the sites of metabolism (SoMs) can steer the model to propose more accurate metabolite predictions, achieving a 30.4% increase in recall and a 16.8% reduction in false positives over the baseline model. The transfer learning strategy was also utilized to tackle the limited availability of metabolic data. For the adaptation to automatic or non-expert prediction, MetaPredictor was designed as a two-stage schema consisting of automatic identification of SoMs followed by metabolite prediction. Compared to four available drug metabolite prediction tools, our method showed comparable performance on the major enzyme families and better generalization that could additionally identify metabolites catalyzed by less common enzymes. The results indicated that MetaPredictor could provide a more comprehensive and accurate prediction of drug metabolism through the effective combination of transfer learning and prompt-based learning strategies.

admetSAR3.0: a comprehensive platform for exploration, prediction and optimization of chemical ADMET properties.

Absorption, distribution, metabolism, excretion and toxicity (ADMET) properties play a crucial role in drug discovery and chemical safety assessment. Built on the achievements of admetSAR and its successor, admetSAR2.0, this paper introduced the new version of the series, admetSAR3.0, as a comprehensive platform for chemical ADMET assessment, including search, prediction and optimization modules. In the search module, admetSAR3.0 hosted over 370 000 high-quality experimental ADMET data for 104 652 unique compounds, and supplemented chemical structure similarity search function to facilitate read-across. In the prediction module, we introduced comprehensive ADMET endpoints and two new sections for environmental and cosmetic risk assessments, empowering admetSAR3.0 to provide prediction for 119 endpoints, more than double numbers compared to the previous version. Furthermore, the advanced multi-task graph neural network framework offered robust and reliable support for ADMET prediction. In particular, a module named ADMETopt was added to automatically optimize the ADMET properties of query molecules through transformation rules or scaffold hopping. Finally, admetSAR3.0 provides user-friendly interfaces for multiple types of input data, such as SMILES string, chemical structure and batch molecule file, and supports various output types, including digital, chart displays and file downloads. In summary, admetSAR3.0 is anticipated to be a valuable and powerful tool in drug discovery and chemical safety assessment at http://lmmd.ecust.edu.cn/admetsar3/.

Controllable synthesis of porous boron nitride fibers modified by cobalt and nickel oxides for efficient ceftriaxone sodium adsorption from aqueous solution.

Antibiotics can easily enter the water environment through direct or indirect approach, causing environmental pollution and endangering the health of organisms. Therefore, development of highly efficient adsorbent materials to adsorb and remove antibiotics is necessary. Here, cobalt oxide and nickel oxide are uniformly and tightly bonded on the surface of porous boron nitride fibers (PBNFs-NiCo), increasing the number of functional groups (B-O and N-H) and hydrogen bond receptors within PBNFs. The total pore volume and specific surface area of resulting PBNFs-NiCo can reach up to 0.48 cm3g-1and 720.3 m2g-1, respectively. Encouraged by the unique micromorphology and chemical composition mentioned above, PBNFs-NiCo exhibits excellent ceftriaxone sodium (CS) adsorption ability, showing the adsorption capacity and removal efficiency up to 410.9 mg g-1and 96.5%, respectively. Chemical adsorption plays an important role in their adsorption behavior, abiding by Langmuir adsorption theory and pseudo-second-order kinetic equation. Importantly, PBNFs-NiCo exhibits fascinating adsorption effects in surroundings with pH ranging from 4 to 6, 25 °C and varying salt concentrations. This work would establish a practical and feasible foundation for the practical application of PBNFs-NiCo for CS adsorption in aqueous solution.

In silico prediction of potential drug-induced nephrotoxicity with machine learning methods.

