AutoAMS: Automated attention-based multi-modal graph learning architecture search.

Multi-modal attention mechanisms have been successfully used in multi-modal graph learning for various tasks. However, existing attention-based multi-modal graph learning (AMGL) architectures heavily rely on manual design, requiring huge effort and expert experience. Meanwhile, graph neural architecture search (GNAS) has made great progress toward automatically designing graph-based learning architectures. However, it is challenging to directly adopt existing GNAS methods to search for better AMGL architectures because of the search spaces that only focus on designing graph neural network architectures and the search objective that ignores multi-modal interactive information between modalities and long-term content dependencies within different modalities. To address these issues, we propose an automated attention-based multi-modal graph learning architecture search (AutoAMS) framework, which can automatically design the optimal AMGL architectures for different multi-modal tasks. Specifically, we design an effective attention-based multi-modal (AM) search space consisting of four sub-spaces, which can jointly support the automatic search of multi-modal attention representation and other components of multi-modal graph learning architecture. In addition, a novel search objective based on an unsupervised multi-modal reconstruction loss and task-specific loss is introduced to search and train AMGL architectures. The search objective can extract the global features and capture multi-modal interactions from multiple modalities. The experimental results on multi-modal tasks show strong evidence that AutoAMS is capable of designing high-performance AMGL architectures.

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.

Sleeve Gastrectomy Surgery makes Obstructive Sleep Apnea Worse or Better?: a Multi-Center Observational Study in Patients with Obesity.

BACKGROUND AND OBJECTIVE: Obstructive sleep apnea (OSA) is highly prevalent in the bariatric surgical population, with rates ranging from 50 to 70%. The impact of laparoscopic sleeve gastrectomy (LSG) on OSA and its associated risk factors remain relatively understudied. The aim of this study is to assess the effect of LSG on OSA and investigate predictors of new or worsening OSA postoperatively. Additionally, the study aims to provide evidence for the individualized selection of LSG procedures based on patient characteristics. METHODS: This multi-center observational study enrolled 119 patients with obesity who underwent LSG and were subdivided into two groups based on their preoperative AHI: AHI < 15 and AHI ≥ 15. The patients were followed up and evaluated before and 30 days after LSG. The study utilized univariate and multivariate analyses to assess risk factors for postoperative AHI development. RESULTS: Following LSG, there was a significant decrease in the mean AHI, leading to the resolution of OSA symptoms in 67.6% of patients with AHI ≥ 15. Neck circumference and the number of METS were also identified as independent risk factors for postoperative OSA. Furthermore, preoperative hypertension was found to be a significant predictor of new or worsened OSA after LSG. CONCLUSION: LSG demonstrated effectiveness in improving OSA among patients with obesity. The study highlights the importance of preoperative hypertension evaluation and postoperative management in patients undergoing LSG. Further long-term, multicenter, and large-scale studies are recommended to validate and generalize these findings to diverse patient populations.

Characterization of immunogenic cell death regulators predicts survival and immunotherapy response in lung adenocarcinoma.

Lung adenocarcinoma (LUAD) is highly lethal tumor; understanding immune response is crucial for current effective treatment. Research investigated immunogenic cell death (ICD) impact on LUAD through 75 ICD-related genes which encompass cell damage, endoplasmic reticulum stress, microenvironment, and immunity. Transcriptome data and clinical info were analyzed, revealing two ICD-related clusters: B, an immune osmotic subgroup, had better prognosis, stronger immune signaling, and higher infiltration, while A represented an immune-deficient subgroup. Univariate Cox analysis identified six prognostic genes (AGER, CD69, CD83, CLEC9A, CTLA4, and NT5E), forming a validated risk score model. It was validated across datasets, showing predictive performance. High-risk group had unfavorable prognosis, lower immune infiltration, and higher chemotherapy sensitivity. Conversely, low-risk group had better prognosis, higher immune infiltration, and favorable immunotherapy response. The key gene NT5E was examined via immunohistochemistry, with higher expression linked to poorer prognosis. NT5E was predominantly expressed in B cells, fibroblasts, and endothelial cells, correlated with immune checkpoints. These outcomes suggest that NT5E can serve as a LUAD therapeutic target. The study highlights gene predictive value, offers an efficient tumor assessment tool, guides clinical treatment strategies, and identifies NT5E as therapeutic target for LUAD.

