Deep multi-view contrastive learning for cancer subtype identification.

Cancer heterogeneity has posed great challenges in exploring precise therapeutic strategies for cancer treatment. The identification of cancer subtypes aims to detect patients with distinct molecular profiles and thus could provide new clues on effective clinical therapies. While great efforts have been made, it remains challenging to develop powerful computational methods that can efficiently integrate multi-omics datasets for the task. In this paper, we propose a novel self-supervised learning model called Deep Multi-view Contrastive Learning (DMCL) for cancer subtype identification. Specifically, by incorporating the reconstruction loss, contrastive loss and clustering loss into a unified framework, our model simultaneously encodes the sample discriminative information into the extracted feature representations and well preserves the sample cluster structures in the embedded space. Moreover, DMCL is an end-to-end framework where the cancer subtypes could be directly obtained from the model outputs. We compare DMCL with eight alternatives ranging from classic cancer subtype identification methods to recently developed state-of-the-art systems on 10 widely used cancer multi-omics datasets as well as an integrated dataset, and the experimental results validate the superior performance of our method. We further conduct a case study on liver cancer and the analysis results indicate that different subtypes might have different responses to the selected chemotherapeutic drugs.

Identifying spatial domains of spatially resolved transcriptomics via multi-view graph convolutional networks.

MOTIVATION: Recent advances in spatially resolved transcriptomics (ST) technologies enable the measurement of gene expression profiles while preserving cellular spatial context. Linking gene expression of cells with their spatial distribution is essential for better understanding of tissue microenvironment and biological progress. However, effectively combining gene expression data with spatial information to identify spatial domains remains challenging. RESULTS: To deal with the above issue, in this paper, we propose a novel unsupervised learning framework named STMGCN for identifying spatial domains using multi-view graph convolution networks (MGCNs). Specifically, to fully exploit spatial information, we first construct multiple neighbor graphs (views) with different similarity measures based on the spatial coordinates. Then, STMGCN learns multiple view-specific embeddings by combining gene expressions with each neighbor graph through graph convolution networks. Finally, to capture the importance of different graphs, we further introduce an attention mechanism to adaptively fuse view-specific embeddings and thus derive the final spot embedding. STMGCN allows for the effective utilization of spatial context to enhance the expressive power of the latent embeddings with multiple graph convolutions. We apply STMGCN on two simulation datasets and five real spatial transcriptomics datasets with different resolutions across distinct platforms. The experimental results demonstrate that STMGCN obtains competitive results in spatial domain identification compared with five state-of-the-art methods, including spatial and non-spatial alternatives. Besides, STMGCN can detect spatially variable genes with enriched expression patterns in the identified domains. Overall, STMGCN is a powerful and efficient computational framework for identifying spatial domains in spatial transcriptomics data.

CasANGCL: pre-training and fine-tuning model based on cascaded attention network and graph contrastive learning for molecular property prediction.

MOTIVATION: Molecular property prediction is a significant requirement in AI-driven drug design and discovery, aiming to predict the molecular property information (e.g. toxicity) based on the mined biomolecular knowledge. Although graph neural networks have been proven powerful in predicting molecular property, unbalanced labeled data and poor generalization capability for new-synthesized molecules are always key issues that hinder further improvement of molecular encoding performance. RESULTS: We propose a novel self-supervised representation learning scheme based on a Cascaded Attention Network and Graph Contrastive Learning (CasANGCL). We design a new graph network variant, designated as cascaded attention network, to encode local-global molecular representations. We construct a two-stage contrast predictor framework to tackle the label imbalance problem of training molecular samples, which is an integrated end-to-end learning scheme. Moreover, we utilize the information-flow scheme for training our network, which explicitly captures the edge information in the node/graph representations and obtains more fine-grained knowledge. Our model achieves an 81.9% ROC-AUC average performance on 661 tasks from seven challenging benchmarks, showing better portability and generalizations. Further visualization studies indicate our model's better representation capacity and provide interpretability.