In recent years, drug-induced nephrotoxicity has been one of the main reasons for the failure of drug development. Early prediction of the nephrotoxicity for drug candidates is critical to the success of clinical trials. Therefore, it is very important to construct an effective model that can predict the potential nephrotoxicity of compounds. Machine learning methods have been widely used to predict the physicochemical properties, biological activities, and safety assessment of compounds. In this study, we manually collected 777 valid drug data and constructed a total of 72 classification models using nine types of molecular fingerprints combined with different machine learning algorithms. From experimental literature and the US FDA Drugs Database, some marketed drugs were screened for external validation of the models. Finally, three models exhibited good performance in the prediction of nephrotoxicity of both chemical drugs and Chinese herbal medicines. The best model was the support vector machine algorithm combined with CDK graph only fingerprint. Furthermore, the applicability domain of the models was analyzed according to the OECD principles, and we also used the SARpy and information gain methods to find eight substructures that might cause nephrotoxicity, so as to attract attention in the future drug discovery.

DBPP-Predictor: a novel strategy for prediction of chemical drug-likeness based on property profiles.

Evaluation of chemical drug-likeness is essential for the discovery of high-quality drug candidates while avoiding unwarranted biological and clinical trial costs. A high-quality drug candidate should have promising drug-like properties, including pharmacological activity, suitable physicochemical and ADMET properties. Hence, in silico prediction of chemical drug-likeness has been proposed while being a challenging task. Although several prediction models have been developed to assess chemical drug-likeness, they have such drawbacks as sample dependence and poor interpretability. In this study, we developed a novel strategy, named DBPP-Predictor, to predict chemical drug-likeness based on property profile representation by integrating physicochemical and ADMET properties. The results demonstrated that DBPP-Predictor exhibited considerable generalization capability with AUC (area under the curve) values from 0.817 to 0.913 on external validation sets. In terms of application feasibility analysis, the results indicated that DBPP-Predictor not only demonstrated consistent and reasonable scoring performance on different data sets, but also was able to guide structural optimization. Moreover, it offered a new drug-likeness assessment perspective, without significant linear correlation with existing methods. We also developed a free standalone software for users to make drug-likeness prediction and property profile visualization for their compounds of interest. In summary, our DBPP-Predictor provided a valuable tool for the prediction of chemical drug-likeness, helping to identify appropriate drug candidates for further development.

XGraphCDS: An explainable deep learning model for predicting drug sensitivity from gene pathways and chemical structures.

Cancer is a highly complex disease characterized by genetic and phenotypic heterogeneity among individuals. In the era of precision medicine, understanding the genetic basis of these individual differences is crucial for developing new drugs and achieving personalized treatment. Despite the increasing abundance of cancer genomics data, predicting the relationship between cancer samples and drug sensitivity remains challenging. In this study, we developed an explainable graph neural network framework for predicting cancer drug sensitivity (XGraphCDS) based on comparative learning by integrating cancer gene expression information and drug chemical structure knowledge. Specifically, XGraphCDS consists of a unified heterogeneous network and multiple sub-networks, with molecular graphs representing drugs and gene enrichment scores representing cell lines. Experimental results showed that XGraphCDS consistently outperformed most state-of-the-art baselines (R2 = 0.863, AUC = 0.858). We also constructed a separate in vivo prediction model by using transfer learning strategies with in vitro experimental data and achieved good predictive power (AUC = 0.808). Simultaneously, our framework is interpretable, providing insights into resistance mechanisms alongside accurate predictions. The excellent performance of XGraphCDS highlights its immense potential in aiding the development of selective anti-tumor drugs and personalized dosing strategies in the field of precision medicine.

In Silico prediction of inhibitors for multiple transporters via machine learning methods.

Transporters play an indispensable role in facilitating the transport of nutrients, signaling molecules and the elimination of metabolites and toxins in human cells. Contemporary computational methods have been employed in the prediction of transporter inhibitors. However, these methods often focus on isolated endpoints, overlooking the interactions between transporters and lacking good interpretation. In this study, we integrated a comprehensive dataset and constructed models to assess the inhibitory effects on seven transporters. Both conventional machine learning and multi-task deep learning methods were employed. The results demonstrated that the MLT-GAT model achieved superior performance with an average AUC value of 0.882. It is noteworthy that our model excels not only in prediction performance but also in achieving robust interpretability, aided by GNN-Explainer. It provided valuable insights into transporter inhibition. The reliability of our model's predictions positioned it as a promising and valuable tool in the field of transporter inhibition research. Related data and code are available at https://gitee.com/wutiantian99/transporter_code.git.

