Secondary traumatic stress and vicarious posttraumatic growth in oncology nurses: the mediating role of empathy.

Background: The relationship between secondary traumatic stress (STS), empathy, and vicarious post-traumatic growth (VPTG) in oncology nurses is unclear. Understanding these links is crucial for improving nurse well-being and patient care quality. Objective: This study aimed to investigate the relationships between STS, empathy, and VPTG among oncology nurses. Methods: This cross-sectional study was conducted in a multicentric setting. Data collection involved oncology nurses completing questionnaires assessing STS, empathy, and VPTG levels. Data analysis included correlation analyses, multiple stepwise regression analysis, and structural equation modeling (SEM) to examine the interrelationships between these variables. The study adhered to the STROBE checklist. Results: A total of 391 oncology nurses participated in the study. They showed moderate to low levels of VPTG and high levels of STS. STS exhibited a negative association with VPTG, while empathy demonstrated a positive direct association with both VPTG and STS. Moreover, SEM indicated that empathy mediated the relationship between STS and VPTG, with a partial mediating effect of 0.127. Factors such as receiving psychological training, educational attainment, STS, and empathy collectively explained 24% of the variance in VPTG. Conclusion: Our findings highlighted the negative correlation between STS and VPTG among oncology nurses. Additionally, empathy was found to mediate the relationship between STS and VPTG, suggesting it plays a significant role in influencing VPTG. Implications for practice: To aid oncology nurses, interventions should focus on reducing STS and enhancing empathy. Strategies like resilience workshops, peer support, and stress management can foster VPTG. Creating a supportive work environment is crucial for nurses' well-being and quality patient care.

Neutral effect of Zishen Yutai Pill on frozen-thawed embryo transfer: a propensity score matching study.

Objective: To investigate whether using Zishen Yutai Pills (ZYP) following embryo transfer would affect the live birth rate in frozen-thawed embryo transfer (FET) cycles. Methods: A retrospective analysis was performed on 15044 FET cycles in the Reproductive Medicine Center of The Affiliated Chenggong Hospital of Xiamen University from January 2013 to December 2020. Patients who used Zishen Yutai Pills were defined as Zishen Yutai Pills Group (ZYP, n=2735), while patients who did not use them were defined as Non- Zishen Yutai Pills Group (Non-ZYP, n=12309). The propensity score matching method was used to control for potential confounders between the two groups, and logistic regression analysis was also used to assess whether using ZYP would affect the live birth rate. Results: After propensity score matching, basic characteristics were similar between the two groups. Using ZYP did not increase the pregnancy rate (51.5% vs. 52.7%, P=0.372), and live birth rate (43.0% vs. 44.7%, P=0.354). This was also confirmed by the logistic regression analysis results (OR=0.95, 95%CI=0.85-1.06). In the subgroup analysis of the endometrial preparation protocols, however, it was found that the use of ZYP in patients with natural cycles increased the live birth rate (47.4% vs. 41.5%, P=0.004). A significant interaction between endometrial preparation and ZYP was found (OR=1.38, 95%CI=1.07-1.79) in the multivariate model. Conclusion: The use of ZYP may not improve the live birth rate of unselected patients in FET cycles. However, a future study is needed on the effect of ZYP in natural cycles for endometrial preparation.

Review of the potential value of serum interleukin levels as prognostic biomarkers of liver failure.

Liver failure (LF) is prevalent in China and is characterized by complex pathogenesis, challenging clinical management, poor prognosis, and rising incidence and mortality rates. The immune status is an important factor affecting LF prognosis. Interleukins (Ils) are a type of cytokine that act and interact with multiple cells, including immune cells. These signaling molecules play important roles in intercellular information transmission, including the regulation of immune cells; mediation of the activation, proliferation, and differentiation of T and B cells; and orchestration of the inflammatory response. To date, many studies have explored the correlation between IL expression and liver disease prognosis, but few studies have evaluated Ils as the prognostic biomarkers of LF. This article reviews the potential use of Ils as the prognostic biomarkers of LF. Particularly, it evaluates the predictive values of IL-21, IL-22, and IL-31, the three often overlooked yet promising prognostic biomarkers, in predicting susceptibility to LF. Harnessing biomarkers for early prognostic insights can facilitate tailored treatment strategies and enhance patient survival. Thus, this article focuses on the identification of IL-21, IL-22, and IL-33 as biomarkers in preclinical and clinical studies on LF and reviews their role as biomarkers in the pathogenesis and diagnosis of LF.