Differentiation of Uc-MSCs into insulin secreting islet-like clusters by trypsin through TGF-beta signaling pathway.

Differentiation of human umbilical cord mesenchymal stem cells (Uc-MSCs) into islet-like clusters which are capable of synthesizing and secreting insulin can potentially serve as donors for islet transplantation in the patient deficiency in islet ß cell function both in type 1 or type 2 diabetic patients. Therefore, we developed an easy and higher efficacy approach by trypsinazing the Uc-MSCs and followed culture in differentiation medium to induce of Uc-MSCs differentiation into islet-like clusters, and the potential mechanism that in the early stage of differentiation was also investigated by using RNA-sequencing and bioinformatics. Results show that induction efficacy was reached to 98% and TGF-ß signaling pathway may play critical role in the early stage differentiation, it was further confirmed that the retardant effect of differentiation progress either in cell morphology or in islet specific genes expression can be observed upon blocking the activation of TGF-ß signaling pathway using specific inhibitor of LY2109761 (TßRI/II kinase inhibitor). Our current study, for the first time, development a protocol for differentiation of Uc-MSCs into islet-like clusters, and revealed the importance of TGF-ß signaling pathway in the early stage of differentiation of Uc-MSCs into islet-like clusters. Our study will provide alternative approach for clinical treatment of either type I or type II diabtes mellitus with dysfunctional pancreatic islets.

Facilitation of Behavioral and Cortical Emergence from Isoflurane Anesthesia by GABAergic Neurons in Basal Forebrain.

General anesthesia shares many similarities with natural sleep in behavior and electroencephalogram (EEG) patterns. The latest evidence suggests that general anesthesia and sleep-wake behavior may share overlapping neural substrates. The GABAergic neurons in the basal forebrain (BF) have recently been demonstrated to play a key role in controlling wakefulness. It was hypothesized that BF GABAergic neurons may participate in the regulation of general anesthesia. Here, using in vivo fiber photometry, we found that the activity of BF GABAergic neurons was generally inhibited during isoflurane anesthesia, having obviously decreased during the induction of anesthesia and being gradually restored during the emergence from anesthesia, in Vgat-Cre mice of both sexes. Activation of BF GABAergic neurons with chemogenetic and optogenetic approaches decreased sensitivity to isoflurane, delayed induction, and accelerated emergence from isoflurane anesthesia. Optogenetic activation of BF GABAergic neurons decreased EEG δ power and the burst suppression ratio (BSR) during 0.8% and 1.4% isoflurane anesthesia, respectively. Similar to the effects of activating BF GABAergic cell bodies, photostimulation of BF GABAergic terminals in the thalamic reticular nucleus (TRN) also strongly promoted cortical activation and behavioral emergence from isoflurane anesthesia. Collectively, these results showed that the GABAergic BF is a key neural substrate for general anesthesia regulation that facilitates behavioral and cortical emergence from general anesthesia via the GABAergic BF-TRN pathway. Our findings may provide a new target for attenuating the depth of anesthesia and accelerating emergence from general anesthesia.SIGNIFICANCE STATEMENT The basal forebrain (BF) is a key brain region controlling sleep-wake behavior. Activation of GABAergic neurons in the BF potently promotes behavioral arousal and cortical activity. Recently, many sleep-wake-related brain structures have been reported to participate in the regulation of general anesthesia. However, it is still unclear what role BF GABAergic neurons play in general anesthesia. In this study, we aim to reveal the role of BF GABAergic neurons in behavioral and cortical emergence from isoflurane anesthesia and elucidate the underlying neural pathways. Understanding the specific role of BF GABAergic neurons in isoflurane anesthesia would improve our understanding of the mechanisms of general anesthesia and may provide a new strategy for accelerating emergence from general anesthesia.

Node-adaptive graph Transformer with structural encoding for accurate and robust lncRNA-disease association prediction.