The E3 ubiquitin ligase RBCK1: Implications in the tumor immune microenvironment and antiangiogenic therapy of glioma.

E3 ubiquitin ligases (E3s) play a pivotal role in regulating the specificity of protein ubiquitination, and their significant functions as regulators of immune responses against tumors are attracting considerable interest. RBCK1-an RBR E3 ligase-is involved in immune regulation and tumor development. However, the potential effect of RBCK1 on glioma remains enigmatic. In the present study, we performed comprehensive analyses of multilevel data, which disclosed distribution characteristics of RBCK1 in pan-cancer, especially in glioma. Functional roles of RBCK1 were further confirmed using immunohistochemistry, cell biological assays, and xenograft experiments. Aberrant ascending of RBCK1 in multiple types of cancer was found to remodel the immunosuppressive microenvironment of glioma by regulating immunomodulators, cancer immunity cycles, and immune cell infiltration. Notably, the MES-like/RBCK1High cell population, a unique subset of cells in the microenvironment, suppressed T cell-mediated cell killing in glioma. Elevated expression levels of RBCK1 suggested a glioma subtype characterized by immunosuppression and hypo-responsiveness to immunotherapy but manifesting surprisingly increased responses to anti-angiogenic therapy. In conclusion, anti-RBCK1 target therapy might be beneficial for glioma treatment. Moreover, RBCK1 assisted in predicting molecular subtypes of glioma and response rates of patients to different clinical treatments, which could guide personalized therapy.

Prevalence and associated factors of vitamin D deficiency after Roux-en-Y gastric bypass: a systematic review and meta-analysis.

OBJECTS: To estimate the prevalence and associated factors of vitamin D deficiency (VDD) after Roux-en-Y gastric bypass (RYGB). METHODS: PubMed, EMBASE, and CENTRAL were searched for relevant records from inception to 17 March 2023, using search terms: vitamin D, vitamin D3, vitamin D deficiency, hypovitaminosis D, gastric bypass, and RYGB. Studies were eligible for inclusion if they provided related data on VDD prevalence after RYGB. RESULTS: Of the 1119 screened studies, 72 studies involving 7688 individuals were enrolled in the final analysis. The prevalence estimates of VDD after RYGB were 42%. Subgroup analyses suggested the pooled prevalence of postoperative VDD was 35% for follow-up duration less than or equal to 1 year, 43% for greater than 1 and less than or equal to 5 years, and 54% for greater than 5 years. Meta-regression showed that VDD prevalence was positively correlated with follow-up time. Also, the prevalence was higher in studies with inadequate vitamin D supplementation than in those with adequate supplementation and in Asia population than in those from South America, Europe, and North America. Other factors associated with high VDD prevalence after RYGB included high presurgical VDD prevalence, noncompliant patients, and black populations. No significant association existed between VDD and alimentary length. CONCLUSION: VDD presented a high prevalence in patients following RYGB. It occurred more frequently with longer postoperative follow-up time. Population-specific vitamin D supplementation measures, targeted treatment for presurgical VDD, improved patient compliance, and periodical follow-ups were necessary to reduce VDD and other adverse outcomes.

Cancer drug response prediction with surrogate modeling-based graph neural architecture search.

MOTIVATION: Understanding drug-response differences in cancer treatments is one of the most challenging aspects of personalized medicine. Recently, graph neural networks (GNNs) have become state-of-the-art methods in many graph representation learning scenarios in bioinformatics. However, building an optimal handcrafted GNN model for a particular drug sensitivity dataset requires manual design and fine-tuning of the hyperparameters for the GNN model, which is time-consuming and requires expert knowledge. RESULTS: In this work, we propose AutoCDRP, a novel framework for automated cancer drug-response predictor using GNNs. Our approach leverages surrogate modeling to efficiently search for the most effective GNN architecture. AutoCDRP uses a surrogate model to predict the performance of GNN architectures sampled from a search space, allowing it to select the optimal architecture based on evaluation performance. Hence, AutoCDRP can efficiently identify the optimal GNN architecture by exploring the performance of all GNN architectures in the search space. Through comprehensive experiments on two benchmark datasets, we demonstrate that the GNN architecture generated by AutoCDRP surpasses state-of-the-art designs. Notably, the optimal GNN architecture identified by AutoCDRP consistently outperforms the best baseline architecture from the first epoch, providing further evidence of its effectiveness. AVAILABILITY AND IMPLEMENTATION: https://github.com/BeObm/AutoCDRP.

