Single-cell transcriptomics dissects the transcriptome alterations of hematopoietic stem cells in myelodysplastic neoplasms.

BACKGROUND: Myelodysplastic neoplasms (MDS) are myeloid neoplasms characterized by disordered differentiation of hematopoietic stem cells and a predisposition to acute myeloid leukemia (AML). The underline pathogenesis remains unclear. METHODS: In this study, the trajectory of differentiation and mechanisms of leukemic transformation were explored through bioinformatics analysis of single-cell RNA-Seq data from hematopoietic stem and progenitor cells (HSPCs) in MDS patients. RESULTS: Among the HSPC clusters, the proportion of common myeloid progenitor (CMP) was the main cell cluster in the patients with excess blasts (EB)/ secondary AML. Cell cycle analysis indicated the CMP of MDS patients were in an active proliferative state. The genes involved in the cell proliferation, such as MAML3 and PLCB1, were up-regulated in MDS CMP. Further validation analysis indicated that the expression levels of MAML3 and PLCB1 in patients with MDS-EB were significantly higher than those without EB. Patients with high expression of PLCB1 had a higher risk of transformation to AML. PLCB1 inhibitor can suppress proliferation, induce cell cycle arrest, and activate apoptosis of leukemic cells in vitro. CONCLUSION: This study revealed the transcriptomic change of HSPCs in MDS patients along the pseudotime and indicated that PLCB1 plays a key role in the transformation of MDS into leukemia.

Progress of lymphocyte activation gene 3 and programmed cell death protein 1 antibodies for cancer treatment: A review.

The application of immune checkpoint inhibitors has proven to be an effective treatment for cancer. Immune checkpoints such as programmed cell death protein 1/programmed cell death protein 1 ligand 1 (PD-1/PD-L1), cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), T-cell immunoglobulin-3 (TIM-3), T-cell immunoglobulin and ITIM domain (TIGIT), and lymphocyte activation gene-3 (LAG-3) have received extensive attention, and the efficacy of antibodies or inhibitors against these checkpoints (either alone or in combination) has been evaluated in many tumors. This paper provides a brief overview of the PD-1 and LAG-3 checkpoints, and then shifts focus to the combined use of PD-1 and LAG-3 antibodies in both in vivo and in vitro experiments. In the in vitro experiments, we examined the correlation between the expression and activation of these inhibitors on T cells, and also assessed toxicity in animals in preparation for in vivo experiments. The effects of the combined use of PD-1 and LAG-3 antibodies were then summarized in animal models of melanoma, MC38 carcinoma, and other tumors. In clinical studies, the combined application of these antibodies was assessed in patients with melanoma, colorectal, breast, and renal cell cancers, as well as other solid tumors. In general, the combination of PD-1 and LAG-3 antibodies has shown promising results in both in vivo and in vitro studies.

Heterointerface regulation of covalent organic framework-anchored graphene via a solvent-free strategy for high-performance supercapacitor and hybrid capacitive deionization electrodes.

Covalent organic frameworks (COFs) with customizable geometry and redox centers are an ideal candidate for supercapacitors and hybrid capacitive deionization (HCDI). However, their poor intrinsic conductivity and micropore-dominated pore structures severely impair their electrochemical performance, and the synthesis process using organic solvents brings serious environmental and cost issues. Herein, a 2D redox-active pyrazine-based COF (BAHC-COF) was anchored on the surface of graphene in a solvent-free strategy for heterointerface regulation. The as-prepared BAHC-COF/graphene (BAHCGO) nanohybrid materials possess high-speed charge transport offered by the graphene carrier and accelerated electrolyte ion migration within the BAHC-COF, allowing ions to effectively occupy ion storage sites inside BAHC. As a result, the BAHCGO//activated carbon asymmetric supercapacitor achieves a high energy output of 61.2 W h kg-1 and a satisfactory long-term cycling life. More importantly, BAHCGO-based HCDI possesses a high salt adsorption capacity (SAC) of 67.5 mg g-1 and excellent long-term desalination/regeneration stability. This work accelerates the application of COF-based materials in the fields of energy storage and water treatment.

Exosomes promote pre-metastatic niche formation in colorectal cancer.