BACKGROUND: Long noncoding RNAs (lncRNAs) are integral to a plethora of critical cellular biological processes, including the regulation of gene expression, cell differentiation, and the development of tumors and cancers. Predicting the relationships between lncRNAs and diseases can contribute to a better understanding of the pathogenic mechanisms of disease and provide strong support for the development of advanced treatment methods. RESULTS: Therefore, we present an innovative Node-Adaptive Graph Transformer model for predicting unknown LncRNA-Disease Associations, named NAGTLDA. First, we utilize the node-adaptive feature smoothing (NAFS) method to learn the local feature information of nodes and encode the structural information of the fusion similarity network of diseases and lncRNAs using Structural Deep Network Embedding (SDNE). Next, the Transformer module is used to capture potential association information between the network nodes. Finally, we employ a Transformer module with two multi-headed attention layers for learning global-level embedding fusion. Network structure coding is added as the structural inductive bias of the network to compensate for the missing message-passing mechanism in Transformer. NAGTLDA achieved an average AUC of 0.9531 and AUPR of 0.9537 significantly higher than state-of-the-art methods in 5-fold cross validation. We perform case studies on 4 diseases; 55 out of 60 associations between lncRNAs and diseases have been validated in the literatures. The results demonstrate the enormous potential of the graph Transformer structure to incorporate graph structural information for uncovering lncRNA-disease unknown correlations. CONCLUSIONS: Our proposed NAGTLDA model can serve as a highly efficient computational method for predicting biological information associations.

CsSHMT3 gene enhances the growth and development in cucumber seedlings under salt stress.

Salt stress is one of the major factors limiting plant growth and productivity. Many studies have shown that serine hydroxymethyltransferase (SHMT) gene play an important role in growth, development and stress response in plants. However, to date, there have been few studies on whether SHMT3 can enhance salt tolerance in plants. Therefore, the effects of overexpression or silencing of CsSHMT3 gene on cucumber seedling growth under salt stress were investigated in this study. The results showed that overexpression of CsSHMT3 gene in cucumber seedlings resulted in a significant increase in chlorophyll content, photosynthetic rate and proline (Pro) content, and antioxidant enzyme activity under salt stress condition; whereas the content of malondialdehyde (MDA), superoxide anion (H2O2), hydrogen peroxide (O2·-) and relative conductivity were significantly decreased when CsSHMT3 gene was overexpressed. However, the content of chlorophyll and Pro, photosynthetic rate, and antioxidant enzyme activity of the silenced CsSHMT3 gene lines under salt stress were significantly reduced, while MDA, H2O2, O2·- content and relative conductivity showed higher level in the silenced CsSHMT3 gene lines. It was further found that the expression of stress-related genes SOD, CAT, SOS1, SOS2, NHX, and HKT was significantly up-regulated by overexpressing CsSHMT3 gene in cucumber seedlings; while stress-related gene expression showed significant decrease in silenced CsSHMT3 gene seedlings under salt stress. This suggests that overexpression of CsSHMT3 gene increased the salt tolerance of cucumber seedlings, while silencing of CsSHMT3 gene decreased the salt tolerance. In conclusion, CsSHMT3 gene might positively regulate salt stress tolerance in cucumber and be involved in regulating antioxidant activity, osmotic adjustment, and photosynthesis under salt stress. KEY MESSAGE: CsSHMT3 gene may positively regulate the expression of osmotic system, photosynthesis, antioxidant system and stress-related genes in cucumber.

Overexpression of COX7A1 Promotes the Resistance of Gastric Cancer to Oxaliplatin and Weakens the Efficacy of Immunotherapy.

