Specific Structure Research Articles

Dissecting the pleiotropic roles of reactive oxygen species (ROS) in lung cancer: From carcinogenesis towardtherapy.

Lung cancer is a major cause of morbidity and mortality. The specific pulmonary structure to directly connect with ambient air makes it more susceptible to damage from airborne toxins. External oxidative stimuli and endogenous reactive oxygen species (ROS) play a crucial role in promoting lung carcinogenesis and development. The biological properties of higher ROS levels in tumor cells than in normal cells make them more sensitive and vulnerable to ROS injury. Therefore, the strategy of targeting ROS has been proposed for cancer therapy for decades. However, it is embarrassing that countless attempts at ROS-based therapies have had very limited success, and no FDA approval in the anticancer list was mechanistically based on ROS manipulation. Even compared with the untargetable proteins, such as transcription factors, ROS are more difficult to be targeted due to their chemical properties. Thus, the pleiotropic roles of ROS provide therapeutic potential for anticancer drug discovery, while a better dissection of the mechanistic action and signaling pathways is a prerequisite for future breakthroughs. This review discusses the critical roles of ROS in cancer carcinogenesis, ROS-inspired signaling pathways, and ROS-based treatment, exemplified by lung cancer. In particular, an eight considerations rule is proposed for ROS-targeting strategies and drug design and development.

Diagnosis-Guided Deep Subspace Clustering Association Study for Pathogenetic Markers Identification of Alzheimer's Disease Based on Comparative Atlases.

The roles of brain region activities and genotypic functions in the pathogenesis of Alzheimer's disease (AD) remain unclear. Meanwhile, current imaging genetics methods are difficult to identify potential pathogenetic markers by correlation analysis between brain network and genetic variation. To discover disease-related brain connectome from the specific brain structure and the fine-grained level, based on the Automated Anatomical Labeling (AAL) and human Brainnetome atlases, the functional brain network is first constructed for each subject. Specifically, the upper triangle elements of the functional connectivity matrix are extracted as connectivity features. The clustering coefficient and the average weighted node degree are developed to assess the significance of every brain area. Since the constructed brain network and genetic data are characterized by non-linearity, high-dimensionality, and few subjects, the deep subspace clustering algorithm is proposed to reconstruct the original data. Our multilayer neural network helps capture the non-linear manifolds, and subspace clustering learns pairwise affinities between samples. Moreover, most approaches in neuroimaging genetics are unsupervised learning, neglecting the diagnostic information related to diseases. We presented a label constraint with diagnostic status to instruct the imaging genetics correlation analysis. To this end, a diagnosis-guided deep subspace clustering association (DDSCA) method is developed to discover brain connectome and risk genetic factors by integrating genotypes with functional network phenotypes. Extensive experiments prove that DDSCA achieves superior performance to most association methods and effectively selects disease-relevant genetic markers and brain connectome at the coarse-grained and fine-grained levels.

The association between childhood trauma and motor structure in obsessive-compulsive rituals: An ethological and evolutionary approach.

Ethological models have highlighted a specific motor structure of compulsions in obsessive-compulsive disorder (OCD), based on the rigid repetitions of acts, and with the adaptive significance of facing unpredictable conditions. Such an evolutionary mechanism might explain the robust association between childhood traumatic experiences (CTEs) and OCD. However, a relationship between CTEs and the motor structure of compulsions has not been investigated yet. The first objective of the study was to confirm a specific motor structure of OCD compulsions with respect to control behaviors; the second objective was to assess a possible association between the motor structure of compulsions and CTEs severity. Thirty-two OCD outpatients (13 female, Mage = 44.50 years, SE = 19.71) and 27 healthy controls (10 female, Mage = 37.62 years, SE = 16.20), matched for sex and age, provided a videotape of their compulsions and corresponding ordinary acts, respectively. Behavior was scored with the software "Observer." Participants were administered the Yale-Brown Obsessive Compulsive Scale and the Childhood Trauma Questionnaire. A dependent t test was used to compare the motor structure of behavior between the groups; Pearson's correlations to investigate associations between motor parameters and CTEs. Compulsions showed a specific motor structure due to the repetition of functional and nonfunctional acts. CTEs severity was especially associated with the repetition of functional acts, independently from OCD severity. Our findings, in confirming a peculiar motor structure for OCD compulsions, hint for the first time at a link between CTEs and compulsive repetition of functional acts, which would represent a plastic developmental response to CTEs unpredictability. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Phenomenologies of loneliness: alone without you, alone without an us