It is well known that colorectal cancer (CRC) has a high morbidity rate, a poor prognosis when metastasized, and a greatly shortened 5-year survival rate. Therefore, understanding the mechanism of tumor metastasis is still important. Based on the "seed and soil" theory, the concept of " premetastatic niche (PMN)" was introduced by Kaplan et al. The complex interaction between primary tumors and the metastatic organ provides a beneficial microenvironment for tumor cells to colonize at a distance. With further exploration of the PMN, exosomes have gradually attracted interest from researchers. Exosomes are extracellular vesicles secreted from cells that include various biological information and are involved in communication between cells. As a key molecule in the PMN, exosomes are closely related to tumor metastasis. In this article, we obtained information by conducting a comprehensive search across academic databases including PubMed and Web of Science using relevant keywords. Only recent, peer-reviewed articles published in the English language were considered for inclusion. This study aims to explore in depth how exosomes promote the formation of pre-metastatic microenvironment (PMN) in colorectal cancer and its related mechanisms.

Mitochondrial-targeting and NIR-responsive Mn3O4@PDA@Pd-SS31 nanozymes reduce oxidative stress and reverse mitochondrial dysfunction to alleviate osteoarthritis.

Mitochondrial reactive oxygen species (mROS) play a crucial role in the process of osteoarthritis (OA), which may be a promising target for therapy of OA. In this study, novel mitochondrial-targeting and SOD-mimic Mn3O4@PDA@Pd-SS31 nanozymes with near-infrared (NIR) responsiveness and synergistic cascade to scavenge mROS were designed for the therapy of OA. Results showed that the nanozymes accelerated the release of Pd and Mn3O4 under NIR irradiation, exhibiting enhanced activities of SOD and CAT mimic enzymes with reversed mitochondrial dysfunction and promoted mitophagy to effectively scavenge mROS from chondrocytes, modulate the microenvironment of oxidative stress, and eventually inhibit the inflammatory response. Nanozymes were excreted in vivo through intestinal metabolic pathway and had good biocompatibility, effectively reducing the inflammatory response and relieving articular cartilage degeneration in OA joints, with a reduction of 93.7 % and 93.8 % in OARSCI scores for 4 and 8 weeks respectively. Thus, this study demonstrated that the mitochondria targeting and NIR responsive Mn3O4@PDA@Pd-SS31 nanozymes could efficiently scavenge mROS, repair damaged mitochondrial function and promote cartilage regeneration, which are promising for the treatment of OA in clinical applications.

GPNMB promotes peripheral nerve regeneration by activating the Erk1/2 and Akt pathways via binding Na+/K+-ATPase α1 in Schwann cells.

Glycoprotein non-metastatic melanoma protein B (GPNMB) is ubiquitously expressed and has protective effects on the central nervous system. In particular, it is also expressed in the peripheral nervous system (PNS) and upregulated after peripheral nerve injury. However, the role and underlying mechanism of GPNMB in the PNS, especially in peripheral nerve regeneration (PNR), are still unknown and need to be further investigated. In this study, recombinant human GPNMB (rhGPNMB) was injected into a sciatic nerve injury model. It was found that rhGPNMB facilitated the regeneration and functional recovery of the injured sciatic nerve in vivo. Moreover, it was also confirmed that GPNMB activated the Erk1/2 and Akt pathways via binding with Na+/K + -ATPase α1 (NKA α1) and promoted the proliferation and migration of Schwann cells (SCs) and their expression and secretion of neurotrophic factors and neural adhesion molecules in vitro. Our findings demonstrate that GPNMB facilitates PNR through activation of the Erk1/2 and Akt pathways in SCs by binding with NKA α1 and may be a novel strategy for PNR.

Endothelial extracellular vesicles induce acute lung injury via follistatin-like protein 1.

