Correction: Halogen-bonded charge-transfer co-crystal scintillators for high-resolution X-ray imaging.

[This corrects the article DOI: 10.1039/D4SC00735B.].

Halogen-bonded charge-transfer co-crystal scintillators for high-resolution X-ray imaging.

The development of high-quality organic scintillators encounters challenges primarily associated with the weak X-ray absorption ability resulting from the presence of low atomic number elements. An effective strategy involves the incorporation of halogen-containing molecules into the system through co-crystal engineering. Herein, we synthesized a highly fluorescent dye, 2,5-di(4-pyridyl)thiazolo[5,4-d]thiazole (Py2TTz), with a fluorescence quantum yield of 12.09%. Subsequently, Py2TTz was co-crystallized with 1,4-diiodotetrafluorobenzene (I2F4B) and 1,3,5-trifluoro-2,4,6-triiodobenzene (I3F3B) obtaining Py2TTz-I2F4 and Py2TTz-I3F3. Among them, Py2TTz-I2F4 exhibited exceptional scintillation properties, including an ultrafast decay time (1.426 ns), a significant radiation luminescence intensity (146% higher than Bi3Ge4O12), and a low detection limit (70.49 nGy s-1), equivalent to 1/78th of the detection limit for medical applications (5.5 µGy s-1). This outstanding scintillation performance can be attributed to the formation of halogen-bonding between I2F4B and Py2TTz. Theoretical calculations and single-crystal structures demonstrate the formation of halogen-bond-induced rather than π-π-induced charge-transfer cocrystals, which not only enhances the X-ray absorption ability and material conductivity under X-ray exposure, but also constrains molecular vibration and rotation, and thereby reducing non-radiative transition rate and sharply increasing its fluorescence quantum yields. Based on this, the flexible X-ray film prepared based on Py2TTz-I2F4 achieved an ultrahigh spatial resolution of 26.8 lp per mm, underscoring the superiority of this strategy in developing high-performance organic scintillators.

Combining data augmentation and deep learning for improved epilepsy detection.

Introduction: In recent years, the use of EEG signals for seizure detection has gained widespread academic attention. Aiming at the problem of overfitting deep learning models due to the small number of EEG signal data during epilepsy detection, this paper proposes an epilepsy detection method that combines data augmentation and deep learning. Methods: First, the Adversarial and Mixup Data Augmentation (AMDA) method is used to realize the data augmentation, which effectively enriches the number of training samples. To further improve the classification accuracy and robustness of epilepsy detection, this paper proposes a one-dimensional convolutional neural network and gated recurrent unit (AM-1D CNN-GRU) network model based on attention mechanism for epilepsy detection. Results and discussion: The experimental results show that the performance of epilepsy detection achieved by using augmented data is significantly improved, and the accuracy, sensitivity, and area under the subject's working characteristic curve are up to 96.06, 95.48%, and 0.9637, respectively. Compared with the non-augmented data, all indicators are increased by more than 6.2%. Meanwhile, the detection performance was significantly improved compared with other epilepsy detection methods. The results of this research can provide a reference for the clinical application of epilepsy detection.

The impact of blood lactic acid levels on retinopathy of prematurity morbidity.

BACKGROUND: Retinopathy of prematurity (ROP) is a common disease in premature infants. In recent years, most researchers have used lactic acid as poor prognosis marker in premature infants. This study aims to explore investigate the impact of blood lactic acid levels on ROP. METHODS: A retrospective case-control study was conducted, and infants with severe ROP born with birth weight (BW) ≤ 1500 g and gestational age (GA) ≤ 32 weeks were enrolled from November 2016 to November 2021. Infants without any stage ROP were included as controls and were matched with ROP infants (1:2) by GA and BW. All selected preterm infants were tested for heel terminal trace blood gas analysis within two weeks of life. Changes in blood lactic acid levels in the two groups were compared and analyzed by using multivariate logistic regression analysis. Sensitivity and specificity were analyzed by receiver operating characteristic (ROC) curve. RESULTS: There were 79 infants in ROP group, and 158 infants in control group. The levels of blood lactic acid were significantly higher in the ROP group on days 1, 3, 5, and 7 compared with control group (all p < 0.05). The blood lactic acid levels on day 5 was an independent risk factor for ROP (p = 0.017). The area under the curve (AUC), sensitivity and specificity were highest on day 5 (AUC 0.716, sensitivity 77.2% and specificity 62.0%, respectively, p < 0.001), and higher on days 1, 3, and 7. CONCLUSION: A high blood lactic acid level in the first seven days of life may be associated with increases ROP occurrence in very preterm infants, and suggest blood lactic acid level may impact the occurrence of ROP.