Loneliness is not simply characterised by a lack of other people in one’s environment or lifeworld, but by a lack of specific forms of relationship, connection or belonging. In this paper I connect the conversation about different types of loneliness with debates happening in the philosophy of intersubjectivity, specifically the phenomenology of intersubjectivity. Philosophers of intersubjectivity tend to characterise experiences of direct encounter as having a second-personal ‘I-thou’ quality. By contrast, experiences of belonging tend to be characterised by first-person plural ‘we’ structures. In this paper I will compare and contrast some of the different phenomenological qualities that come with the absence or poverty of these two types of intersubjective structure respectively. These different types of loneliness are characterised by different phenomenological properties, but also by (the absence of) specific interpersonal and social structures. I argue that the absence of these different interpersonal structures respectively gives rise to loneliness as the experience of being unseen and loneliness as the experience of not-being-at-home. This is significant for loneliness studies, as understanding and combating loneliness is consequently likely to take variegated rather than homogeneous forms. This is particularly significant for a psychosocial approach to loneliness which seeks to understand, hold together and integrate both the structural features of loneliness (in this case the relevant interpersonal and social structures) and the lived experience of individuals and groups (in this case the phenomenology of loneliness).

Characterization of defects in additively manufactured materials from mechanical properties

Minimization of defects in additively manufactured components is essential for material development and the determination of fabrication process windows. Here, a purely mechanical means of quantifying porosity in a material is developed based on the Mori–Tanaka mean field theory. The micromechanics based defect quantification approach developed here links the elastic properties of a material with the specific defect structures that must exist within it to obtain those properties. Laser powder bed fusion is used to fabricate aluminum 6061 with three different processing conditions to produce nominally identical samples with unique defect microstructures. Test samples are imaged with microcomputed X-ray tomography (μCT) and the images are classified with a machine learning method to identify the volume fraction of pores, solidification cracks, and the orientation distribution function of the solidification cracks. Tensile tests are then conducted in the z-axis and within the xy-plane to measure the respective elastic moduli and the Poisson’s ratio. These three commonly measured material properties are the only inputs to the new micromechanics based approach used to infer the same defect descriptors that were extracted from μCT. Excellent agreement is found between defect descriptors obtained with both methods, and quantities derived from the mechanical properties are found to carry substantially lower uncertainty. Full error propagation maps for the micromechanics based technique are calculated. Notably, defect descriptors from the micromechanics based approach are calculated from strain and load measurements, which are measured on much larger length scales than the defect size. This enables simple quantification of defects for materials containing small defects that are challenging to image.

고등학교 영어 교과서 속 칭찬과 칭찬응답

Objectives The purpose of the paper was to analyze two speech acts (compliment and compliment response) in high school English textbooks and to find whether they were presented authentically and accurately according to L2 pragmatic norms.
 Methods Compliments in 8 differing textbooks were analyzed in terms of three aspects of compliment (syntactic structures, lexicals and topics), and compliment responses in textbooks were categorized according to four types of response (acceptance, amendment, rejection and non-acknowledgement). Then the analyzed compliments and compliment responses were compared with those acts made by English native speakers.
 Results Compliments in textbooks heavily relied on specific syntactic structures, used a variety of adjectives rather than a limited set of adjectives, and focused more on personality than physical appearance as compliment topics, which showed clear differences from what English native speakers do in their performance of compliments. Also unlike native speakers who prefer acceptance by responding positively to compliments, compliment responses in textbooks over-used non-acceptance strategies through refusal or non-acknowledgement.
 Conclusions Given these results, English textbooks should be developed based only on pragmatic norms of target language speakers, not intuitions or L1 knowledge of textbook developers.

Confined Transport Behavior of Biomolecules within Tilted Nanopores.