Gastric cancer (GC) is one of the most common clinical malignant tumors worldwide, with high morbidity and mortality. Presently, the overall response rate to immunotherapy is low, and current methods for predicting the prognosis of GC are not optimal. Therefore, novel biomarkers with accuracy, efficiency, stability, performance ratio, and wide clinical application are needed. Based on public data sets, the chemotherapy cohort and immunotherapy cohort from Sun Yat-sen University Cancer Center, a series of bioinformatics analyses, such as differential expression analysis, survival analysis, drug sensitivity prediction, enrichment analysis, tumor immune dysfunction and exclusion analysis, single-sample gene set enrichment analysis, stemness index calculation, and immune cell infiltration analysis, were performed for screening and preliminary exploration. Immunohistochemical staining and in vitro experiments were performed for further verification. Overexpression of COX7A1 promoted the resistance of GC cells to Oxaliplatin. COX7A1 may induce immune escape by regulating the number of fibroblasts and their cellular communication with immune cells. In summary, measuring the expression levels of COX7A1 in the clinic may be useful in predicting the prognosis of GC patients, the degree of chemotherapy resistance, and the efficacy of immunotherapy.

A knowledge-driven network for fine-grained relationship detection between miRNA and disease.

Increasing biological evidence indicated that microRNAs (miRNAs) play a vital role in exploring the pathogenesis of various human diseases (especially in tumors). Mining disease-related miRNAs is of great significance for the clinical diagnosis and treatment of diseases. Compared with the traditional experimental methods with the significant limitations of high cost, long cycle and small scale, the methods based on computing have the advantages of being cost-effective. However, although the current methods based on computational biology can accurately predict the correlation between miRNAs and disease, they can not predict the detailed association information at a fine level. We propose a knowledge-driven approach to the fine-grained prediction of disease-related miRNAs (KDFGMDA). Different from the previous methods, this method can finely predict the clear associations between miRNA and disease, such as upregulation, downregulation or dysregulation. Specifically, KDFGMDA extracts triple information from massive experimental data and existing datasets to construct a knowledge graph and then trains a depth graph representation learning model based on knowledge graph to complete fine-grained prediction tasks. Experimental results show that KDFGMDA can predict the relationship between miRNA and disease accurately, which is of far-reaching significance for medical clinical research and early diagnosis, prevention and treatment of diseases. Additionally, the results of case studies on three types of cancers, Kaplan-Meier survival analysis and expression difference analysis further provide the effectiveness and feasibility of KDFGMDA to detect potential candidate miRNAs. Availability: Our work can be downloaded from https://github.com/ShengPengYu/KDFGMDA.

Cyanopyridoimidazole/1,2-Aminothiol Click Reaction: A Novel Bioorthogonal Reaction for Synthesis of Radiotracers.

We herein described the design and synthesis of the cyanopyridoimidazoles (CPIs) as new bioorthogonal click reagents toward 1,2-aminothiol groups. Kinetic and density functional theory-based studies of the synthetic compounds revealed that incorporating an electron-withdrawing substituent into the CPI scaffold lowers its lowest unoccupied molecular orbital energy, consequently increasing reactivity. Optimized CPI 8a showed rapid reactivity and high stability in physiological conditions and has been demonstrated to be suitable for various radiotracer synthetic methods. Based on the new bioorthogonal reaction, a [67Ga]Ga-labeled prostate-specific membrane antigen-targeted probe was successfully prepared for in vivo imaging of prostate cancer in an animal model.

Long-term oncological outcomes of robotic versus laparoscopic gastrectomy for gastric cancer: multicentre cohort study.