Mechanochemical Synthesis of Cuprous Complexes for X-ray Scintillation and Imaging.

Cuprous complex scintillators show promise for X-ray detection with abundant raw materials, diverse luminescent mechanisms, and adjustable structures. However, their synthesis typically requires a significant amount of organic solvents, which conflict with green chemistry principles. Herein, we present the synthesis of two high-performance cuprous complex scintillators using a simple mechanochemical method for the first time, namely [CuI(PPh3)2R] (R = 4-phenylpyridine hydroiodide (PH, Cu-1) and 4-(4-bromophenyl)pyridine hydroiodide (PH-Br, Cu-2). Both materials demonstrated remarkable scintillation performances, exhibiting radioluminescence (RL) intensities 1.52 times (Cu-1) and 2.52 times (Cu-2) greater than those of Bi4Ge3O12 (BGO), respectively. Compared to Cu-1, the enhanced RL performance of Cu-2 can be ascribed to its elevated quantum yield of 51.54%, significantly surpassing that of Cu-1 at 37.75%. This excellent luminescent performance is derived from the introduction of PH-Br, providing a more diverse array of intermolecular interactions that effectively constrain molecular vibration and rotation, further suppressing the nonradiative transition process. Furthermore, Cu-2 powder can be prepared into scintillator film with excellent X-ray imaging capabilities. This work establishes a pathway for the rapid, eco-friendly, and cost-effective synthesis of high-performance cuprous complex scintillators.

GRELinker: A Graph-Based Generative Model for Molecular Linker Design with Reinforcement and Curriculum Learning.

Fragment-based drug discovery (FBDD) is widely used in drug design. One useful strategy in FBDD is designing linkers for linking fragments to optimize their molecular properties. In the current study, we present a novel generative fragment linking model, GRELinker, which utilizes a gated-graph neural network combined with reinforcement and curriculum learning to generate molecules with desirable attributes. The model has been shown to be efficient in multiple tasks, including controlling log P, optimizing synthesizability or predicted bioactivity of compounds, and generating molecules with high 3D similarity but low 2D similarity to the lead compound. Specifically, our model outperforms the previously reported reinforcement learning (RL) built-in method DRlinker on these benchmark tasks. Moreover, GRELinker has been successfully used in an actual FBDD case to generate optimized molecules with enhanced affinities by employing the docking score as the scoring function in RL. Besides, the implementation of curriculum learning in our framework enables the generation of structurally complex linkers more efficiently. These results demonstrate the benefits and feasibility of GRELinker in linker design for molecular optimization and drug discovery.

Enhancing Generic Reaction Yield Prediction through Reaction Condition-Based Contrastive Learning.

Deep learning (DL)-driven efficient synthesis planning may profoundly transform the paradigm for designing novel pharmaceuticals and materials. However, the progress of many DL-assisted synthesis planning (DASP) algorithms has suffered from the lack of reliable automated pathway evaluation tools. As a critical metric for evaluating chemical reactions, accurate prediction of reaction yields helps improve the practicality of DASP algorithms in the real-world scenarios. Currently, accurately predicting yields of interesting reactions still faces numerous challenges, mainly including the absence of high-quality generic reaction yield datasets and robust generic yield predictors. To compensate for the limitations of high-throughput yield datasets, we curated a generic reaction yield dataset containing 12 reaction categories and rich reaction condition information. Subsequently, by utilizing 2 pretraining tasks based on chemical reaction masked language modeling and contrastive learning, we proposed a powerful bidirectional encoder representations from transformers (BERT)-based reaction yield predictor named Egret. It achieved comparable or even superior performance to the best previous models on 4 benchmark datasets and established state-of-the-art performance on the newly curated dataset. We found that reaction-condition-based contrastive learning enhances the model's sensitivity to reaction conditions, and Egret is capable of capturing subtle differences between reactions involving identical reactants and products but different reaction conditions. Furthermore, we proposed a new scoring function that incorporated Egret into the evaluation of multistep synthesis routes. Test results showed that yield-incorporated scoring facilitated the prioritization of literature-supported high-yield reaction pathways for target molecules. In addition, through meta-learning strategy, we further improved the reliability of the model's prediction for reaction types with limited data and lower data quality. Our results suggest that Egret holds the potential to become an essential component of the next-generation DASP tools.