Up-regulation of sortase-dependent pili in Bifidobacterium longum BBMN68 in response to bile stress enhances its adhesion to HT-29 cells.

Adhesion to gastrointestinal tract is crucial for bifidobacteria to exert their probiotic effects. Our previous work found that bile salts significantly enhance the adhesion ability of Bifidobacterium longum BBMN68 to HT-29 cells. In this study, trypsin-shaving and LC-MS/MS-based surface proteomics were employed to identify surface proteins involved in bile stress response. Among the 829 differentially expressed proteins, 56 up-regulated proteins with a fold change >1.5 were subjected to further analysis. Notably, the minor pilin subunit FimB was 4.98-fold up-regulated in response to bile stress. In silico analysis and RT-PCR confirmed that gene fimB, fimA and srtC were co-transcribed and contributed to the biosynthesis of sortase-dependent pili Pil1. Moreover, scanning electron microscopy and immunogold electron microscopy assays showed increased abundance and length of Pil1 on BBMN68 under bile stress. As the major pilin subunit FimA serves as adhesion component of Pil1, an inhibition assay using anti-FimA antibodies further confirmed the critical role of Pil1 in mediating the adhesion of BBMN68 to HT-29 cells under bile stress. Our findings suggest that the up-regulation of Pil1 in response to bile stress enhances the adhesion of BBMN68 to intestinal epithelial cells, highlighting a novel mechanism of gut persistence in B. longum strains.

Polyvinyl alcohol/chitosan based nanocomposite organohydrogel flexible wearable strain sensors for sports monitoring and underwater communication rescue.

Hydrogel-based flexible wearable sensors have garnered significant attention in recent years. However, the use of hydrogel, a biomaterial known for its high toughness, environmental friendliness, and frost resistance, poses a considerable challenge. In this study, we propose a stepwise construction and multiple non-covalent interaction matching strategy to successfully prepare dynamically physically crosslinked multifunctional conductive hydrogels. These hydrogels self-assembled to form a rigid crosslinked network through intermolecular hydrogen bonding and metal ion coordination chelation. Furthermore, the freeze-thawing process promoted the formation of poly(vinyl alcohol) microcrystalline domains within the amorphous hydrogel network system, resulting in exceptional mechanical properties, including a tensile strength (2.09 ± 0.01 MPa) and elongation at break of 562 ± 12 %. It can lift 10,000 times its own weight. Additionally, these hydrogels exhibit excellent resistance to swelling and maintain good toughness even at temperatures as low as -60 °C. As a wearable strain sensor with remarkable sensing ability (GF = 1.46), it can be effectively utilized in water and underwater environments. Moreover, it demonstrates excellent antimicrobial properties against Escherichia coli (Gram-negative bacteria). Leveraging its impressive sensing ability, we combine signal recognition with a deep learning model by incorporating Morse code for encryption and decryption, enabling information transmission.

Functional Effects of Probiotic Lactiplantibacillus plantarum in Alleviation Multidrug-Resistant Escherichia coli-Associated Colitis in BALB/c Mice Model.