Cardiopulmonary bypass has been speculated to elicit systemic inflammation to initiate acute lung injury (ALI), including acute respiratory distress syndrome (ARDS), in patients after cardiac surgery. We previously found that post-operative patients showed an increase in endothelial cell-derived extracellular vesicles (eEVs) with components of coagulation and acute inflammatory responses. However, the mechanism underlying the onset of ALI owing to the release of eEVs after cardiopulmonary bypass, remains unclear. Plasma plasminogen-activated inhibitor-1 (PAI-1) and eEV levels were measured in patients with cardiopulmonary bypass. Endothelial cells and mice (C57BL/6, Toll-like receptor 4 knockout (TLR4-/-) and inducible nitric oxide synthase knockout (iNOS-/-)) were challenged with eEVs isolated from PAI-1-stimulated endothelial cells. Plasma PAI-1 and eEVs were remarkably enhanced after cardiopulmonary bypass. Plasma PAI-1 elevation was positively correlated with the increase in eEVs. The increase in plasma PAI-1 and eEV levels was associated with post-operative ARDS. The eEVs derived from PAI-1-stimulated endothelial cells could recognize TLR4 to stimulate a downstream signaling cascade identified as the Janus kinase 2/3 (JAK2/3)-signal transducer and activator of transcription 3 (STAT3)-interferon regulatory factor 1 (IRF-1) pathway, along with iNOS induction, and cytokine/chemokine production in vascular endothelial cells and C57BL/6 mice, ultimately contributing to ALI. ALI could be attenuated by JAK2/3 or STAT3 inhibitors (AG490 or S3I-201, respectively), and was relieved in TLR4-/- and iNOS-/- mice. eEVs activate the TLR4/JAK3/STAT3/IRF-1 signaling pathway to induce ALI/ARDS by delivering follistatin-like protein 1 (FSTL1), and FSTL1 knockdown in eEVs alleviates eEV-induced ALI/ARDS. Our data thus demonstrate that cardiopulmonary bypass may increase plasma PAI-1 levels to induce FSTL1-enriched eEVs, which target the TLR4-mediated JAK2/3/STAT3/IRF-1 signaling cascade and form a positive feedback loop, leading to ALI/ARDS after cardiac surgery. Our findings provide new insight into the molecular mechanisms and therapeutic targets for ALI/ARDS after cardiac surgery.

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.

CODD-Pred: A Web Server for Efficient Target Identification and Bioactivity Prediction of Small Molecules.

Target identification and bioactivity prediction are critical steps in the drug discovery process. Here we introduce CODD-Pred (COmprehensive Drug Design Predictor), an online web server with well-curated data sets from the GOSTAR database, which is designed with a dual purpose of predicting potential protein drug targets and computing bioactivity values of small molecules. We first designed a double molecular graph perception (DMGP) framework for target prediction based on a large library of 646 498 small molecules interacting with 640 human targets. The framework achieved a top-5 accuracy of over 80% for hitting at least one target on both external validation sets. Additionally, its performance on the external validation set comprising 200 molecules surpassed that of four existing target prediction servers. Second, we collected 56 targets closely related to the occurrence and development of cancer, metabolic diseases, and inflammatory immune diseases and developed a multi-model self-validation activity prediction (MSAP) framework that enables accurate bioactivity quantification predictions for small-molecule ligands of these 56 targets. CODD-Pred is a handy tool for rapid evaluation and optimization of small molecules with specific target activity. CODD-Pred is freely accessible at http://codd.iddd.group/.

Neutrophil to Lymphocyte ratio as a predictor for immune-related adverse events in cancer patients treated with immune checkpoint inhibitors: a systematic review and meta-analysis.

Background: The use of immune checkpoint inhibitors (ICIs) in cancer treatment has led to an increase in immune-related adverse events (irAEs), which can cause treatment discontinuation and even fatal reactions. The purpose of this study was to evaluate the usefulness of the peripheral biomarker neutrophil to lymphocyte ratio (NLR) in predicting irAEs. Methods: A systematic search of databases was conducted to identify studies on the predictive value of NLR for irAEs. The standardized mean difference (SMD) was used to compare continuous NLR, while crude odds ratios (ORs) were calculated for categorized NLR if adjusted ORs and 95% confidence intervals (CIs) were not provided in the original study. Results: The meta-analysis included 47 studies with a total of 11,491 cancer patients treated with ICIs. The baseline continuous NLR was significantly lower in patients with irAEs compared to those without (SMD=-1.55, 95%CI=-2.64 to -0.46, P=0.006). Similarly, categorized NLR showed that lower baseline NLR was associated with increased irAEs (OR=0.55, 95%CI=0.41-0.73, P<0.001). Subgroup analysis revealed that the OR for predicting irAEs with NLR cut-off values of 3 and 5 was 0.4 and 0.59, respectively. Interestingly, increased baseline NLR was associated with a higher incidence of immune-related liver injury (OR=2.44, 95%CI=1.23-4.84, I2 = 0%, P=0.010). Conclusion: Our study suggests that lower baseline NLR is associated with a higher risk of overall irAEs. However, further studies are needed to determine the best cut-off value and explore the efficacy of NLR in predicting specific types of irAEs.

Real-World Effectiveness and Safety of Hydrogen Inhalation in Chinese Patients with Type 2 Diabetes: A Single-Arm, Retrospective Study.