Halogen-bonding boosting the high performance X-ray imaging of organic scintillators.

Organic scintillators with efficient X-ray excited luminescence are essential for medical diagnostics and security screening. However, achieving excellent organic scintillation materials is challenging due to low X-ray absorption coefficients and inferior radioluminescence (RL) intensity. Herein, supramolecular interactions are incorporated, particularly halogen bonding, into organic scintillators to enhance their radioluminescence properties. By introducing heavy atoms (X = Cl, Br, I) into 9,10-bis(4-pyridyl)anthracene (BPA), the formation of halogen bonding (BPA-X) enhances their X-ray absorption coefficient and restricts the molecular vibration and rotation, which boosts their RL intensity. The RL intensity of BPA-Cl and BPA-Br fluorochromes increased by over 2 and 6.3 times compared to BPA, respectively. Especially, BPA-Br exhibits an ultrafast decay time of 8.25 ns and low detection limits of 25.95 ± 2.49 nGy s-1. The flexible film constructed with BPA-Br exhibited excellent X-ray imaging capabilities. Furthermore, this approach is also applicable to organic phosphors. The formation of halogen bonding in bromophenyl-methylpyridinium iodide (PYI) led to a fourfold increase in RL intensity compared to bromophenyl-methyl-pyridinium (PY). It suggests that halogen bonding serves as a promising and effective molecular design strategy for the development of high-performance organic scintillator materials, presenting new opportunities for their applications in radiology and security screening.

Role of the autonomic nervous system in young, middle-aged, and older individuals with essential hypertension and sleep-related changes in neurocardiac regulation.

Essential hypertension involves complex cardiovascular regulation. The autonomic nervous system function fluctuates throughout the sleep-wake cycle and changes with advancing age. However, the precise role of the autonomic nervous system in the development of hypertension during aging remains unclear. In this study, we characterized autonomic function during the sleep-wake cycle in different age groups with essential hypertension. This study included 97 men (53 with and 44 without hypertension) aged 30-79 years. They were stratified by age into young (< 40 years), middle-aged (40-59 years), and older (60-79 years) groups. Polysomnography and blood pressure data were recorded for 2 min before and during an hour-long nap. Autonomic function was assessed by measuring heart rate variability and blood pressure variability. Data were analyzed using t tests, correlation analyses, and two-way analysis of variance. During nonrapid eye movement (nREM), a main effect of age was observed on cardiac parasympathetic measures and baroreflex sensitivity (BRS), with the highest and lowest levels noted in the younger and older groups, respectively. The coefficients of the correlations between these measures and age were lower in patients with hypertension than in normotensive controls. The BRS of young patients with hypertension was similar to that of their middle-aged and older counterparts. However, cardiac sympathetic activity was significantly higher (p = 0.023) and BRS was significantly lower (p = 0.022) in the hypertension group than in the control group. During wakefulness, the results were similar although some of the above findings were absent. Autonomic imbalance, particularly impaired baroreflex, plays a more significant role in younger patients with hypertension. The nREM stage may be suitable for gaining insights into the relevant mechanisms.

Tree-Invent: A Novel Multipurpose Molecular Generative Model Constrained with a Topological Tree.

