Negative logarithm of the acid dissociation constant (pK _a) significantly influences the absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties of molecules and is a crucial indicator in drug research. Given the rapid and accurate characteristics of computational methods, their role in predicting drug properties is increasingly important. Although many pK _a prediction models currently exist, they often focus on enhancing model precision while neglecting interpretability. In this study, we present GraFpK _a, a pK _a prediction model using graph neural networks (GNNs) and molecular fingerprints. The results show that our acidic and basic models achieved mean absolute errors (MAEs) of 0.621 and 0.402, respectively, on the test set, demonstrating good predictive performance. Notably, to improve interpretability, GraFpK _a also incorporates Integrated Gradients (IGs), providing a clearer visual description of the atoms significantly affecting the pK _a values. The high reliability and interpretability of GraFpK _a ensure accurate pK _a predictions while also facilitating a deeper understanding of the relationship between molecular structure and pK _a values, making it a valuable tool in the field of pK _a prediction.

Accurate prediction of drug-target interactions (DTIs) plays a pivotal role in drug discovery, facilitating optimization of lead compounds, drug repurposing and elucidation of drug side effects. However, traditional DTI prediction methods are often limited by incomplete biological data and insufficient representation of protein features. In this study, we proposed KG-CNNDTI, a novel knowledge graph-enhanced framework for DTI prediction, which integrates heterogeneous biological information to improve model generalizability and predictive performance. The proposed model utilized protein embeddings derived from a biomedical knowledge graph via the Node2Vec algorithm, which were further enriched with contextualized sequence representations obtained from ProteinBERT. For compound representation, multiple molecular fingerprint schemes alongside the Uni-Mol pre-trained model were evaluated. The fused representations served as inputs to both classical machine learning models and a convolutional neural network-based predictor. Experimental evaluations across benchmark datasets demonstrated that KG-CNNDTI achieved superior performance compared to state-of-the-art methods, particularly in terms of Precision, Recall, F1-Score and area under the precision-recall curve (AUPR). Ablation analysis highlighted the substantial contribution of knowledge graph-derived features. Moreover, KG-CNNDTI was employed for virtual screening of natural products against Alzheimer's disease, resulting in 40 candidate compounds. 5 were supported by literature evidence, among which 3 were further validated in vitro assays.
Alzheimer Disease/drug therapy* ; Biological Products/therapeutic use* ; Humans ; Neural Networks, Computer ; Machine Learning ; Drug Discovery/methods* ; Algorithms ; Drug Evaluation, Preclinical/methods*

OBJECTIVE: To investigate the early effectiveness of transiliac-transsacral screws internal fixation assisted by augmented reality navigation system HoloSight (hereinafter referred to as "computer navigation system") in the treatment of posterior pelvic ring injuries. METHODS: A retrospective analysis was made in the 41 patients with posterior pelvic ring injuries who had been treated surgically with transiliac-transsacral screws between June 2022 and June 2023. The patients were divided into navigation group (18 cases, using computer navigation system to assist screw implantation) and freehand group (23 cases, using C-arm X-ray fluoroscopy to guide screw implantation) according to the different methods of transiliac-transsacral screws placement. There was no significant difference in gender, age, body mass index, causes of injuries, Tile classification of pelvic fracture, days from injury to operation, usage of unlocking closed reduction technique between the two groups ( P>0.05). The time of screw implantation, the fluoroscopy times, the guide wire adjustment times of each screw, and the incidence of complications were recorded and compared between the two groups. The position of the transiliac-transsacral screw was scanned by CT within 2 days after operation, and the position of the screw was classified according to Gras standard. RESULTS: The operation was successfully completed in both groups. The time of screw implantation, the fluoroscopy times, and the guide wire adjustment times of each screw in the navigation group were significantly less than those in the freehand group ( P<0.05). There were 2 cases of incision infection in the freehand group, and the incision healed by first intention after active dressing change; there was no screw-related complication in the navigation group during operation and early period after operation; the difference in incidence of complications between the two groups (8.7% vs. 0) was not significant ( P=0.495). According to the Gras standard, the screw position of the navigation group was significantly better than that of the freehand group ( P<0.05). CONCLUSION Compared with the traditional freehand method, the computer navigation system assisted transiliac-transsacral screws internal fixation in the treatment of posterior pelvic ring injuries has advantages of improving the accuracy of screw implantation and reducing radiation damage and the time of screw implantation.
Humans ; Retrospective Studies ; Surgical Wound Infection ; Replantation ; Body Mass Index

In order to investigate the growth of in vitro cultured microencapsulated adre nal cortical tissues of rats, we produced the microcapsules by using sodium alginate, calcium chloride and poly-L-lysine, etc. The adrenal glands were taken out of 24 rats, half of the adrenal glands were microencap sulated (capsule group) while the other half were left free (noncapsule group). After 36 hours' in vitro culture, the concentrations of aldosterone and cortisol in culture solution were measured by radioimmunoassay. The two groups con tinued to be cultured for 24 and 36 hours with ACTH stimulation. The concentrations of aldosterone and cortisol at 24 and 36 hours were detected respectively. Parts of adrenal cortical tissues were observed under light microscope and electron micros cope. The results showed that the concentrations of aldoste rone and cortisol in the microencapsulated adrenal cortical tissues were signifi cantly higher than those of noncapsule group (P＜0.01). In the capsule group , the concentration of cortisol after 36 hours' ACTH stimulation was significantly higher than that before ACTH stimulation (P＜0.01). After 24 and 36 hours' ACTH stimulation, the concentrations of aldosterone and cortisol were sign if icantly higher in the capsule group than in the noncapsule group. In the nonca psule group, regardless of ACTH stimulation, no significant change was seen in the concentration of aldosterone and cortisol (P＞0.05). The cellular structure of adrenal tissues remained intact under light microscopic and electron microscopic observations, and the cells remained alive well. These data suggest that in vitro cultured cells of microenc ap sulated adrenal cortical tissues remain alive well. Microcapsule has not influen ced the secretory function of adrenal tissues. Microencapsulated tissues have a good response to ACTH stimulation.