Multidrug-resistant Escherichia coli (MDR-E. coli) is a global health concern. Lactic acid bacteria (LAB) are important probiotics that have beneficial effects on health, and in recent years, their influences in preventing foodborne pathogens-induced colitis have attracted much attention. Therefore, this study aimed to investigate the oral administration of Lactiplantibacillus plantarum NWAFU-BIO-BS29 as an emerging approach to alleviate MDR-E. coli-induced colitis in BALB/c mice model. To illustrate the mode of action of NWAFU-BIO-BS29 interventions with the gut microbiota and immune responses, the changes on the colonic mucosal barrier, regulatory of the gene expressions of inflammatory cytokines, re-modulating the intestinal microflora, and changes in physiological parameters were studied. The results indicated that daily supplementation of 200 µL fresh bacteria for 7 days had ameliorated the associated colitis and partially prevented the infection. The modes of action by ameliorating the inflammatory response, which destructed villous and then affected the intestinal barrier integrity, reducing the secretion of interleukins (6 and ß) and tumor necrosis factor (TNF-α) in serum by 87.88-89.93%, 30.73-35.98%, and 19.14-22.32%, respectively, enhancing the expressions of some epithelial integrity-related proteins in the mouse mucous layer of mucins 2 and 3, Claudin-1, and Occludin by 130.00-661.85%, 27.64-57.35%, 75.52-162.51%, and 139.36-177.73%, respectively, and 56.09-73.58% for toll-like receptor (TLR4) in colon tissues. Notably, the mouse gut microbiota analysis showed an increase in the relative abundance of beneficial bacteria, including Lactobacillus, Bacteriodales bacterium, Candidatus Saccharimonas, Enterorhabdus, and Bacilli. Furthermore, the probiotic promoted the proliferation of epithelia and goblet cells by increasing short-chain fatty acids (SCFAs) levels by 19.23-31.39%. In conclusion, L. plantarum NWAFU-BIO-BS29 has potential applications and can be considered a safe dietary supplement to ameliorate the colitis inflammation symptoms of MDR-E. coli infection.

In silico prediction of drug-induced liver injury with a complementary integration strategy based on hybrid representation.

Drug-induced liver injury (DILI) is one of the major causes of drug withdrawals, acute liver injury and blackbox warnings. Clinical diagnosis of DILI is a huge challenge due to the complex pathogenesis and lack of specific biomarkers. In recent years, machine learning methods have been used for DILI risk assessment, but the model generalization does not perform satisfactorily. In this study, we constructed a large DILI data set and proposed an integration strategy based on hybrid representations for DILI prediction (HR-DILI). Benefited from feature integration, the hybrid graph neural network models outperformed single representation-based models, among which hybrid-GraphSAGE showed balanced performance in cross-validation with AUC (area under the curve) as 0.804±0.019. In the external validation set, HR-DILI improved the AUC by 6.4 %-35.9 % compared to the base model with a single representation. Compared with published DILI prediction models, HR-DILI had better and balanced performance. The performance of local models for natural products and synthetic compounds were also explored. Furthermore, eight key descriptors and six structural alerts associated with DILI were analyzed to increase the interpretability of the models. The improved performance of HR-DILI indicated that it would provide reliable guidance for DILI risk assessment.

Protective effect of plantaricin bio-LP1 bacteriocin on multidrug-resistance Escherichia Coli infection by alleviate the inflammation and modulate of gut-microbiota in BALB/c mice model.

The rapid spread of multidrug-resistant pathogens with the low efficacy of common antibiotics for humans and animals in its clinical therapeutics are a global health concern. Therefore, there is a need to develop new treatment strategies to control them clinically. The study aimed to evaluate the effects of Plantaricin Bio-LP1 bacteriocin produced from Lactiplantibacillus plantarum NWAFU-BIO-BS29 to alleviate the inflammation caused by multidrug-resistance Escherichia Coli (MDR-E. coli) infection in BALB/c mice-model. The focus was given on aspects linked to the mechanism of the immune response. Results indicated that Bio-LP1 had highly promising effects on partially ameliorating MDR-E. coli infection by reducing the inflammatory response through inhibiting the overexpression of proinflammatory-cytokines such as secretion of tumor necrosis factor (TNF-α) and interleukin (IL-6 and IL-ß) and strongly regulated theTLR4 signaling-pathway. Additionally, avoided the villous destruct, colon length shortening, loss of intestinal barrier integrity, and increased disease activity index. Furthermore, significantly increased the relative abundance of beneficial-intestinal-bacteria including Ligilactobacillus, Enterorhabdus, Pervotellaceae, etc. Finally, improved the intestinal mucosal barrier to alleviate the pathological damages and promote the production of short-chain fatty acids (SCFAs) a source of energy for the proliferation. In conclusion, plantaricin Bio-LP1 bacteriocin can be considered a safe alternative to antibiotics against MDR-E. coli-induced intestinal inflammation.