De novo molecular design plays an important role in drug discovery. Here, a novel generative model, Tree-Invent, was proposed to integrate topological constraints in the generation of a molecular graph. In this model, a molecular graph is represented as a topological tree in which a ring system, a nonring atom, and a chemical bond are regarded as the ring node, single node, and edge, respectively. The molecule generation is driven by three independent submodels for carrying out operations of node addition, ring generation, and node connection. One unique feature of the generative model is that the topological tree structure can be specified as a constraint for structure generation, which provides more precise control of structure generation. Combined with reinforcement learning, the Tree-Invent model could efficiently explore targeted chemical space. Moreover, the Tree-Invent model is flexible enough to be used in versatile molecule design settings such as scaffold decoration, scaffold hopping, and linker generation.

Employing feature engineering strategies to improve the performance of machine learning algorithms on echocardiogram dataset.

Objectives: This study mainly uses machine learning (ML) to make predictions by inputting features during training and inference. The method of feature selection is an important factor affecting the accuracy of ML models, and the process includes data extraction, which is the collection of all data required for ML. It also needs to import the concept of feature engineering, namely, this study needs to label the raw data of the cardiac ultrasound dataset with one or more meaningful and informative labels so that the ML model can learn from it and predict more accurate target values. Therefore, this study will enhance the strategies of feature selection methods from the raw dataset, as well as the issue of data scrubbing. Methods: In this study, the ultrasound dataset was cleaned and critical features were selected through data standardization, normalization, and missing features imputation in the field of feature engineering. The aim of data scrubbing was to retain and select critical features of the echocardiogram dataset while making the prediction of the ML algorithm more accurate. Results: This paper mainly utilizes commonly used methods in feature engineering and finally selects four important feature values. With the ML algorithms available on the Azure platform, namely, Random Forest and CatBoost, a Voting Ensemble method is used as the training algorithm, and this study also uses visual tools to gain a clearer understanding of the raw data and to improve the accuracy of the predictive model. Conclusion: This paper emphasizes feature engineering, specifically on the cleaning and analysis of missing values in the raw dataset of echocardiography and the identification of important critical features in the raw dataset. The Azure platform is used to predict patients with a history of heart disease (individuals who have been under surveillance in the past three years and those who haven't). Through data scrubbing and preprocessing methods in feature engineering, the model can more accurately predict the future occurrence of heart disease in patients.

Generic Interpretable Reaction Condition Predictions with Open Reaction Condition Datasets and Unsupervised Learning of Reaction Center.

Effective synthesis planning powered by deep learning (DL) can significantly accelerate the discovery of new drugs and materials. However, most DL-assisted synthesis planning methods offer either none or very limited capability to recommend suitable reaction conditions (RCs) for their reaction predictions. Currently, the prediction of RCs with a DL framework is hindered by several factors, including: (a) lack of a standardized dataset for benchmarking, (b) lack of a general prediction model with powerful representation, and (c) lack of interpretability. To address these issues, we first created 2 standardized RC datasets covering a broad range of reaction classes and then proposed a powerful and interpretable Transformer-based RC predictor named Parrot. Through careful design of the model architecture, pretraining method, and training strategy, Parrot improved the overall top-3 prediction accuracy on catalysis, solvents, and other reagents by as much as 13.44%, compared to the best previous model on a newly curated dataset. Additionally, the mean absolute error of the predicted temperatures was reduced by about 4 °C. Furthermore, Parrot manifests strong generalization capacity with superior cross-chemical-space prediction accuracy. Attention analysis indicates that Parrot effectively captures crucial chemical information and exhibits a high level of interpretability in the prediction of RCs. The proposed model Parrot exemplifies how modern neural network architecture when appropriately pretrained can be versatile in making reliable, generalizable, and interpretable recommendation for RCs even when the underlying training dataset may still be limited in diversity.

Nonvolatile Memory Organic Light-Emitting Transistors.