Unlocking reproducible transcriptomic signatures for acute myeloid leukaemia: Integration, classification and drug repurposing.

Acute myeloid leukaemia (AML) is a highly heterogeneous disease, which lead to various findings in transcriptomic research. This study addresses these challenges by integrating 34 datasets, including 26 control groups, 6 prognostic datasets and 2 single-cell RNA sequencing (scRNA-seq) datasets to identify 10,000 AML-related genes (ARGs). We focused on genes with low variability and high consistency and successfully discovered 191 AML signatures (ASs). Leveraging machine learning techniques, specifically the XGBoost model and our custom framework, we classified AML subtypes with both scRNA-seq and bulk RNA-seq data, complementing the ELN2022 classification approach. Our research also identified promising treatments for AML through drug repurposing, with solasonine showing potential efficacy for high-risk AML patients, supported by molecular docking and transcriptomic analyses. To enhance reproducibility and customizability, we developed CSAMLdb, a user-friendly database platform. It facilitates the reuse and personalized analysis of nearly all results obtained in this research, including single-gene prognostics, multi-gene scoring, enrichment analysis, machine learning risk assessment, drug repositioning analysis and literature abstract named entity recognition. CSAMLdb is available at http://www.csamldb.com.

Study on VOCs of Fishmeal during Storage Based on HS-SPME-GC-MS.

Fishmeal is widely used in the feed industry as the main protein material. The freshness grade directly affects the quality of the fishmeal. During the storage of fishmeal, the odor would change accordingly as the freshness grades decreased. To study the characteristic volatile organic compounds (VOCs) of fishmeal, stored at 25 °C and 80%RH with different freshness grades, headspace solid-phase microextraction combined with gas chromatography-mass spectrometry (HS-SPME-GC-MS) was used to analyze. The single-factor test was chosen for 50/30 µm divinylbenzene/carboxe/polydimethylsiloxane (DVB/CAR/PDMS) fiber. The equilibration time of 24 min, the extraction time of 60 min, the extraction temperature of 87 °C, and the addition of a saturated saline volume of 4 mL were determined by Box-Behnken design. There were 15 common VOCs detected during storage, the relative contents of acids increased significantly, ketones, aldehydes, esters, and nitrogen-containing compounds increased, and aromatic compounds and alcohols decreased. Combined with freshness indexes, volatile base nitrogen (VBN) and acid value (AV), hexadecanoic acid, tetradecanoic acid, methyl (Z)-N-hydroxybenzenecarboximidate, (Z)-hexadec-9-enoic acid, 6-ethoxy-2,2,4-trimethyl-3,4-dihydro-1H-quinoline, octadecanal, and [(Z)-octadec-9-enyl] acetate were determined as the characteristic VOCs based on the PLS-DA model. This study may provide data support for the development of fishmeal freshness-detecting instruments.

Numerical Simulation and Analysis of the Airflow Field in the Crushing Chamber of the Hammer Mill.

The airflow dynamics within hammer mills' crushing chambers significantly affect material crushing and screening. Understanding the crushing mechanism necessitates studying the airflow distribution. Using a self-built crushing test platform and computational fluid dynamics (CFD) simulations, we investigated the impact of screen aperture size, rotor speed, hammer-screen clearance, hammer quantity, and mass flow rate on airflow distribution within the rotor region, circulation layer, and screen apertures. Results indicated generally uniform axial static pressure distribution within the rotor region, with radial gradients. Increased rotor speed improved radial static pressure gradients, while higher mass flow rates reduced them. The highest airflow velocity within the circulation layer reached approximately 83.46% of the hammer tip's tangential velocity. Greater rotor speed and hammer quantity intensified circulation airflow, whereas increased mass flow rate decreased it. Eddies formed within screen apertures with higher rotor speeds and hammer quantities but diminished with larger apertures and higher mass flow rates. Static pressure differences across screen apertures increased with mass flow rate and rotor speed but decreased significantly with larger apertures. This systematic examination provides insights into airflow distribution within hammer mill crushing chambers, offering a theoretical foundation for improving and designing hammer mills.

Non-invasive ventilation restores the gut microbiota in rats with acute heart failure.