BACKGROUND: The aim of this multicentre cohort study was to compare the long-term oncological outcomes of robotic gastrectomy (RG) and laparoscopic gastrectomy (LG) for patients with gastric cancer. METHODS: Patients with gastric cancer who underwent radical gastrectomy by robotic or laparoscopic approaches from 1 March 2010 to 31 December 2018 at 10 high-volume centres in China were selected from institutional databases. Patients receiving RG were matched 1 : 1 by propensity score with patients undergoing LG. The primary outcome was 3-year disease-free survival. Secondary outcomes were overall survival and disease recurrence. RESULTS: Some 2055 patients who underwent RG and 4309 patients who had LG were included. The propensity score-matched cohort comprised 2026 RGs and 2026 LGs. Median follow-up was 41 (i.q.r. 39-58) months for the RG group and 39 (38-56) months for the LG group. The 3-year disease-free survival rates were 80.8% in the RG group and 79.5% in the LG group (log rank P = 0.240; HR 0.92, 95% c.i. 0.80 to 1.06; P = 0.242). Three-year OS rates were 83.9 and 81.8% respectively (log rank P = 0.068; HR 0.87, 0.75 to 1.01; P = 0.068) and the cumulative incidence of recurrence over 3 years was 19.3% versus 20.8% (HR 0.95, 0.88 to 1.03; P = 0.219), with no difference between groups. CONCLUSION: RG and LG in patients with gastric cancer are associated with comparable disease-free and overall survival.

Predictors for new-onset conduction block in patients with pure native aortic regurgitation after transcatheter aortic valve replacement with a new-generation self-expanding valve (VitaFlow Liberty™): a retrospective cohort study.

BACKGROUND: New-generation self-expanding valves can improve the success rate of transcatheter aortic valve replacement (TAVR) for severe pure native aortic regurgitation (PNAR). However, predictors of new-onset conduction block post-TAVR using new-generation self-expanding valves in patients with PNAR remain to be established. Therefore, this study aimed to identify predictors of new-onset conduction block post-TAVR using new-generation self-expanding valves (VitaFlow Liberty™) in patients with PNAR. METHODS: In this retrospective cohort study, patients were categorized into pacemaker and non-pacemaker groups based on their need for new postoperative permanent pacemaker implantation (PPI). Based on the postoperative presence of either new-onset complete left bundle branch block (cLBBB) or high-grade atrioventricular block (AVB), patients were further classified into conduction disorder and non-conduction disorder groups. Laboratory, echocardiographic, computed tomography, preoperative and postoperative electrocardiography, and procedural and clinical data were collected immediately after TAVR and during hospitalization and compared between the groups. Multivariate logistic regression analysis was performed incorporating the significant variables from the univariate analyses. RESULTS: This study examined 68 consecutive patients with severe PNAR who underwent TAVR. In 20 patients, a permanent pacemaker was fitted postoperatively. Multivariate logistic regression analysis revealed an association between the need for postoperative PPI and preoperative complete right bundle branch block (cRBBB) or first-degree AVB, as well as a non-tubular left ventricular outflow tract (LVOT). In addition, valve implantation depth and angle of aortic root were independent predictors of new-onset cLBBB or high-grade AVB developing post-TAVR. The predictive value of valve implantation depth and angle of aortic root was further supported by receiver operating characteristic curve analysis results. CONCLUSIONS: In patients with PNAR undergoing TAVR using self-expanding valves, preoperative cRBBB or first-degree AVB and a non-tubular LVOT were indicators of a higher likelihood of PPI requirement. Moreover, deeper valve implantation depth and greater angle of aortic root may be independent risk factors for new-onset cLBBB or high-grade AVB post-TAVR. Valve implantation depth and angle of aortic root values may be used to predict the possibility of new cLBBB or high-grade AVB post-TAVR.

Cardiac ischemic preconditioning promotes cMyBP-C phosphorylation by inhibiting the calpain-mediated proteolysis.

Ischemic preconditioning (IPC) has been considered as the most important mean to protect against ischemia/reperfusion (I/R) induced heart injury. It has been reported that cardiac myosin binding protein-C (cMyBP-C) phosphorylation plays an essential role in cardiac protection against I/R-induced heart injury. However, it is still obscured whether IPC-mediated cardiac protection is causally related to cMyBP-C phosphorylation and proteolysis and, if so, what the underlying mechanism is. In this study, IPC was found to increase the phosphorylation level of cMyBP-C, companying with the decreased calpain activity in the collected perfusate samples. Mechanistically, we confirmed that IPC promoted cMyBP-C phosphorylation and inhibited calpain-mediated cMyBP-C proteolysis. Moreover, inhibition of calpain activity significantly increased the phosphorylated cMyBP-C level by using calpain inhibitor (MG-101), and subsequently promoted stabilization and secretion of cMyBP-C. Functionally, adeno-associated virus (AAV)-mediated overexpression of mutated phosphorylation motif site of cMyBP-C exhibited impaired IPC-mediated cardiac protection via proteolysis of the full-length cMyBP-C protein. We concluded that IPC promoted cMyBP-C phosphorylation via inhibition of calpain-mediated proteolysis and participated in IPC-mediated protection against I/R induced heart injury.