In the field of active-matrix organic light emitting display (AMOLED), large-size and ultra-high-definition AMOLED applications have escalated the demand for the integration density of driver chips. However, as Moore's Law approaches the limit, the traditional technology of improving integration density that relies on scaling down device dimension is facing a huge challenge. Thus, developing a multifunctional and highly integrated device is a promising route for improving the integration density of pixel circuits. Here, a novel nonvolatile memory ferroelectric organic light-emitting transistor (Fe-OLET) device which integrates the switching capability, light-emitting capability and nonvolatile memory function into a single device is reported. The nonvolatile memory function of Fe-OLET is achieved through the remnant polarization property of ferroelectric polymer, enabling the device to maintain light emission at zero gate bias. The reliable nonvolatile memory operations are also demonstrated. The proof-of-concept device optimized through interfacial modification approach exhibits 20 times improved field-effect mobility and five times increased luminance. The integration of nonvolatile memory, switching and light-emitting capabilities within Fe-OLET provides a promising internal-storage-driving paradigm, thus creating a new pathway for deploying storage capacitor-free circuitry to improve the pixel aperture ratio and the integration density of circuits toward the on-chip advanced display applications.

Branched montbretin A mimics allow derivatisation and potent amylase inhibition.

Mimics of the complex flavonol glycoside montbretin A in which a flavonol moiety is coupled to a caffeic acid via partially peptidic linkers have proved to be potent inhibitors of human pancreatic alpha-amylase with potential as therapeutics for control of blood glucose levels. After exploring optimal linker length, a synthetic route to a version with a branched linker was devised based on the structure of the enzyme/inhibitor complex. The resultant branched inhibitors were shown to retain nanomolar potency even when decorated with polymers as a means of modifying solubility. Similar improvements, along with nanomolar affinity, could also be achieved through conjugation to cyclodextrins which have the potential to bind to starch binding sites found on the surface of human amylase. Incorporation of a conjugatable branch into this unusual pharmacophore thereby affords considerable flexibility for further modifications to improve pharmacokinetic behaviour or as a site for attachment of capture tags or fluorophores.

3D based generative PROTAC linker design with reinforcement learning.

Proteolysis targeting chimera (PROTAC), has emerged as an effective modality to selectively degrade disease-related proteins by harnessing the ubiquitin-proteasome system. Due to PROTACs' hetero-bifunctional characteristics, in which a linker joins a warhead binding to a protein of interest (POI), conferring specificity and a E3-ligand binding to an E3 ubiquitin ligase, this could trigger the ubiquitination and transportation of POI to the proteasome, followed by degradation. The rational PROTAC linker design is challenging due to its relatively large molecular weight and the complexity of maintaining the binding mode of warhead and E3-ligand in the binding pockets of counterpart. Conventional linker generation method can only generate linkers in either 1D SMILES or 2D graph, without taking into account the information of ternary structures. Here we propose a novel 3D linker generative model PROTAC-INVENT which can not only generate SMILES of PROTAC but also its 3D putative binding conformation coupled with the target protein and the E3 ligase. The model is trained jointly with the RL approach to bias the generation of PROTAC structures toward pre-defined 2D and 3D based properties. Examples were provided to demonstrate the utility of the model for generating reasonable 3D conformation of PROTACs. On the other hand, our results show that the associated workflow for 3D PROTAC conformation generation can also be used as an efficient docking protocol for PROTACs.

An extensive benchmark study on biomedical text generation and mining with ChatGPT.

MOTIVATION: In recent years, the development of natural language process (NLP) technologies and deep learning hardware has led to significant improvement in large language models (LLMs). The ChatGPT, the state-of-the-art LLM built on GPT-3.5 and GPT-4, shows excellent capabilities in general language understanding and reasoning. Researchers also tested the GPTs on a variety of NLP-related tasks and benchmarks and got excellent results. With exciting performance on daily chat, researchers began to explore the capacity of ChatGPT on expertise that requires professional education for human and we are interested in the biomedical domain. RESULTS: To evaluate the performance of ChatGPT on biomedical-related tasks, this article presents a comprehensive benchmark study on the use of ChatGPT for biomedical corpus, including article abstracts, clinical trials description, biomedical questions, and so on. Typical NLP tasks like named entity recognization, relation extraction, sentence similarity, question and answering, and document classification are included. Overall, ChatGPT got a BLURB score of 58.50 while the state-of-the-art model had a score of 84.30. Through a series of experiments, we demonstrated the effectiveness and versatility of ChatGPT in biomedical text understanding, reasoning and generation, and the limitation of ChatGPT build on GPT-3.5. AVAILABILITY AND IMPLEMENTATION: All the datasets are available from BLURB benchmark https://microsoft.github.io/BLURB/index.html. The prompts are described in the article.