Long-Term Results of Bariatric Surgery in Adolescents with at Least 5 Years of Follow-up: a Systematic Review and Meta-Analysis.

OBJECTS: The purpose of this study was to investigate the long-term outcomes of bariatric surgery in adolescents with obesity by including studies with a follow-up of at least 5 years. METHODS: PubMed, EMBASE, and CENTRAL were systematically searched. Studies that met the criteria were included in the analysis. RESULT: We identified 29 cohort studies with a total population of 4970. Preoperative age ranged from 12 to 21 years; body mass index (BMI) ranged from 38.9 to 58.5 kg/m2. Females were the predominant gender (60.3%). After at least 5-year of follow-up, the pooled BMI decline was 13.09 kg/m2 (95%CI 11.75-14.43), with sleeve gastrectomy (SG) was 15.27 kg/m2, Roux-en-Y gastric bypass (RYGB) was 12.86 kg/m2, and adjustable gastric banding (AGB) was 7.64 kg/m2. The combined remission rates of type 2 diabetes mellitus (T2DM), dyslipidemia, hypertension (HTN), obstructive sleep apnea (OSA), and asthma were 90.0%, 76.6%, 80.7%, 80.8%, and 92.5%, (95%CI 83.2-95.6, 62.0-88.9, 71.5-88.8, 36.4-100, and 48.5-100), respectively. Postoperative complications were underreported. Combined with the current study, we found a low level of postoperative complications. Iron and vitamin B12 deficiencies were the main nutritional deficiency complications identified so far. CONCLUSION: For adolescents with severe obesity, bariatric surgery (especially RYGB and SG) is the independent and effective treatment option. After at least 5 years of follow-up, bariatric surgery in adolescents showed a desirable reduction in BMI and significant remission of T2DM, dyslipidemia, and HTN. Surgical and nutrition-related complications still need to be further explored by more long-term studies.

Prevalence of Rhabdomyolysis Following Bariatric Surgery and its Associated Risk Factors: a Meta-Analysis.

PURPOSE: This study aimed to evaluate the prevalence of rhabdomyolysis (RML) following bariatric surgery and potential associated factors. MATERIALS AND METHODS: We systematically searched PubMed, Embase, and CENTRAL for relevant trials from database inception through August 2022. Articles were eligible for inclusion if they reported the prevalence of RML after bariatric surgery and provided at least one of the following outcome indicators: preoperative mean BMI/mean operative time for the included population. RESULTS: Sixteen studies with a total of 1540 patients were analyzed. The mean preoperative age distribution of the included patients was centered between 32.9 and 47.0 years, and the mean preoperative BMI ranged from 42.3 to 60.0 kg/m2. The operative time varied between 126.7 and 403.3 min. The overall pooled crude prevalence of post-bariatric surgery RML was 19.4%. Subgroup analyses showed the pooled prevalence of RML was 8.1% for operative duration > 120 and ≤ 180 min, 32.8% for > 180 and ≤ 240 min, and 47.4% for > 240 min. Meta-regression revealed that operation time was an independent risk factor for developing RML. Besides, BMI > 50 kg/m2 and open Roux-en-Y gastric bypass (RYGB) indicated a higher risk of RML. CONCLUSION: Post-bariatric surgery RML prevalence occurred more frequently with the extension of the operation time. For bariatric subjects with surgery time > 180 min, open RYGB, or BMI > 50 kg/m2, CKP could be routinely measured early to verify the presence of RML and to actively prevent its fatal complications.

lncRNA PCGEM1 Regulates the Progress of Colorectal Cancer through Targeting miR-129-5p/SOX4.