Condensation-assembly synthesis of three-dimensionally porous boron nitride for effective oil removal.

A template-free pyrolysis route has been developed using condensation-assembly precursors made of trimethoxyboroxane (TMB) and melamine (M) to cater the requirements of an industrial real-world environment. The precursors contain abundant B-N bonds and exhibit a high level of interconnectivity, resulting in 3D-PBN with enhanced mechanical properties and the ability to be easily customized in terms of shape. Moreover, 3D-PBN demonstrates rapid adsorption kinetics and excellent reusability, efficiently removing up to 270% of its own weight of fuel within 30 s and being readily regenerated through simple calcination. Even after undergoing 50 cycles, the mechanical properties remain at a remarkable 80%, while the adsorption performance exceed 95%. Furthermore, a comprehensive analysis of thermal behavior from precursor to 3D-PBN has been conducted, leading to the proposal of a molecular-scale evolution process comprising four major steps. This understanding enables us to control the phase reaction and regulate the composition of the products, which is crucial for determining the characteristics of the final product.

Functional Characterization of the Nuclear Receptor Gene SaE75 in the Grain Aphid, Sitobion avenae.

Ecdysteroid hormones are key regulators of insect development and metamorphosis. Ecdysone-inducible E75, a major component of insect ecdysone signaling pathway, has been well characterized in holometabolous insects, however, barely in hemimetabolous species. In this study, a total of four full-length E75 cDNAs from the English grain aphid, Sitobion avenae, were identified, cloned, and characterized. The four SaE75 cDNAs contained 3048, 2625, 2505, and 2179 bp open reading frames (ORF), encoding 1015, 874, 856, and 835 amino acids, respectively. Temporal expression profiles showed that SaE75 expression was low in adult stages, while high in pseudo embryo and nymphal stages. SaE75 was differentially expressed between winged and wingless morphs. RNAi-mediated suppression of SaE75 led to substantial biological impacts, including mortality and molting defects. As for the pleiotropic effects on downstream ecdysone pathway genes, SaHr3 (hormone receptor like in 46) was significantly up-regulated, while Sabr-c (broad-complex core protein gene) and Saftz-f1 (transcription factor 1) were significantly down-regulated. These combined results not only shed light on the regulatory role of E75 in the ecdysone signaling pathway, but also provide a potential novel target for the long-term sustainable management of S. avenae, a devastating global grain pest.

Analytical solutions for time-dependent kinematic three-dimensional magnetic reconnection.

Magnetic reconnection is a process that can rapidly convert magnetic field energy into plasma thermal energy and kinetic energy, and it is also an important energy conversion mechanism in space physics, astrophysics and plasma physics. Research related to analytical solutions for time-dependent three-dimensional magnetic reconnection is extremely difficult. For decades, several mathematical descriptions have been developed regarding different reconnection mechanisms, in which the equations based on magnetohydrodynamics theory outside the reconnection diffusion region are widely accepted. However, the equation set cannot be analytically solved unless specified constraints are imposed or the equations are reduced. Based on previous analytical methods for kinematic stationary reconnection, here the analytical solutions for time-dependent kinematic three-dimensional magnetic reconnection are discussed. In contrast to the counter-rotating plasma flows that existed in steady-state reconnection, it is found that spiral plasma flows, which have never been reported before, can be generated if the magnetic field changes exponentially with time. These analyses reveal new scenarios for time-dependent kinematic three-dimensional magnetic reconnection, and the deduced analytical solutions could improve our understanding of the dynamics involved in reconnection processes, as well as the interactions between the magnetic field and plasma flows during magnetic reconnection.

A multitask GNN-based interpretable model for discovery of selective JAK inhibitors.