The transport behavior of biomolecules at the confined nanoscale is very different from that of the bulk state. Numerous disease diagnostics and targeted drug treatments are performed based on nanochannels in cells. The specific structure and shape of nanochannels play an important role in the behavior and efficiency of substance transport. In this paper, we fabricated nanopores with different tilt angles and the same diameters using focused ion beam. The capture frequency and the blocking current amplitude of λ-DNA within large-angle nanopores decrease obviously, suggesting an increase in the energy barrier of large-angle nanopores and the fact that they stretch biomolecules to thinness. Most importantly, large-angle nanopores slow down λ-DNA transport by 2-4 times. MD simulations find that the sloped electroosmotic flow inside the tilted nanopores is the main factor contributing to the transport phenomena. The increase in the capture time of biomolecules by nanopores assists in obtaining more biological information from the current trajectories. Our study provides a new understanding of substance transport in specially shaped nanopores, which can be instrumental in providing fresh inspiration and approaches to the biomedical field.

Hierarchical porous poly (L-lactic acid) fibrous vascular graft with controllable architectures and stable structure

Electrospun fibre has shown great potential in tissue engineering and regenerative medicine due to its high specific surface area and extracellular matrix-mimicking structure. However, fabricating an electrospun fibrous scaffold with controllable complex 3D macroscopic configuration remains a challenge. In the present study, a novel method was designed to transform 2D electrospun poly (L-lactic acid) (PLLA) fibrous membrane to tubular PLLA fibrous scaffolds with 3D complex but tailored configuration. The electrospun PLLA fibrous membrane was rolled around a designed mould and then treated with acetone. Treated vascular grafts’ length, diameter, and shape can be tailored by the mould parameters. Moreover, treated vascular grafts achieve favourable mechanical properties (Young’s modulus = 155 MPa, tensile stress = 8.79 MPa and radial force = 2.2 N) and the mechanical properties could be engineered on demand. In addition, treated vascular grafts kept their initial structure and size during long-term in vitro experiments once they were formed. In addition, with the acetone-induced recrystallization of PLLA, pristine solid PLLA fibres were changed to hierarchical porous PLLA fibres with ultra-high specific surface area (28.9 m2/g) and wettability (water contact angle = 101.32°), which has positive effects on cell adhesion and proliferation ability. A7r5 in vitro experiment shows that the proliferation rate of treated vascular grafts increased 153% at day 4 and 170.6% at day 7 compared with pristine vascular grafts.

3D Printing Bio‐Inspired Micro Soft Robot with Programming Magnetic Elastic Composites

AbstractBio‐inspired micro soft robots mimic biologically specific body structures and movement mechanisms by utilizing bionic principles. Because of its soft body, it can adapt to complex external environments. However, the existing polymer material system is single and the fabrication process is limited. How to further reduce soft robot size has become a key issue. In this work, a Programming Magnetic Elastic Composites (PMEC) is proposed for 3D printing to manufacture micro soft robots. The PMEC has the advantage of changing the direction of the internal magnetic field in response to changes in the external magnetic field. A Microsoft robot is designed and fabricated using PMEC inspired by an inchworm. Numerical simulations are also used to study the effect of soft robot size parameters on local stresses and optimize the structure. The results show that the microscale soft robot can crawl with a load of 2 times its weight at a speed of 6.67 mm s−1, crawl on slopes from 0° to 90°, and crawl over obstacles with a maximum height of 7 mm. In addition, active adaptation of soft robots to complex tunnel models based on external stimuli is achieved by magnetic field control.

Recent Progress in Nucleic Acid Pulmonary Delivery toward Overcoming Physiological Barriers and Improving Transfection Efficiency.

Pulmonary delivery of therapeutic agents has been considered the desirable administration route for local lung disease treatment. As the latest generation of therapeutic agents, nucleic acid has been gradually developed as gene therapy for local diseases such as asthma, chronic obstructive pulmonary diseases, and lung fibrosis. The features of nucleic acid, specific physiological structure, and pathophysiological barriers of the respiratory tract have strongly affected the delivery efficiency and pulmonary bioavailability of nucleic acid, directly related to the treatment outcomes. The development of pharmaceutics and material science provides the potential for highly effective pulmonary medicine delivery. In this review, the key factors and barriers are first introduced that affect the pulmonary delivery and bioavailability of nucleic acids. The advanced inhaled materials for nucleic acid delivery are further summarized. The recent progress of platform designs for improving the pulmonary delivery efficiency of nucleic acids and their therapeutic outcomes have been systematically analyzed, with the application and the perspectives of advanced vectors for pulmonary gene delivery.