Aim: To evaluate the real-life effectiveness and safety of Chinese patients with type 2 diabetes mellitus (T2DM) receiving hydrogen inhalation (HI) treatment as a supplementary treatment. Methods: This retrospective, multicenter, observational 6-months clinical study included T2DM patients maintaining HI, visited at 4 time points. The primary outcome is the mean change in glycated hemoglobin (HbA1c) at the end of the study compared to baseline. The secondary outcome is analyzing the mean change of fasting plasma glucose (FPG), weight, lipid profile, insulin dose and homeostasis model assessment. Linear regression and logistics regression are applied to evaluate the effect of HI after the treatment. Results: Of the 431 patients comprised, it is observed a significant decrease in HbA1c level (9.04±0.82% at baseline to 8.30±0.99% and 8.00±0.80% at the end, p<0.001), FPG (165.6±40.2 mg/dL at baseline to 157.1±36.3mg/dL and 143.6±32.3mg/dL at the end, p<0.001), weight (74.7±7.1kg at baseline to 74.8±10.0kg and 73.6±8.1kg at the end, p<0.001), insulin dose (49.3±10.8U/d at baseline to 46.7±8.0U/d and 45.2±8.7U/d, p<0.001). The individuals in subgroup with higher baseline HbA1c and longer daily HI time duration gain greater HbA1c decrease after 6 months. Linear regression shows that higher baseline HbA1c level and shorter diabetes duration are significantly in relation to greater HbA1c reduction. Logistics regression reveals that lower weight is associated with a higher possibility of reaching HbA1c<7%. The most common adverse event is hypoglycemia. Conclusion: HI therapy significantly improves glycemic control, weight, insulin dose, lipid metabolism, ß-cell function and insulin resistance of patients with type 2 diabetes after 6 months. Higher baseline HbA1c level and shorter diabetes duration is related to greater clinical response to HI.

A propensity score-matched analysis of neoadjuvant chemoimmunotherapy versus surgery alone for locally advanced esophageal squamous cell carcinoma.

BACKGROUND: The aim of this study was to evaluate the safety, efficacy, and oncologic outcomes of neoadjuvant immunotherapy combined with chemotherapy (NICT) group and surgery alone group in the treatment of patients with locally advanced esophageal squamous cell carcinoma (ESCC). METHODS: A series of 232 consecutive patients who underwent surgery with or without NICT from June 2019 to August 2022 were evaluated. We performed propensity score matching between the NICT and surgery alone groups on the basis of estimated propensity scores for each patient. RESULTS: After propensity score matching, data of 137 patients with clinical stages II-IV ESCC, including 85 receiving surgery alone and 52 receiving NICT, were analyzed. Compared with the surgery alone group (301.7 ± 94.4 min), the operation time was significantly longer in the NICT group (333.4 ± 79.7 min). However, there was no significant difference between the two groups in the postoperative complications, intraoperative blood loss, thoracic fluid volume, chest tube duration, lengths of intensive care unit stay and postoperative hospitalization. Additionally, 90-day mortality rate and 30-day readmission were similar in both groups. CONCLUSIONS: Overall, NICT followed by esophagectomy appears to be safe and feasible for locally advanced ESCC. However, further multicenter prospective clinical trials are needed to validate our results.

Improving drug-target affinity prediction via feature fusion and knowledge distillation.

Rapid and accurate prediction of drug-target affinity can accelerate and improve the drug discovery process. Recent studies show that deep learning models may have the potential to provide fast and accurate drug-target affinity prediction. However, the existing deep learning models still have their own disadvantages that make it difficult to complete the task satisfactorily. Complex-based models rely heavily on the time-consuming docking process, and complex-free models lacks interpretability. In this study, we introduced a novel knowledge-distillation insights drug-target affinity prediction model with feature fusion inputs to make fast, accurate and explainable predictions. We benchmarked the model on public affinity prediction and virtual screening dataset. The results show that it outperformed previous state-of-the-art models and achieved comparable performance to previous complex-based models. Finally, we study the interpretability of this model through visualization and find it can provide meaningful explanations for pairwise interaction. We believe this model can further improve the drug-target affinity prediction for its higher accuracy and reliable interpretability.

The comparison of manual and mechanical anastomosis after total pharyngolaryngoesophagectomy.