The Janus kinase (JAK) family plays a pivotal role in most cytokine-mediated inflammatory and autoimmune responses via JAK/STAT signaling, and administration of JAK inhibitors is a promising therapeutic strategy for several diseases including COVID-19. However, to screen and design selective JAK inhibitors is a daunting task due to the extremely high homology among four JAK isoforms. In this study, we aimed to simultaneously predict pIC50 values of compounds for all JAK subtypes by constructing an interpretable GNN multitask regression model. The final model performance was positive, with R2 values of 0.96, 0.79 and 0.78 on the training, validation and test sets, respectively. Meanwhile, we calculated and visualized atom weights, followed by the rank sum tests and local mean comparisons to obtain key atoms and substructures that could be fine-tuned to design selective JAK inhibitors. Several successful case studies have demonstrated that our approach is feasible and our model could learn the interactions between proteins and small molecules well, which could provide practitioners with a novel way to discover and design JAK inhibitors with selectivity.

Lactic acid bacteria isolated from Chinese traditional fermented milk as novel probiotic strains and their potential therapeutic applications.

Lactic acid bacteria (LAB) are believed to have health-promoting properties to the host and can be used in therapeutics interventions; intriguingly, they have the property to produce bio-preservatives substances. Therefore, this study aimed to mine probiotics and evaluate their safety, functional properties, and cholesterol-lowering capability. Seven potential probiotic strains were compared from 56 LAB strains isolated from traditional Chinese fermented milk. The results showed that all tested strains are tolerant to gastric acidity (45.5-83.26) and bile salts (11.92-92.91%) and have antibacterial activity against Staphylococcus aureus ATCC25923 and Escherichia coli ATCC25922. Likewise, it lowered the cholesterol levels in vitro by live cells (26.57-45.76%) and dead cells (29.53-50.97%) with remarkable aggregation ability (13.8-43.71%). Antioxidant properties and produce short chain fatty acids (SCFAs) were strain-dependent features. Upon assessment of the safety, Enterococcus faecium NWAFU-BIO-AS14 exhibited virulence factors genes (VFs) of (mur-2ed, odc, and tet(K)) and + hemolysis activity. While Enterococcus faecium NWAFU-BIO-A-B24 and Limosilactobacillus fermentum NWAFU-BIO-B-S6 have VFs of (odc, vanC2, and ant(6)-Ia). Limosilactobacillus fermentum NWAFU-BIO-D-B2 has only (odc). Thus, they are not considered as safe probiotics. In contrast, Lactiplantibacillus plantarum NWAFU-BIO-BS29, Companilactobacillus crustorum NWAFU-BIO-AS16, and Lactobacillus gallinarum NWAFU-BIO-D-S7 are the safest and best strains, respectively, due to the absence of 16 VFs and their sensitivity to antibiotics such as kanamycin, erythromycin, tetracycline, gentamycin, vancomycin, streptomycin, chloramphenicol, and ampicillin. Accordingly, these strains have a high potentiality to be used as starter cultures or safely applied as perfect probiotics in functionals food and feed. Supplementary Information: The online version contains supplementary material available at 10.1007/s13205-022-03403-z.

Probiotic of Lactiplantibacillus plantarum NWAFU-BIO-BS29 Isolated from Chinese Traditional Fermented Milk and Its Potential Therapeutic Applications Based on Gut Microbiota Regulation.