Measurements of the gravitational constant using two independent methods.

The Newtonian gravitational constant, G, is one of the most fundamental constants of nature, but we still do not have an accurate value for it. Despite two centuries of experimental effort, the value of G remains the least precisely known of the fundamental constants. A discrepancy of up to 0.05 per cent in recent determinations of G suggests that there may be undiscovered systematic errors in the various existing methods. One way to resolve this issue is to measure G using a number of methods that are unlikely to involve the same systematic effects. Here we report two independent determinations of G using torsion pendulum experiments with the time-of-swing method and the angular-acceleration-feedback method. We obtain G values of 6.674184 × 10-11 and 6.674484 × 10-11 cubic metres per kilogram per second squared, with relative standard uncertainties of 11.64 and 11.61 parts per million, respectively. These values have the smallest uncertainties reported until now, and both agree with the latest recommended value within two standard deviations.

Intramuscular Pulsed Radiofrequency Upregulates BNDF-TrKB Expression in the Spinal Cord in Rats as an Alternative Treatment for Complicated Pain.

Two cases of complicated pain exist: posterior screw fixation and myofascial pain. Intramuscular pulsed radiofrequency (PRF) may be an alternative treatment for such patients. This is a two-stage animal study. In the first stage, two muscle groups and two nerve groups were subdivided into a high-temperature group with PRF at 58 °C and a regular temperature with PRF at 42 °C in rats. In the second stage, two nerve injury groups were subdivided into nerve injury with PRF 42 °C on the sciatic nerve and muscle. Blood and spinal cord samples were collected. In the first stage, the immunohistochemical analysis showed that PRF upregulated brain-derived neurotrophic factor (BDNF) in the spinal cord in both groups of rats. In the second stage, the immunohistochemical analysis showed significant BDNF and tropomyosin receptor kinase B (TrkB) expression within the spinal cord after PRF in muscles and nerves after nerve injury. The blood biomarkers showed a significant increase in BDNF levels. PRF in the muscle in rats could upregulate BDNF-TrkB in the spinal cord, similar to PRF on the sciatica nerve for pain relief in rats. PRF could be considered clinically for patients with complicated pain and this study also demonstrated the role of BDNF in pain modulation. The optimal temperature for PRF was 42 °C.

LDL Receptor-Related Protein 1B Polymorphisms Associated with Increased Risk of Lymph Node Metastasis in Oral Cancer Group with Diabetes Mellitus.

Oral cancer ranks fourth among malignancies among Taiwanese men and is the eighth most common cancer among men worldwide in terms of general diagnosis. The purpose of the current study was to investigate how low-density lipoprotein receptor-related protein 1B (LDL receptor related protein 1B; LRP1B) gene polymorphisms affect oral squamous cell carcinoma (OSCC) risk and progression in individuals with diabetes mellitus (DM). Three LRP1B single-nucleotide polymorphisms (SNPs), including rs10496915, rs431809, and rs6742944, were evaluated in 311 OSCC cases and 300 controls. Between the case and control groups, we found no evidence of a significant correlation between the risk of OSCC and any of the three specific SNPs. Nevertheless, in evaluating the clinicopathological criteria, individuals with DM who possess a minimum of one minor allele of rs10496915 (AC + CC; p = 0.046) were significantly associated with tumor size compared with those with homozygous major alleles (AA). Similarly, compared to genotypes homologous for the main allele (GG), rs6742944 genotypes (GA + AA; p = 0.010) were more likely to develop lymph node metastases. The tongue and the rs6742944 genotypes (GA + AA) exhibited higher rates of advanced clinical stages (p = 0.024) and lymph node metastases (p = 0.007) when compared to homozygous alleles (GG). LRP1B genetic polymorphisms appear to be prognostic and diagnostic markers for OSCC and DM, as well as contributing to genetic profiling research for personalized medicine.