Heart failure (HF) is an increasingly prevalent disease in humans; it induces multiple symptoms and damages health. The animal gut microbiota has critical roles in host health, which might be related to HF symptoms. Currently, several options are used to treat HF, including non-invasive ventilation (NIV). However, studies on gut microbiota responses to acute HF and associated treatments effects on gut communities in patients are scarce. Here, short-term (1 week after treatments) and long-term (3 months after treatment) variations in gut microbiota variations in rats with acute HF treated were examined NIV through high-throughput sequencing of the bacterial 16S rRNA gene. Through comparison of gut microbiota alpha diversity, it was observed lower gut microbiota richness and diversity in animals with acute HF than in normal animals. Additionally, beta-diversity analysis revealed significant alterations in the gut microbiota composition induced by acute HF, as reflected by increased Firmicutes/Bacteroidetes (F/B) ratios and Proteobacteria enrichment. When network analysis results were combined with the null model, decreased stability and elevated deterministic gut microbiota assemblies were observed in animals with acute HF. Importantly, in both short- and long-term periods, NIV was found to restore gut microbiota dysbiosis to normal states in acute HF rats. Finally, it was shown that considerable gut microbiota variations existed in rats with acute HF, that underlying microbiota mechanisms regulated these changes, and confirmed that NIV is suitable for HF treatment. In future studies, these findings should be validated with different model systems or clinical samples.

A data-driven Bayesian network of management and organizational factors for human reliability analysis in the process industry.

According to historical statistical data, management and organizational factors (MOFs) contribute more to process accidents than technique factors. Under the umbrella of socio-tech system theory, human reliability analysis (HRA) has become a critical part of systemic probability risk analysis. In many HRA techniques, MOFs are among the performance shaping factors (PSFs). However, the interactions and causality of MOFs to human errors are still difficult to quantify and lack validation. To fill these gaps, a framework is proposed, considering data source selection, CBN construction algorithm comparison, and results validation. The case study employed the open access eMARS database as a data source. The optimized hybrid structure learning algorithm and Bayesian criteria parameter learning algorithm are employed to build a Causal Bayesian Network (CBN) of (MOFs) that lead to human error. The proposed kernel CBN is validated through prediction accuracy and sensitivity analysis. For theoretical contribution, the validated kernel BN could generally serve as the heart part of more specific CBNs as a basis for future works. For practical applications, an application shows the model's ability to quantify the contribution of MOFs to system reliability. The results show that human-machine interacting system reliability is most sensitive to organizational factors such as adequate training and procedures.

Ovalbumin/sodium alginate Pickering emulsion: Structural characteristics and its contribution to enhancing the gel properties of Hairtail (Trichiurus haumela) surimi.

In this study, ovalbumin (OV) and sodium alginate (SA), two macromolecular complexes, were coagulated into the emulsifier (OV/SA), which stabilized soybean oil by electrostatic interaction, hydrophobic interactions, and hydrogen bonding. The structure of OV/SA and properties of OV/SA Pickering emulsion were investigated. Additionally, the effect of emulsions on the gel and protein properties of hairtail surimi was studied. The results revealed that with the increasing concentration of OV/SA, the particle size and zeta potential value (negative value) of the emulsion initially decreased and then increased, while the rheological properties gradually improved. Compared with the surimi gel directly supplemented with soybean oil, the addition of emulsion enhanced gel strength, whiteness, water holding capacity, and hydrophobic interactions, resulting in a more stable gel network structure. In summary, incorporating emulsion into surimi at the same lipid content not only maintained its gel properties but also improved its color and compensated for lipid loss.

Clotrimazole inhibits growth of multiple myeloma cells in vitro via G0/G1 arrest and mitochondrial apoptosis.

Patients with multiple myeloma (MM) experience relapse and drug resistance; therefore, novel treatments are essential. Clotrimazole (CTZ) is a wide-spectrum antifungal drug with antitumor activity. However, CTZ's effects on MM are unclear. We investigated CTZ's effect on MM cell proliferation and apoptosis induction mechanisms. CTZ's effects on MM.1S, NCI- H929, KMS-11, and U266 cell growth were investigated using Cell Counting Kit-8 (CCK-8) assay. The apoptotic cell percentage was quantified with annexin V-fluorescein isothiocyanate/7-amino actinomycin D staining. Mitochondrial membrane potential (MMP) and cell cycle progression were evaluated. Reactive oxygen species (ROS) levels were measured via fluorescence microscopy. Expression of apoptosis-related and nuclear factor (NF)-κB signaling proteins was analyzed using western blotting. The CCK-8 assay indicated that CTZ inhibited cell proliferation based on both dose and exposure time. Flow cytometry revealed that CTZ decreased apoptosis and MMP and induced G0/G1 arrest. Immunofluorescence demonstrated that CTZ dose-dependently elevated in both total and mitochondrial ROS production. Western blotting showed that CTZ enhanced Bax and cleaved poly ADP-ribose polymerase and caspase-3 while decreasing Bcl-2, p-p65, and p-IκBα. Therefore, CTZ inhibits MM cell proliferation by promoting ROS-mediated mitochondrial apoptosis, inducing G0/G1 arrest, inhibiting the NF-κB pathway, and has the potential for treating MM.