Unifying Structured Data as Graph for Data-to-Text Pre-Training

Abstract Data-to-text (D2T) generation aims to transform structured data into natural language text. Data-to-text pre-training has proved to be powerful in enhancing D2T generation and yields impressive performance. However, previous pre-training methods either oversimplified structured data into a sequence without considering input structures or designed training objectives tailored for a specific data structure (e.g., table or knowledge graph). In this paper, we unify different types of structured data (i.e., table, key-value data, knowledge graph) into the graph format and cast different D2T generation tasks as graph-to-text generation. To effectively exploit the structural information of the input graph, we propose a structure-enhanced pre-training method for D2T generation by designing a structure-enhanced Transformer. Concretely, we devise a position matrix for the Transformer, encoding relative positional information of connected nodes in the input graph. In addition, we propose a new attention matrix to incorporate graph structures into the original Transformer by taking the available explicit connectivity structure into account. Extensive experiments on six benchmark datasets show the effectiveness of our model. Our source codes are available at https://github.com/AlibabaResearch/DAMO-ConvAI/tree/main/unid2t.

Engineered biosorbents of pomelo (Citrus maxima (Burm.f.) Merr) peels modified with zinc oxide and titanium dioxide for methylene blue dye sorption

The pomelo-doped zinc oxide beads (PZB), pomelo-doped titanium dioxide beads (PTB), and pomelo-doped zinc oxide and titanium dioxide beads (PZTB) were synthesized for sorbing methylene blue (MB) dye. Their characterizations were explored by X-Ray Diffractometer (XRD), Field Emission Scanning Electron Microscopy and Focus Ion Beam (FESEM-FIB), Energy Dispersive X-Ray Spectrometer (EDX), and Fourier Transform Infrared Spectroscopy (FT-IR). In addition, their sorbent efficiencies for sorbing MB dye were investigated through batch experiments, sorbent reusability studies, sorption isotherms, kinetics, and thermodynamic studies. They were crystalline phases presenting the specific peaks of zinc oxide (ZnO) or titanium dioxide (TiO2). Their surfaces had lamella structures with coarse surfaces, and they also found specific structures of ZnO or TiO2 on the surfaces. Zn–O or Ti–O–Ti was also detected in PZB or PTB or, PZTB depending upon metal oxide types added into pomelo beaded sorbents. For batch experiments, they could adsorb MB dye of more than 86%, and PZTB showed the highest MB dye removal efficiency. In addition, they could be reused for more than three cycles with high MB dye sorptions of more than 72%. They corresponded to Freundlich and pseudo-second-order kinetic models. Moreover, the increasing temperature affected their decreasing MB dye sorptions which were exothermic processes.

NEURAL NETWORK-BASED ADVERSE DRUG REACTION PREDICTION USING MOLECULAR SUBSTRUCTURE ANALYSES

Objective: This study aims to enhance early detection and prediction by exploiting drug molecular substructures, overcoming challenges posed by limited authentic patient data in the medical domain. Methods: The study implemented a neural network approach to optimize molecular fingerprint algorithms and employed various machine learning algorithms for predictions. Additionally, the study identified and extracted substructures associated with severe Adverse Drug Reactions (ADRs), validating their presence within drug structures through a comparison with a random set of drug structures. Predictions were made for specific molecular structures, and results were validated using clinical evidence from the literature. Results: Optimized molecular fingerprint algorithms and diverse machine-learning models yielded promising outcomes. The Area Under Curve (AUC) value for the fingerprint dataset was obtained at approximately 65%, and integrating it with patient data significantly improved the performance by about 30%. Substructure analysis pinpointed key components linked to severe ADRs, reinforcing the predictive prowess of the model. Predictions for specific molecular structures were corroborated using clinical evidence from the literature, fortifying the credibility of the proposed approach. Conclusion: In conclusion, this research effectively tackles challenges in the early detection and prediction of ADRs by leveraging machine learning algorithms, focusing on drug molecular substructures. The optimized model, incorporating both fingerprint and patient datasets, demonstrated significant improvements in predictive performance. Identifying and validating substructures linked to severe ADRs contribute to the model's reliability. The study's findings are vital for advancing drug safety and laying the groundwork for further strides in predictive modeling within the medical domain.