The change of state-trait anxiety among patients undergoing orthognathic surgery: A longitudinal study.

INTRODUCTION: This study aimed to assess state-trait anxiety level changes in Chinese patients with dentofacial discrepancies before and after orthognathic surgery and to explore the feasibility of developing a reference index for the preoperative screening of postoperative patients with high anxiety. METHODS: A total of 96 Chinese patients with dentofacial discrepancies who underwent orthognathic surgery were included in this study. Data were collected before orthognathic surgery and at 2 weeks (T2), 3 months, and 6 months (T4) after surgery using the State-Trait Anxiety Inventory. Receiver operating characteristic and linear regression analyses were performed to screen for preoperative indicators of postoperative high-state anxiety. RESULTS: State-trait anxiety levels in patients with dentofacial discrepancies decreased after surgery (F = 18.95, P <0.01; F = 6.90, P <0.01). Trait Anxiety Inventory can be used to screen patients with high-state anxiety from T2 to T4 (area under cover 95% confidence interval: T2, 0.74 [0.62-0.86]; 3 months, 0.79 [0.69-0.90]; T4, 0.77 [0.66-0.87], P <0.01), corresponding to cutoff values of 48.5, 46.5, and 45.5, respectively. CONCLUSIONS: All participants' state-trait anxiety levels improved after surgery compared with their preoperative levels. Preoperative trait anxiety levels can be used as a reference indicator to screen patients who may have high-state anxiety levels after orthognathic surgery. The creation of a screening scale will assist health care professionals to more pertinently help patients with high anxiety.

Protein-protein interaction site prediction by model ensembling with hybrid feature and self-attention.

BACKGROUND: Protein-protein interactions (PPIs) are crucial in various biological functions and cellular processes. Thus, many computational approaches have been proposed to predict PPI sites. Although significant progress has been made, these methods still have limitations in encoding the characteristics of each amino acid in sequences. Many feature extraction methods rely on the sliding window technique, which simply merges all the features of residues into a vector. The importance of some key residues may be weakened in the feature vector, leading to poor performance. RESULTS: We propose a novel sequence-based method for PPI sites prediction. The new network model, PPINet, contains multiple feature processing paths. For a residue, the PPINet extracts the features of the targeted residue and its context separately. These two types of features are processed by two paths in the network and combined to form a protein representation, where the two types of features are of relatively equal importance. The model ensembling technique is applied to make use of more features. The base models are trained with different features and then ensembled via stacking. In addition, a data balancing strategy is presented, by which our model can get significant improvement on highly unbalanced data. CONCLUSION: The proposed method is evaluated on a fused dataset constructed from Dset186, Dset_72, and PDBset_164, as well as the public Dset_448 dataset. Compared with current state-of-the-art methods, the performance of our method is better than the others. In the most important metrics, such as AUPRC and recall, it surpasses the second-best programmer on the latter dataset by 6.9% and 4.7%, respectively. We also demonstrated that the improvement is essentially due to using the ensemble model, especially, the hybrid feature. We share our code for reproducibility and future research at https://github.com/CandiceCong/StackingPPINet .

Molecular subtypes of breast cancer predicting clinical benefits of radiotherapy after breast-conserving surgery: a propensity-score-matched cohort study.