Background: Total pharyngolaryngoesophagectomy (TPLE) is considered as a curative treatment for hypopharynx cancer and cervical esophageal carcinomas (HPCECs). Traditional pharyngo-gastric anastomosis is usually performed manually, and postoperative complications are common. The aim of this study was to introduce a new technique for mechanical anastomosis and to evaluate perioperative outcomes and prognosis. Methods: From May 1995 to Nov 2021, a series of 75 consecutive patients who received TPLE for a pathological diagnosis of HPCECs at Sun Yat-sen Memorial Hospital were evaluated. Mechanical anastomosis was performed in 28 cases and manual anastomosis was performed in 47 cases. The data from these patients were retrospectively analyzed. Results: The mean age was 57.6 years, and 20% of the patients were female. The rate of anastomotic fistula and wound infection in the mechanical group were significantly lower than that in the manual group. The operation time, intraoperative blood loss and postoperative hospital stays were significantly higher in the manual group than that in the mechanical group. The R0 resection rate and the tumor characteristics were not significantly different between groups. There was no significant difference in overall survival and disease-free survival between the two groups. Conclusion: The mechanical anastomosis technology adopted by this study was shown to be a safer and more effective procedure with similar survival comparable to that of manual anastomosis for the HPCECs patients.

Feeding Strategies for Adapting Lake Sturgeon (Acipenser fulvescens) Larvae to Formulated Diets at Early Life Stages.

Cost-effective feeding management is required to support conservation hatcheries for lake sturgeon (Acipenser fulvescens), an ecologically important species in the Great Lakes region. This study investigated an approach to transition lake sturgeon larvae from live feed (Artemia) to formulated feed and its effect on growth performance, survival, and response to acute hypoxia stress. The first experiment showed that sturgeon had similar (p > 0.05) growth and survival when fed Artemia or the combined feeding of Artemia with the commercial diet (crude protein, 551 g/kg diet). Feeding solely on the commercial or lab-made (crude protein, 491 g/kg diet) diet significantly reduced growth and survival (p < 0.05). In the second experiment, the growth performance of sturgeon (14 days post-hatch, DPH) fed with either Artemia only or combined feeding different feeding durations of two, three, and four weeks followed by a complete transition to the commercial diet. At the end of six weeks, the 3- and 4-week combined feeding periods resulted in significantly higher body weight and survival compared to the 2-week combined and the Artemia only feeding treatments. In the last experiment, sturgeons (27 DPH) were fed only with Artemia or combined feeding of Artemia with the commercial diet for four weeks followed by the complete transition to the commercial diet for two weeks. Eighteen fish from each treatment were investigated the response to acute hypoxic conditions (gradual decrease in dissolved oxygen level from 8 to 2.3 mg/L at the rate of 1 mg/L per hour). When the dissolved oxygen was between 3 and 4 mg/L, the mortality rate of the combination-fed sturgeon (11.7%) was significantly lower than those fed only Artemia (83.3%). These results clearly demonstrate that a commercial diet can partially replace Artemia at early life stages to improve growth, survival, and hypoxia tolerance and thus its co-feeding with Artemia is recommended.

Several non-salt and solid thienopyridine derivatives as oral P2Y12 receptor inhibitors with good stability.

A series of novel thienopyridine derivatives were designed and synthesized as P2Y12 receptor inhibitors. Several solid compounds were assessed for inhibitory effect where they exhibited stronger potency than clopidogrel. Compound 6b and 6g were evaluated for metabolism to verify that they could overcome clopidogrel resistance and for toxicity where they showed lower toxicity than prasugrel. Compound 6b exhibited lower risk of bleeding than prasugrel and showed good stability under stress testing. Overall, as a promising antiplatelet agent, representative compound 6b showed the following advantages: (1) no drug resistance for CYP2C19 poor metabolizers; (2) higher potency than clopidogrel; (3) lower toxicity than prasugrel; (4) lower risk of bleeding than prasugrel; (5) good stability as a non-salt solid.

Acute ammonia stress-induced oxidative and heat shock responses modulated by transcription factors in Litopenaeus vannamei.