The Genetic Architecture of the Root System during Seedling Emergence in Populus euphratica under Salt Stress and Control Environments

The growth of the Populus euphratica root system is of great significance for its survival under adverse environmental stress. In harsh saline-stress environments, the proportion, morphology, and functionality of the taproots and lateral roots and how they manifest specific adaptive structures, growth strategies, and potential genetic controls are still subjects for further exploration. In this study, we delve into the fundamental patterns and trade-offs of root morphology and functionality by constructing an environment-induced differential interaction equation (EDIE) to model the independent and interactive growth of the root system while considering the influence of environmental conditions. We identify 93 key QTLs in the control group and 44 key QTLs in the salt-stress group, of which 2 QTLs are significant in both environments. By constructing ODE-based QTL networks, we explore in depth how these loci regulate the growth of the root system under different environmental conditions while considering their independent direct effects and epistatic effects among loci. This study elucidates the intrinsic factors that influence the variations in taproots and lateral roots, providing crucial insights into the relationship between root morphology and functionality.

Construction of SU-102 for adsorption and photocatalytic synergistic removal of tetracycline.

Tetracycline (TC) is a significant group of broad-spectrum antibiotics that are frequently employed in medical health and animal husbandry. However, the problem of TC residues has been increasing globally with the large-scale production and widespread use, posing a serious threat to the human health and ecological environment. In this paper, a green plant-based MOF SU-102 was prepared, and the adsorption characteristics of SU-102 on TC were investigated. SU-102 was columnar crystal with considerable specific surface area and pore structure, and it could adsorb TC quickly and effectively. And compared to SU-102-a, the adsorption rate of TC by SU-102-b has increased by nearly four times. The adsorption reaction was a spontaneous, entropy-gaining, heat-absorbing process. The adsorption mechanisms between SU-102 and TC were π-π interaction and hydrogen bonding. In addition, SU-102 also had considerable photocatalytic properties, and its application in adsorbent desorption treatment effectively solved the problem of secondary pollution.

Effects of different thawing methods on physical and physicochemical properties of frozen dough and quality of corresponding steamed bread

The thawing method is critical for the final quality of products based on the frozen dough. The effects of ultrasound thawing, proofer thawing, refrigerator thawing, water bath thawing, ambient thawing, and microwave thawing on the rheology, texture, water distribution, fermentation characteristics, and microstructure of frozen dough and the properties of steamed bread were investigated. The results indicated that the ultrasound thawing dough had better physicochemical properties than other doughs. It was found that ultrasound thawing restrained the water migration of dough, improved its rheological properties and fermentation capacity. The total gas volume value of the ultrasound thawing dough was reduced by 21.35% compared with that of unfrozen dough. The ultrasound thawing dough displayed a thoroughly uniform starch-gluten network, and an enhanced the specific volume and internal structure of the steamed bread. In conclusion, ultrasound thawing effectively mitigated the degradation of the frozen dough and enhanced the quality of steamed bread.

G4-Ligand-Conjugated Oligonucleotides Mediate Selective Binding and Stabilization of Individual G4 DNA Structures.

G-quadruplex (G4) DNA structures are prevalent secondary DNA structures implicated in fundamental cellular functions, such as replication and transcription. Furthermore, G4 structures are directly correlated to human diseases such as cancer and have been highlighted as promising therapeutic targets for their ability to regulate disease-causing genes, e.g., oncogenes. Small molecules that bind and stabilize these structures are thus valuable from a therapeutic perspective and helpful in studying the biological functions of the G4 structures. However, there are hundreds of thousands of G4 DNA motifs in the human genome, and a long-standing problem in the field is how to achieve specificity among these different G4 structures. Here, we developed a strategy to selectively target an individual G4 DNA structure. The strategy is based on a ligand that binds and stabilizes G4s without selectivity, conjugated to a guide oligonucleotide, that specifically directs the G4-Ligand-conjugated oligo (GL-O) to the single target G4 structure. By employing various biophysical and biochemical techniques, we show that the developed method enables the targeting of a unique, specific G4 structure without impacting other off-target G4 formations. Considering the vast amount of G4s in the human genome, this represents a promising strategy to study the presence and functions of individual G4s but may also hold potential as a future therapeutic modality.