Lactic acid bacteria are one of the bioresources that can promote the host's health and have potential therapeutic applications. This study aimed to evaluate the probiotic properties of novel Lactiplantibacillus plantarum NWAFU-BISO-BS29 isolated in vitro from traditional Chinese fermented milk, assess its safety, and study its interaction with the gut microbiota using a BALB/c mouse model. The findings reveal that this strain had a high tolerance to gastric acidity (64.4%) and bile salts (19.83-87.92%) with remarkable auto-aggregation and co-aggregation abilities (33.01-83.96%), respectively. Furthermore, it lowered the cholesterol levels in dead cells (44.02%) and live cells (34.95%) and produced short-chain fatty acids (SCFAs). Likewise, it showed good antioxidant properties and strong antipathogen activity against Escherichia coli and Staphylococcus aureus with inhibition zones at 21 and 25 mm, respectively. The safety assessment results indicate that all of the virulence factor genes were not detected in the whole DNA; additionally, no hemolysis or resistance to antibiotics commonly used in food and feed was observed. Interestingly, the 16S rRNA gene sequencing of the mouse gut microbiota showed a marked alteration in the microbial composition of the administrated group, with a noticeable increase in Firmicutes, Patescibacteria, Campylobacterota, Deferribacterota, Proteobacteria, and Cyanobacteria at the phylum level. The modulation of gut microbial diversity significantly improved the production of SCFCs due to the abundance of lactobacillus genera, which was consistent with the functional gene predictive analysis and is believed to have health-promoting properties. Based on these results, our novel strain is considered a safe and good probiotic and could hold high potential to be used as a starter culture or to safely supplement functional foods as a probiotic and may provide new insights into therapeutic interventions.

DGL-LifeSci: An Open-Source Toolkit for Deep Learning on Graphs in Life Science.

Graph neural networks (GNNs) constitute a class of deep learning methods for graph data. They have wide applications in chemistry and biology, such as molecular property prediction, reaction prediction, and drug-target interaction prediction. Despite the interest, GNN-based modeling is challenging as it requires graph data preprocessing and modeling in addition to programming and deep learning. Here, we present Deep Graph Library (DGL)-LifeSci, an open-source package for deep learning on graphs in life science. Deep Graph Library (DGL)-LifeSci is a python toolkit based on RDKit, PyTorch, and Deep Graph Library (DGL). DGL-LifeSci allows GNN-based modeling on custom datasets for molecular property prediction, reaction prediction, and molecule generation. With its command-line interfaces, users can perform modeling without any background in programming and deep learning. We test the command-line interfaces using standard benchmarks MoleculeNet, USPTO, and ZINC. Compared with previous implementations, DGL-LifeSci achieves a speed up by up to 6×. For modeling flexibility, DGL-LifeSci provides well-optimized modules for various stages of the modeling pipeline. In addition, DGL-LifeSci provides pretrained models for reproducing the test experiment results and applying models without training. The code is distributed under an Apache-2.0 License and is freely accessible at https://github.com/awslabs/dgl-lifesci.

Anti-Adhesion Effects of Lactobacillus Strains on Caco-2 Cells Against Escherichia Coli and Their Application in Ameliorating the Symptoms of Dextran Sulfate Sodium-Induced Colitis in Mice.

The beneficial effects of probiotics on ameliorating ulcerative colitis (UC) have attracted much attention in recent years. Nevertheless, the number of these identified probiotics is still limited. In addition, the adhesion abilities of probiotics are considered to be a key determinant for probiotic efficacy. However, the relationship between the adhesion abilities of probiotics and their role in ameliorating UC has been poorly studied to date. This study measured the adhesion abilities of four Lactobacillus strains to Caco-2 cells and their anti-adhesion effects on Caco-2 cells against pathogenic bacteria, as well as their application in ameliorating the symptoms of dextran sulfate sodium-induced UC, and further illustrated the relationship between these two potential probiotic properties of probiotics and their beneficial effects on UC. Results suggested that the adhesion abilities of the four tested Lactobacillus strains exists highly strain-specific and the mechanisms of their anti-adhesion effect on Caco-2 cells against Escherichia coli may be different. Moreover, all these strains had promising effects on ameliorating UC by reducing inflammatory response and improving the intestinal mucosal barrier function, as well as promoting the production of SCFAs. In conclusion, the four tested Lactobacillus strains can be considered as alternative dietary supplements in alleviating UC. In addition, it could be concluded that there is no significant correlation between the adhesion abilities of probiotics and their role in ameliorating UC, which further illustrated that the adhesion properties of probiotics in vitro may not be suitable as the key criterion for screening potential strains with UC-alleviating effects.