A Health Monitoring Model for Circulation Water Pumps in a Nuclear Power Plant Based on Graph Neural Network Observer.

The health monitoring of CRF (circulation water) pumps is essential for prognostics and management in nuclear power plants. However, the operational status of CRF pumps can vary due to environmental factors and human intervention, and the interrelationships between monitoring parameters are often complex. Consequently, the existing methods face challenges in effectively assessing the health status of CRF pumps. In this study, we propose a health monitoring model for CRF pumps utilizing a meta graph transformer (MGT) observer. Initially, the meta graph transformer, a temporal-spatial graph learning model, is employed to predict trends across the various monitoring parameters of the CRF pump. Subsequently, a fault observer is constructed to generate early warnings of potential faults. The proposed model was validated using real data from CRF pumps in a nuclear power plant. The results demonstrate that the average Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) of normal predictions were reduced to 1.2385, 0.5614, and 2.6554, respectively. These findings indicate that our model achieves higher prediction accuracy compared to the existing methods and can provide fault warnings at least one week in advance.

Multi role ChatGPT framework for transforming medical data analysis.

The application of ChatGPTin the medical field has sparked debate regarding its accuracy. To address this issue, we present a Multi-Role ChatGPT Framework (MRCF), designed to improve ChatGPT's performance in medical data analysis by optimizing prompt words, integrating real-world data, and implementing quality control protocols. Compared to the singular ChatGPT model, MRCF significantly outperforms traditional manual analysis in interpreting medical data, exhibiting fewer random errors, higher accuracy, and better identification of incorrect information. Notably, MRCF is over 600 times more time-efficient than conventional manual annotation methods and costs only one-tenth as much. Leveraging MRCF, we have established two user-friendly databases for efficient and straightforward drug repositioning analysis. This research not only enhances the accuracy and efficiency of ChatGPT in medical data science applications but also offers valuable insights for data analysis models across various professional domains.

Evaluation of Dynamic Changes of Volatile Organic Components for Fishmeal during Storage by HS-SPME-GC-MS with PLS-DA.

Headspace solid-phase microextraction, combined with gas chromatography-mass spectrometry and partial least squares discriminant analysis, was adopted to study the rule of change in volatile organic compounds (VOCs) for domestic and imported fishmeal during storage with different freshness grades. The results showed that 318 kinds of VOCs were detected in domestic fishmeal, while 194 VOCs were detected in imported fishmeal. The total relative content of VOCs increased with storage time, among which acids and nitrogen-containing compounds increased significantly, esters and ketones increased slightly, and phenolic and ether compounds were detected only in domestic fishmeal. Regarding the volatile base nitrogen, acid value, pH value, and mold counts as freshness indexes, the freshness indexes were significantly correlated with nine kinds of VOCs (p < 0.05) through the correlation analysis. Among them, volatile base nitrogen had a significant correlation with VOCs containing nitrogen, acid value with VOCs containing carboxyl group and hydrocarbons, pH value with acids which could be used to adjust pH value, and mold counts with part of acids adjusting pH value and VOCs containing nitrogen. Due to the fact that the value of all freshness indexes increased with freshness degradation during storage, based on volatile base nitrogen and acid value, the fishmeal was divided into three freshness grades, superior freshness, corrupting, and completely corrupted. By using partial least squares discriminant analysis, this study revealed the differences in flavor of the domestic and imported fishmeal during storage with different freshness grades, and it identified four common characteristic VOCs, namely ethoxyquinoline, 6,7,8,9-tetrahydro-3H-benzo[e]indole-1,2-dione, hexadecanoic acid, and heptadecane, produced by the fishmeal samples during storage, as well as the characteristic VOCs of fishmeal at each freshness grade.

Body mass index modified the effectiveness of low dose aspirin treatment on frozen-thawed embryo transfer outcome: a propensity score-matched study.