The present study aimed to examine the effects of short-term exposure to ammonia on stress and oxidative responses in shrimp (Litopenaeus vannamei) and to determine whether the antioxidant system related to the regulatory role of transcription factors and stress proteins was activated. Shrimp were exposed ammonia-N at four concentrations: 0 (control), 5, 10, and 15 mg/L, for 48 h. The hepatopancreas was sampled to measure the levels of glutathione (GSH), malondialdehyde (MDA), nitric oxide (NO); the activities of superoxide dismutase (SOD), catalase (CAT), nitric oxide synthase (NOS); and the expression levels of GSH-px (encoding glutathione peroxidase), GST (encoding glutathione-S-transferase), HSP70 (encoding heat shock protein 70), HSP90 (encoding heat shock protein 90), p53, RELISH, and AKIRIN. We observed that exposure to a high ammonia content increased the abundance of oxidative factors (MDA, CAT, SOD, NOS, and NO), reduced the levels of GSH, and upregulated the mRNA expression levels of antioxidant genes (GSH-px and GST), stress-related genes (HSP70 and HSP90), and transcription factor genes (p53, RELISH, and AKIRIN). These results indicated that ammonia induced oxidative stress and inflammation. Both enzymatic and nonenzymatic antioxidant defense systems are involved, which might be regulated by HSPs, as well as certain transcription factors, such as p53 and nuclear factor kappa B (NF-κB), thus mounting an adaptive response to help rebalance redox homoeostasis.

Size Distribution of Microparticles: A New Parameter to Predict Acute Lung Injury After Cardiac Surgery With Cardiopulmonary Bypass.

Background: Acute lung injury (ALI) is a common complication after cardiac surgery with cardiopulmonary bypass (CPB). No precise way, however, is currently available to predict its occurrence. We and others have demonstrated that microparticles (MPs) can induce ALI and were increased in patients with ALI. However, whether MPs can be used to predict ALI after cardiac surgery with CPB remains unknown. Methods: In this prospective study, 103 patients undergoing cardiac surgery with CPB and 53 healthy subjects were enrolled. MPs were isolated from the plasma before, 12 h after, and 3 d after surgery. The size distributions of MPs were measured by the LitesizerTM 500 Particle Analyzer. The patients were divided into two subgroups (ALI and non-ALI) according to the diagnosis of ALI. Descriptive and correlational analyzes were conducted between the size distribution of MPs and clinical data. Results: Compared to the non-ALI group, the size at peak and interquartile range (IQR) of MPs in patients with ALI were smaller, but the peak intensity of MPs is higher. Multivariate logistic regression analysis indicated that the size at peak of MPs at postoperative 12 h was an independent risk factor for ALI. The area under the curve (AUC) of peak diameter at postoperative 12 h was 0.803. The best cutoff value of peak diameter to diagnose ALI was 223.05 nm with a sensitivity of 88.0% and a negative predictive value of 94.5%. The AUC of IQR at postoperative 12 h was 0.717. The best cutoff value of IQR to diagnose ALI was 132.65 nm with a sensitivity of 88.0% and a negative predictive value of 92.5%. Combining these two parameters, the sensitivity reached 92% and the negative predictive value was 96%. Conclusions: Our findings suggested that the size distribution of MPs could be a novel biomarker to predict and exclude ALI after cardiac surgery with CPB.

Out-of-the-box deep learning prediction of quantum-mechanical partial charges by graph representation and transfer learning.

Accurate prediction of atomic partial charges with high-level quantum mechanics (QM) methods suffers from high computational cost. Numerous feature-engineered machine learning (ML)-based predictors with favorable computability and reliability have been developed as alternatives. However, extensive expertise effort was needed for feature engineering of atom chemical environment, which may consequently introduce domain bias. In this study, SuperAtomicCharge, a data-driven deep graph learning framework, was proposed to predict three important types of partial charges (i.e. RESP, DDEC4 and DDEC78) derived from high-level QM calculations based on the structures of molecules. SuperAtomicCharge was designed to simultaneously exploit the 2D and 3D structural information of molecules, which was proved to be an effective way to improve the prediction accuracy of the model. Moreover, a simple transfer learning strategy and a multitask learning strategy based on self-supervised descriptors were also employed to further improve the prediction accuracy of the proposed model. Compared with the latest baselines, including one GNN-based predictor and two ML-based predictors, SuperAtomicCharge showed better performance on all the three external test sets and had better usability and portability. Furthermore, the QM partial charges of new molecules predicted by SuperAtomicCharge can be efficiently used in drug design applications such as structure-based virtual screening, where the predicted RESP and DDEC4 charges of new molecules showed more robust scoring and screening power than the commonly used partial charges. Finally, two tools including an online server (http://cadd.zju.edu.cn/deepchargepredictor) and the source code command lines (https://github.com/zjujdj/SuperAtomicCharge) were developed for the easy access of the SuperAtomicCharge services.