Segmentation of Tree-Like Data Structures and Their Parallel Processing in Authentication Methods Based on Block Coupling Encoding

Purpose of reseach. In the tasks of authenticating groups of messages encoded in the mode of chaining blocks, there is a need for the formation and processing of specific tree-like structures. The contents of such structures, in addition to information about the placement of data in the internal memory of the calculators, describes the relative location of messages in the data stream between subscribers of a peer-to-peer network. This information is necessary to isolate a structured set from the entire message stream to the receiver, for which its source is uniquely determined. Using approaches to segmentation of tree structures allows you to parallelize the processes of adding elements to it and searching for areas corresponding to an authentication error.Methods. The division of the tree structure into areas subject to modification and areas for analysis is based on a metric dynamically formed from message authentication codes – the position of a specific message in a structured set of messages transmitted from the source to the receiver. The value of this metric determines the distance from the root of the tree, which defines the boundary between the two named areasResults. By isolating the modified and analyzed sections of the tree structure, races of processes implementing independent algorithms for working with it are excluded. The possibility of detecting authentication errors before receiving the last message in a structured set of messages is shown. As a result, there is no need to transmit those group messages that were supposed to be sent after the error was detected. Formulas for estimating the average transmission time of multiple messages with sequential and parallel implementation of procedures for the formation and processing of a tree structure containing descriptors of incoming messages to the receiver are given.Conclusion. The paper shows that the parallel implementation of algorithms for adding elements to the tree structure and the algorithm for searching for areas corresponding to an error reduces the average transmission time of a group of messages by 5-12% compared with the sequential implementation of these algorithms. This reduces the load on the communication channel for the target class of systems using block coupling encoding for authentication.

Porous poly(bismaleimide-co-divinylbenzene) microspheres as dispersive solid-phase extraction adsorbent coupled to high-performance liquid chromatography for the determination of triazine herbicide residues in vegetable samples.

In this work, monodisperse and nano-porous poly(bismaleimide-co-divinylbenzene) microspheres with large specific surface area (427.6 m2 /g) and rich pore structure were prepared by one-pot self-stable precipitation polymerization of 2,2'-bis[4-(4-maleimidophenoxy) phenyl] propane and divinylbenzene. The prepared poly(bismaleimide-co-divinylbenzene) microspheres were employed as dispersive solid-phase extraction (DSPE) adsorbent for the extraction of triazine herbicides. Under optimized conditions, good linearities were obtained between the peak area and the concentration of triazine herbicides in the range of 1-400µg/L (R2 ≥ 0.9987) with the limits of detection of 0.12-0.31µg/L. Triazine herbicides were detected using the described approach in vegetable samples (i.e., cucumber, tomato, and maize) with recoveries of 93.6%-117.3% and relative standard deviations of 0.4%-3.5%. In addition, the recoveries of triazine herbicides remained above 80.7% after being used for nine DSPE cycles, showing excellent reusability of poly(bismaleimide-co-divinylbenzene) microspheres. The adsorption of poly(bismaleimide-co-divinylbenzene) microspheres toward triazine herbicides was a monolayer and chemical adsorption. The adsorption mechanism between triazine herbicides and adsorbents might be a combination of hydrogen bonding, electrostatic interaction, and π-π conjugation. The results confirmed the potential use of the poly(bismaleimide-co-divinylbenzene) microspheres-based DSPE coupled to the high-performance liquid chromatography method for the detection of triazine herbicide residues in vegetable samples.

AutoDDI: Drug-Drug Interaction Prediction With Automated Graph Neural Network.

Drug-drug interaction (DDI) has attracted widespread attention because when incompatible drugs are taken together, DDI will lead to adverse effects on the body, such as drug poisoning or reduced drug efficacy. The adverse effects of DDI are closely determined by the molecular structures of the drugs involved. To represent drug data effectively, researchers usually treat the molecular structure of drugs as a molecule graph. Then, previous studies can use the handcrafted graph neural network (GNN) model to learn the molecular graph representations of drugs for DDI prediction. However, in the field of bioinformatics, manually designing GNN architectures for specific molecular structure datasets is time-consuming and depends on expert experience. To address this problem, we propose an automatic drug-drug interaction prediction method named AutoDDI that can efficiently and automatically design the GNN architecture for drug-drug interaction prediction without manual intervention. To this end, we first design an effective search space for drug-drug interaction prediction by revisiting various handcrafted GNN architectures. Then, to efficiently and automatically design the optimal GNN architecture for each drug dataset from the search space, a reinforcement learning search algorithm is adopted. The experiment results show that AutoDDI can achieve the best performance on two real-world datasets. Moreover, the visual interpretation results of the case study show that AutoDDI can effectively capture drug substructure for drug-drug interaction prediction.