Background: Low-dose aspirin is one of the widely used adjuvants in assisted reproductive technologies with the hope of improving the live birth rate. However, the studies regarding its effects are conflicting. The study aimed to investigate the association between aspirin administration and live birth following frozen-thawed embryo transfer (FET) in patients with different body mass index (BMI). Methods: A retrospective cohort study was performed on 11,993 patients receiving FET treatments. 644 of which received a low-dose aspirin (100 mg/day) during endometrial preparation until 10 weeks after transfer. Propensity score matching was performed to avoid selection biases and potential confounders. Results: The clinical pregnancy rate and live birth rate were similar before matching (54.4% versus 55.4%, RR: 1.02, 95%CI: 0.95-1.09, and 46.3 versus 47.8, RR: 1.03, 95%CI: 0.95-1.12 respectively). A weak association in favor of aspirin administration was found in the matched cohort (49.5% versus 55.4%, RR: 1.12, 95%CI: 1.01-1.24, and 41.9% versus 47.8%, RR: 1.14, 95%CI: 1.01-1.29 respectively). However, when stratified the patients with WHO BMI criteria, a significant increase in live birth rate associated with aspirin treatment was found only in patients with low BMI (<18.5 kg/m2) in either unmatched (46.4% versus 59.8%, RR:1.29, 95%CI:1.07-1.55) or matched cohort (44% versus 59.8%, RR: 1.36, 95%CI: 1.01-1.83) but not in patients with higher BMI categories. With the interaction analysis, less association between aspirin and live birth appeared in patients with normal BMI (Ratio of OR:0.49, 95%CI: 0.29-0.81) and high BMI (Ratio of OR:0.57, 95%CI: 0.27-1.2) compared with patients with low BMI. Conclusion: BMI may be considered when evaluating aspirin's effect in FET cycles.

A simple protocol to establish a conditionally immortalized mouse podocyte cell line.

Podocytes are specialized terminally differentiated cells in the glomerulus that are the primary target cells in many glomerular diseases. However, the current podocyte cell lines suffer from prolonged in vitro differentiation and limited survival time, which impede research progress. Therefore, it is necessary to establish a cell line that exhibits superior performance and characteristics. We propose a simple protocol to obtain an immortalized mouse podocyte cell (MPC) line from suckling mouse kidneys. Primary podocytes were cultured in vitro and infected with the SV40 tsA58 gene to obtain immortalized MPCs. The podocytes were characterized using Western blotting and quantitative real-time PCR. Podocyte injury was examined using the Cell Counting Kit-8 assay and flow cytometry. First, we successfully isolated an MPC line and identified 39 °C as the optimal differentiation temperature. Compared to undifferentiated MPCs, the expression of WT1 and synaptopodin was upregulated in differentiated MPCs. Second, the MPCs ceased proliferating at a nonpermissive temperature after day 4, and podocyte-specific proteins were expressed normally after at least 15 passages. Finally, podocyte injury models were induced to simulate podocyte injury in vitro. In summary, we provide a simple and popularized protocol to establish a conditionally immortalized MPC, which is a powerful tool for the study of podocytes.

Vericiguat improves cardiac function and microcirculation of a male patient with Fabry disease: A case report.

Fabry disease (FD) is a rare X chromosome-linked disorder and can be easily misdiagnosed. Here, we report the case of a 69-year-old male patient with FD who developed heart failure and showed extremely high pulmonary artery pressure. His initial symptom was recurrent atrial fibrillation. The left and right atrial inner diameters were large, and the ventricular wall was thick. Gene analysis which showed GLA c.215T>C p.Met72Thr mutation and single photon emission computed tomography indicated the diagnosis of FD with coronary microvascular dysfunction. The patient was prescribed anti-heart failure drugs, including vericiguat. Following the treatment, his heart function and microvascular perfusion significantly improved, which might be due to the beneficial effects of vericiguat.

Prediction of anti-cancer drug synergy based on cross-matching network and cancer molecular subtypes.

At present, anti-cancer drug synergy therapy is one of the most important methods to overcome drug resistance and reduce drug toxicity in cancer treatment. High-throughput screening through deep learning can effectively improve the efficiency of discovering synergistic drugs. Nowadays, most of the existing deep learning algorithms for anti-cancer drug synergy prediction use deep neural networks and can only implicitly perform feature interaction. This study proposes a deep learning algorithm, named MolCross, which combines implicit feature interaction with explicit features to improve the accuracy of prediction of the anti-cancer drug synergy score. MolCross uses a deep autoencoder to extract features from high-dimensional input, uses the drug-specific subnetworks and cross-network to perform implicit feature interaction and explicit feature interaction respectively, and finally uses a synergy prediction network to combine the two feature interaction methods to obtain the final prediction results. We adopted a five-fold cross validation and compared MolCross with other four anti-cancer drug synergy prediction models. The results show that MolCross has better prediction performance than other models. MolCross also has good performance in terms of cross-cell line and cross-tissue type. Existing studies have demonstrated that cancer molecular subtypes have different sensitivities to targeted therapy. In this study, the features of cancer molecular subtype were introduced in the model using an embedding layer in MolCross to explore the effect of cancer molecular subtype on anti-cancer drug synergy. We also found that the cancer molecular subtype is one of the main factors affecting the synergy between drugs.