Prostate cancer gene expression marker 1 (PCGEM1) has abnormal expression level in a variety of malignant tumor. However, the relationship between PCGEM1 and colorectal cancer is still unclear yet. This study is aimed at identifying the role of PCGEM1 in colorectal cancer. qRT-PCR was used to examine the expressions of the expression of lncRNA PCGEM1 and SOX4 in CRC tissues and cell lines. The biological functions of lncRNA PCGEM1 and SOX4 were examined by CCK-8 assay, Transwell assay, immunohistochemistry, western blotting, RNA interference, and gene overexpression techniques. Bioinformatics analysis was used to find the potential downstream molecule of PCGEM1 and miR-129-5p. The relationship between PCGEM1, miR-129-5p, and SOX4 was assessed by dual luciferase activity assay. We found that PCGEM1 is overexpressed in colorectal cancer cells and tissues, while miR-129-5p is underexpressed. SOX4 is overexpressed in colorectal cancer cells and tissues. Functionally, PCGEM1 silencing can significantly inhibit the proliferation, invasion, and migration of colorectal cancer cells. Mechanically, PCGEM1 acted as a sponge for miR-129-5p and absorbed its expression, and miR-129-5p was found to target SOX4, constructing the axis of PCGEM1/miR-129-5p/SOX4 in colorectal cancer. In conclusion, PCGEM1 mediates the proliferation, invasion, and migration of colorectal cancer cells by targeting miR-129-5p/SOX4 axis.

Prevalence and associated factors of secondary hyperparathyroidism after Roux-en-Y gastric bypass: A meta-analysis.

This study aimed to investigate the prevalence and factors associated with secondary hyperparathyroidism (SHPT) after Roux-en-Y gastric bypass (RYGB). We searched PubMed, EMBASE, and CENTRAL for relevant studies using search terms gastric bypass, RYGB and hyperparathyroidism. Thirty-four cohort studies with 4331 patients were incorporated into the final meta-analysis. Overall estimates of the prevalence of SHPT following RYGB were 39%. Subgroup analyses indicated the pooled prevalences of SHPT were 25%, 42%, 48%, and 54% for ≤1 year, >1 and ≤5 years, >5 and ≤10 years, and >10 years, respectively, after RYGB. Meta-regression showed that SHPT occurred was positively related to follow-up durations (p = 0.001). Additionally, SHPT prevalence was higher in studies in which calcium and vitamin D supplementation were considered inadequate than in those which were adequate (p = 0.002). SHPT is highly prevalent in individuals with obesity after RYGB. It seems to progress with time after surgery. Routine calcium and vitamin D supplementation post-RYGB together with targeted treatment of vitamin D deficiency, reasonable adjustment of the doses of supplementation with regular follow-up, and improved patient compliance, as well as long-term screening, are necessary to prevent the development of SHPT.

Characterization of m6A RNA Methylation Regulators Predicts Survival and Immunotherapy in Lung Adenocarcinoma.

N6-methyladenosine (m6A) RNA modification is a reversible mechanism that regulates eukaryotic gene expression. Growing evidence has demonstrated an association between m6A modification and tumorigenesis and response to immunotherapy. However, the overall influence of m6A regulators on the tumor microenvironment and their effect on the response to immunotherapy in lung adenocarcinoma remains to be explored. Here, we comprehensively analyzed the m6A modification patterns of 936 lung adenocarcinoma samples based on 24 m6A regulators. First, we described the features of genetic variation in these m6A regulators. Many m6A regulators were aberrantly expressed in tumors and negatively correlated with most tumor-infiltrating immune cell types. Furthermore, we identified three m6A modification patterns using a consensus clustering method. m6A cluster B was preferentially associated with a favorable prognosis and enriched in metabolism-associated pathways. In contrast, m6A cluster A was associated with the worst prognosis and was enriched in the process of DNA repair. m6A cluster C was characterized by activation of the immune system and a higher stromal cell score. Surprisingly, patients who received radiotherapy had a better prognosis than patients without radiotherapy only in the m6A cluster C group. Subsequently, we constructed an m6A score model that qualified the m6A modification level of individual samples by using principal component analysis algorithms. Patients with high m6A score were characterized by enhanced immune cell infiltration and prolonged survival time and were associated with lower tumor mutation burden and PD-1/CTLA4 expression. The combination of the m6A score and tumor mutation burden could accurately predict the prognosis of patients with lung adenocarcinoma. Furthermore, patients with high m6A score exhibited greater prognostic benefits from radiotherapy and immunotherapy. This study demonstrates that m6A modification is significantly associated with tumor microenvironment diversity and prognosis. A comprehensive evaluation of m6A modification patterns in single tumors will expand our understanding of the tumor immune landscape. In addition, our m6A score model demonstrated that the level of immune cell infiltration plays a significant role in cancer immunotherapy and provides a basis to increase the efficiency of current immune therapies and promote the clinical success of immunotherapy.