BACKGROUND: Based on the molecular expression of cancer cells, molecular subtypes of breast cancer have been applied to classify patients for predicting clinical outcomes and prognosis. However, further evidence is needed regarding the influence of molecular subtypes on the efficacy of radiotherapy (RT) after breast-conserving surgery (BCS), particularly in a population-based context. Hence, the present study employed a propensity-score-matched cohort design to investigate the potential role of molecular subtypes in stratifying patient outcomes for post-BCS RT and to identify the specific clinical benefits that may emerge. METHODS: From 2006 to 2019, the present study included 59,502 breast cancer patients who underwent BCS from the Taiwan National Health Insurance Research Database. Propensity scores were utilized to match confounding variables between patients with and without RT within each subtype of breast cancer, namely luminal A, luminal B/HER2-negative, luminal B/HER2-positive, basal-like, and HER2-enriched ones. Several clinical outcomes were assessed, in terms of local recurrence (LR), regional recurrence (RR), distant metastasis (DM), disease-free survival (DFS), and overall survival (OS). RESULTS: After post-BCS RT, patients with luminal A and luminal B/HER2-positive breast cancers exhibited a decrease in LR (adjusted hazard ratio [aHR] = 0.18, p < 0.0001; and, 0.24, p = 0.0049, respectively). Furthermore, reduced RR and improved DFS were observed in patients with luminal A (aHR = 0.15, p = 0.0004; and 0.29, p < 0.0001), luminal B/HER2-negative (aHR = 0.06, p = 0.0093; and, 0.46, p = 0.028), and luminal B/HER2-positive (aHR = 0.14, p = 0.01; and, 0.38, p < 0.0001) breast cancers. Notably, OS benefits were found in patients with luminal A (aHR = 0.62, p = 0.002), luminal B/HER2-negative (aHR = 0.30, p < 0.0001), basal-like (aHR = 0.40, p < 0.0001), and HER2-enriched (aHR = 0.50, p = 0.03), but not luminal B/HER2-positive diseases. Remarkably, when considering DM, luminal A patients who received RT demonstrated a lower cumulative incidence of DM than those without RT (p = 0.02). CONCLUSION: In patients with luminal A breast cancer who undergo BCS, RT could decrease the likelihood of tumor metastasis. After RT, the tumor's hormone receptor status may predict tumor control regarding LR, RR, and DFS. Besides, the HER2 status of luminal breast cancer patients may serve as an additional predictor of OS after post-BCS RT. However, further prospective studies are required to validate these findings.

Robotic Gastrectomy Versus Laparoscopic Gastrectomy for Gastric Cancer: A Multicenter Cohort Study of 5402 Patients in China.

OBJECTIVE: A large-scale multicenter retrospective cohort study was conducted to compare the short- and long-term outcomes of robotic gastrectomy (RG) and laparoscopic gastrectomy (LG) for gastric cancer. SUMMARY OF BACKGROUND DATA: RG is being increasingly used worldwide, but data from large-scale multicenter studies on the short- and long-term oncologic outcomes of RG versus LG are limited. The potential benefits of RG compared with LG for gastric cancer remain controversial. METHODS: Data from eligible patients who underwent RG or LG for gastric cancer of 11 experienced surgeons from 7 centers in China between March 2010 and October 2019 were collected. The RG group was matched 1:1 with the LG group by using propensity score matching. The primary outcome was postoperative complications. RESULTS: After propensity score matching, a well-balanced cohort of 3552 patients was included for further analysis. The occurrence of overall complications (12.6% vs 15.2%, P = 0.023) was lower in the RG group than in the LG group. RG was associated with less blood loss (126.8 vs 142.5 mL, P < 0.001) and more retrieved lymph nodes in total (32.5 vs 30.7, P < 0.001) and in suprapancreatic areas (13.3 vs 11.6, P < 0.001).The long-term oncological outcomes were comparable between the two groups. CONCLUSIONS: The results of this multicenter study demonstrate that RG is a safe and effective treatment for gastric cancer when performed by experienced surgeons, although longer operation time and higher costs are still concerns about RG. This study provides evidence suggesting that RG may represent an alternative surgical treatment to LG.