Indications affect neonatal outcomes following early rescue ICSI: a retrospective study.

PURPOSE: To investigate the impact of heterogeneity in patient indications or insemination protocols on neonatal outcomes of singletons following early rescue ICSI (rICSI) treatments. METHODS: A retrospective study was conducted. Propensity score matching and multivariable logistic regression were used to adjust for confounders and biases. RESULTS: A total of 9095 IVF patients, 2063 ICSI patients, and 642 early rICSI patients were included in the study. No differences were detected in neonatal outcomes except small for gestational age (SGA) which increased in early rICSI patients compared with both unmatched and matched IVF groups with the risk ratio (RR) of 1.31 (95% CI: 1.05, 1.64) and 1.49 (95% CI: 1.05, 2.12). Further analysis showed that SGA increased significantly in partial fertilization failure (PFF) cycles with RRs of 1.56 (95% CI: 1.08, 2.27) and 1.78 (95% CI: 1.22, 2.59) compared with both unmatched and matched IVF patients but not in TFF patients. A positive association between fertilization rate via IVF and birth weight z-score was revealed in the PFF patients. CONCLUSION: Early rICSI in patients with total fertilization failure (TFF) appeared to be safe in terms of neonatal outcomes. However, when expanding the indications of rICSI to PFF patients, the SGA in the offspring increased, suggesting a potential effect on long-term health. Since other treatment options, such as using only the IVF-origin embryos still exist for these patients, further studies were needed to confirm the optimal decision for these patients.

Paramyosin from field snail (Bellamya quadrata): Structural characteristics and its contribution to enhanced the gel properties of myofibrillar protein.

To assess the blending effect of field snails with grass carp muscle, the effects of paramyosin (PM) and actomyosin (AM) with different mixture ratios on the gel properties of the binary blend system were investigated in our work. The purified PM from field snail muscle was about 95 kDa on SDS-PAGE. Its main secondary structure was α-helix, which reached to 97.97 %. When the amount of PM increased in the binary blend system, their rheological indices and gel strength were improved. The water holding capacity (WHC) increased to 86.30 % at a mixture ratio of 2:8. However, the WHC and the area of immobile water (P22) dramatically decreased, and the area of free water (P23) increased when the mixture ratio exceeded 4:6. The low level of PM in binary blend system promoted the formation of a homogenous and dense gel network through non-covalent interactions as observed results of SEM and FTIR. When there were redundant PM molecules, the development of heterostructure via hydrophobic interaction of tail-tail contributed to the reduced gel properties of the binary blend system. These findings provided new insight into the binary blend system of PM and AM with different ratios to change the gel properties of myofibrillar protein.

An effective framework for predicting drug-drug interactions based on molecular substructures and knowledge graph neural network.

Drug-drug interactions (DDIs) play a central role in drug research, as the simultaneous administration of multiple drugs can have harmful or beneficial effects. Harmful interactions lead to adverse reactions, some of which can be life-threatening, while beneficial interactions can promote efficacy. Therefore, it is crucial for physicians, patients, and the research community to identify potential DDIs. Although many AI-based techniques have been proposed for predicting DDIs, most existing computational models primarily focus on integrating multiple data sources or combining popular embedding methods. Researchers often overlook the valuable information within the molecular structure of drugs or only consider the structural information of drugs, neglecting the relationship or topological information between drugs and other biological objects. In this study, we propose MSKG-DDI - a two-component framework that incorporates the Drug Chemical Structure Graph-based component and the Drug Knowledge Graph-based component to capture multimodal characteristics of drugs. Subsequently, a multimodal fusion neural layer is utilized to explore the complementarity between multimodal representations of drugs. Extensive experiments were conducted using two real-world datasets, and the results demonstrate that MSKG-DDI outperforms other state-of-the-art models in binary-class, multi-class, and multi-label prediction tasks under both transductive and inductive settings. Furthermore, the ablation analysis further confirms the practical usefulness of MSKG-DDI.