A social theory-enhanced graph representation learning framework for multitask prediction of drug-drug interactions.

Current machine learning-based methods have achieved inspiring predictions in the scenarios of mono-type and multi-type drug-drug interactions (DDIs), but they all ignore enhancive and depressive pharmacological changes triggered by DDIs. In addition, these pharmacological changes are asymmetric since the roles of two drugs in an interaction are different. More importantly, these pharmacological changes imply significant topological patterns among DDIs. To address the above issues, we first leverage Balance theory and Status theory in social networks to reveal the topological patterns among directed pharmacological DDIs, which are modeled as a signed and directed network. Then, we design a novel graph representation learning model named SGRL-DDI (social theory-enhanced graph representation learning for DDI) to realize the multitask prediction of DDIs. SGRL-DDI model can capture the task-joint information by integrating relation graph convolutional networks with Balance and Status patterns. Moreover, we utilize task-specific deep neural networks to perform two tasks, including the prediction of enhancive/depressive DDIs and the prediction of directed DDIs. Based on DDI entries collected from DrugBank, the superiority of our model is demonstrated by the comparison with other state-of-the-art methods. Furthermore, the ablation study verifies that Balance and Status patterns help characterize directed pharmacological DDIs, and that the joint of two tasks provides better DDI representations than individual tasks. Last, we demonstrate the practical effectiveness of our model by a version-dependent test, where 88.47 and 81.38% DDI out of newly added entries provided by the latest release of DrugBank are validated in two predicting tasks respectively.

Directed graph attention networks for predicting asymmetric drug-drug interactions.

It is tough to detect unexpected drug-drug interactions (DDIs) in poly-drug treatments because of high costs and clinical limitations. Computational approaches, such as deep learning-based approaches, are promising to screen potential DDIs among numerous drug pairs. Nevertheless, existing approaches neglect the asymmetric roles of two drugs in interaction. Such an asymmetry is crucial to poly-drug treatments since it determines drug priority in co-prescription. This paper designs a directed graph attention network (DGAT-DDI) to predict asymmetric DDIs. First, its encoder learns the embeddings of the source role, the target role and the self-roles of a drug. The source role embedding represents how a drug influences other drugs in DDIs. In contrast, the target role embedding represents how it is influenced by others. The self-role embedding encodes its chemical structure in a role-specific manner. Besides, two role-specific items, aggressiveness and impressionability, capture how the number of interaction partners of a drug affects its interaction tendency. Furthermore, the predictor of DGAT-DDI discriminates direction-specific interactions by the combination between two proximities and the above two role-specific items. The proximities measure the similarity between source/target embeddings and self-role embeddings. In the designated experiments, the comparison with state-of-the-art deep learning models demonstrates the superiority of DGAT-DDI across a direction-specific predicting task and a direction-blinded predicting task. An ablation study reveals how well each component of DGAT-DDI contributes to its ability. Moreover, a case study of finding novel DDIs confirms its practical ability, where 7 out of the top 10 candidates are validated in DrugBank.

Laser stripe extraction method in industrial environments utilizing self-adaptive convolution technique.

A line-structured laser scanner is widely applied for 3D reconstruction in industrial environments with ubiquitous various luminance, complicated background, diverse objects, and instable lasers. These elements will show up as noise in the obtained laser stripe images. Therefore, the basic and key point for a line-structured laser scanner is to accurately extract the laser stripe from noise. This paper proposes an effective laser stripe extraction procedure with two steps. First, a novel laser stripe center extraction method based on the geometry information and correlation in the laser stripe, is designed to significantly eliminate noise and accurately extract the laser stripe centers. In addition, the robustness, speed, and accuracy of this method are respectively analyzed in detail. Second, piecewise fitting is adopted to acquire a smooth and continuous laser stripe centerline. In order to select the optimal fitting method, the characteristics of two spline methods, Akima spline and cubic Hermite spline, are deeply analyzed and compared. Finally, an experiment is carried out by using a rough metal step and a line-structured laser scanning system. The experiment results demonstrate that the proposed self-adaptive convolution-mass method can significantly eliminate noise in industrial environments. In addition, the cubic Hermite spline is a better choice for 3D reconstruction, rather than the Akima spline.

Enhanced adsorption of phosphorus in soil by lanthanum-modified biochar: improving phosphorus retention and storage capacity.

Use of soil adsorbent is an effective method for the promotion of phosphorus adsorption capacity of soil, though most of the soil adsorbents have weak phosphorus retention ability. Herein, we compared the traditional gypsum (GP) and zeolite (ZP) adsorbents to explore the phosphorus retention ability of lanthanum modified walnut shell biochar (La-BC) in soil. The results showed that with the increase of exogenous phosphorus concentration, the adsorption amount of phosphorus by adsorbents in soil increased at first and then tended to be stable. The maximum adsorption capacity of soil to phosphorus is gypsum, lanthanum-modified biochar > zeolite, and the addition of lanthanum-modified biochar can improve the adsorption capacity of soil to phosphorus, enhance the binding strength of soil and phosphorus, improve the ability of soil to store phosphorus, reducing phosphorus adsorption saturation, and is beneficial to control the leaching of soil phosphorus. FTIR and XRD analysis showed that the adsorption of phosphorus by each adsorbent in soil was mainly chemical precipitation. The response surface analysis showed that the adsorption performance of La-BC+S was the best when the concentration of exogenous phosphorus was 50.0 mg/L, pH was 6.47, and the reaction time was 436.98 min. This study provides a reference for soil adsorbents to hold phosphorus and reduce the risk of phosphorus leaching to avoid groundwater pollution.

Reticulocalbin-1 knockdown increases the sensitivity of cells to Adriamycin in nasopharyngeal carcinoma and promotes endoplasmic reticulum stress-induced cell apoptosis.

Nasopharyngeal carcinoma (NPC) mainly appears in southeastern Asian countries, including China. Adriamycin (ADM), a type of antitumor drug, is widely applied in treatments against various cancers. Nevertheless, cancer cells will eventually develop drug resistance to ADM. The present study aims to explore the potential role of reticulocalbin-1 (RCN1) in NPC cells resistance to ADM. Microarray-based analysis was used to screen NPC-related genes, with RCN1 acquired for this current study. RCN1 expression in NPC tissues and cells was determined. The biological function of RCN1 on NPC cell apoptosis was evaluated via gain- and loss-of-function experiments in 5-8 F/ADM and 5-8 F cells by delivering si-RCN1 and RCN1-vector. The function of endoplasmic reticulum (ER) stress on cell apoptosis was measured with the involvement of the PERK-CHOP signaling pathway. Furthermore, tumor formation in nude mice was performed to evaluate the survival condition and RCN1 effects in vivo. RCN1 was highly expressed in NPC tissues and cell lines. The increased expression of ER-related proteins ATF4, CHOP, and the extents of IRE1 and PERK phosphorylation were observed. RCN1 knockdown was found to reduce resistance of NPC cells/tissues to ADM while activating ER stress through the activated PERK-CHOP signaling pathway, which further promoted NPC cell apoptosis. These in vitro findings were detected in vivo on tumor formation in nude mice. In conclusion, the present study provides evidence that RCN1 knockdown stimulates ADM sensitivity in NPC by promoting ER stress-induced cell apoptosis, highlighting a theoretical basis for NPC treatment.