A model for predicting physical function upon discharge of hospitalized older adults in Taiwan-a machine learning approach based on both electronic health records and comprehensive geriatric assessment.

Background: Predicting physical function upon discharge among hospitalized older adults is important. This study has aimed to develop a prediction model of physical function upon discharge through use of a machine learning algorithm using electronic health records (EHRs) and comprehensive geriatrics assessments (CGAs) among hospitalized older adults in Taiwan. Methods: Data was retrieved from the clinical database of a tertiary medical center in central Taiwan. Older adults admitted to the acute geriatric unit during the period from January 2012 to December 2018 were included for analysis, while those with missing data were excluded. From data of the EHRs and CGAs, a total of 52 clinical features were input for model building. We used 3 different machine learning algorithms, XGBoost, random forest and logistic regression. Results: In total, 1,755 older adults were included in final analysis, with a mean age of 80.68 years. For linear models on physical function upon discharge, the accuracy of prediction was 87% for XGBoost, 85% for random forest, and 32% for logistic regression. For classification models on physical function upon discharge, the accuracy for random forest, logistic regression and XGBoost were 94, 92 and 92%, respectively. The auROC reached 98% for XGBoost and random forest, while logistic regression had an auROC of 97%. The top 3 features of importance were activity of daily living (ADL) at baseline, ADL during admission, and mini nutritional status (MNA) during admission. Conclusion: The results showed that physical function upon discharge among hospitalized older adults can be predicted accurately during admission through use of a machine learning model with data taken from EHRs and CGAs.

A deep learning framework for accurate reaction prediction and its application on high-throughput experimentation data.

In recent years, it has been seen that artificial intelligence (AI) starts to bring revolutionary changes to chemical synthesis. However, the lack of suitable ways of representing chemical reactions and the scarceness of reaction data has limited the wider application of AI to reaction prediction. Here, we introduce a novel reaction representation, GraphRXN, for reaction prediction. It utilizes a universal graph-based neural network framework to encode chemical reactions by directly taking two-dimension reaction structures as inputs. The GraphRXN model was evaluated by three publically available chemical reaction datasets and gave on-par or superior results compared with other baseline models. To further evaluate the effectiveness of GraphRXN, wet-lab experiments were carried out for the purpose of generating reaction data. GraphRXN model was then built on high-throughput experimentation data and a decent accuracy (R2 of 0.712) was obtained on our in-house data. This highlights that the GraphRXN model can be deployed in an integrated workflow which combines robotics and AI technologies for forward reaction prediction.

Novel Semiconductive Ternary Hybrid Heterostructures for Artificial Optoelectronic Synapses.

Synaptic devices that mimic biological synapses are considered as promising candidates for brain-inspired devices, offering the functionalities in neuromorphic computing. However, modulation of emerging optoelectronic synaptic devices has rarely been reported. Herein, a semiconductive ternary hybrid heterostructure is prepared with a D-D'-A configuration by introducing polyoxometalate (POM) as an additional electroactive donor (D') into a metalloviologen-based D-A framework. The obtained material features an unprecedented porous 8-connected bcu-net that accommodates nanoscale [α-SiW12 O40 ]4- counterions, displaying uncommon optoelectronic responses. Besides, the fabricated synaptic device based on this material can achieve dual-modulation of synaptic plasticity due to the synergetic effect of electron reservoir POM and photoinduced electron transfer. And it can successfully simulate learning and memory processes similar to those in biological systems. The result provides a facile and effective strategy to customize multi-modality artificial synapses in the field of crystal engineering, which opens a new direction for developing high-performance neuromorphic devices.

Application of message passing neural networks for molecular property prediction.

Accurate molecular property prediction, as one of the classical cheminformatics topics, plays a prominent role in the fields of computer-aided drug design. For instance, property prediction models can be used to quickly screen large molecular libraries to find lead compounds. Message-passing neural networks (MPNNs), a sub-class of Graph neural networks (GNNs), have recently been demonstrated to outperform other deep learning methods on a variety of tasks, including the prediction of molecular characteristics. In this survey, we provide a brief review of the MPNN models and their applications on molecular property prediction.