SC-AIR-BERT: a pre-trained single-cell model for predicting the antigen-binding specificity of the adaptive immune receptor.

Accurately predicting the antigen-binding specificity of adaptive immune receptors (AIRs), such as T-cell receptors (TCRs) and B-cell receptors (BCRs), is essential for discovering new immune therapies. However, the diversity of AIR chain sequences limits the accuracy of current prediction methods. This study introduces SC-AIR-BERT, a pre-trained model that learns comprehensive sequence representations of paired AIR chains to improve binding specificity prediction. SC-AIR-BERT first learns the 'language' of AIR sequences through self-supervised pre-training on a large cohort of paired AIR chains from multiple single-cell resources. The model is then fine-tuned with a multilayer perceptron head for binding specificity prediction, employing the K-mer strategy to enhance sequence representation learning. Extensive experiments demonstrate the superior AUC performance of SC-AIR-BERT compared with current methods for TCR- and BCR-binding specificity prediction.

HE2Gene: image-to-RNA translation via multi-task learning for spatial transcriptomics data.

MOTIVATION: Tissue context and molecular profiling are commonly used measures in understanding normal development and disease pathology. In recent years, the development of spatial molecular profiling technologies (e.g. spatial resolved transcriptomics) has enabled the exploration of quantitative links between tissue morphology and gene expression. However, these technologies remain expensive and time-consuming, with subsequent analyses necessitating high-throughput pathological annotations. On the other hand, existing computational tools are limited to predicting only a few dozen to several hundred genes, and the majority of the methods are designed for bulk RNA-seq. RESULTS: In this context, we propose HE2Gene, the first multi-task learning-based method capable of predicting tens of thousands of spot-level gene expressions along with pathological annotations from H&E-stained images. Experimental results demonstrate that HE2Gene is comparable to state-of-the-art methods and generalizes well on an external dataset without the need for re-training. Moreover, HE2Gene preserves the annotated spatial domains and has the potential to identify biomarkers. This capability facilitates cancer diagnosis and broadens its applicability to investigate gene-disease associations. AVAILABILITY AND IMPLEMENTATION: The source code and data information has been deposited at https://github.com/Microbiods/HE2Gene.

Harnessing natural product polysaccharides against lung cancer and revisit its novel mechanism.

The incidence and mortality of lung cancer are on the rise worldwide. However, the benefit of clinical treatment in lung cancer is limited. Owning to important sources of drug development, natural products have received constant attention around the world. Main ingredient polysaccharides in natural products have been found to have various activities in pharmacological research. In recent years, more and more scientists are looking for the effects and mechanisms of different natural product polysaccharides on lung cancer. In this review, we focus on the following aspects: First, natural product polysaccharides have been discovered to directly suppress the growth of lung cancer cells, which can be effective in limiting tumor progression. Additionally, polysaccharides have been considered to enhance immune function, which can play a pivotal role in fighting lung cancer. Lastly, polysaccharides can improve the efficacy of drugs in lung cancer treatment by regulating the gut microbiota. Overall, the research of natural product polysaccharides in the treatment of lung cancer is a promising area that has the potential to lead to new clinical treatments. With better understanding, natural product polysaccharides have the potential to become important components of future lung cancer treatments.

Human host status inference from temporal microbiome changes via recurrent neural networks.

With the rapid increase in sequencing data, human host status inference (e.g. healthy or sick) from microbiome data has become an important issue. Existing studies are mostly based on single-point microbiome composition, while it is rare that the host status is predicted from longitudinal microbiome data. However, single-point-based methods cannot capture the dynamic patterns between the temporal changes and host status. Therefore, it remains challenging to build good predictive models as well as scaling to different microbiome contexts. On the other hand, existing methods are mainly targeted for disease prediction and seldom investigate other host statuses. To fill the gap, we propose a comprehensive deep learning-based framework that utilizes longitudinal microbiome data as input to infer the human host status. Specifically, the framework is composed of specific data preparation strategies and a recurrent neural network tailored for longitudinal microbiome data. In experiments, we evaluated the proposed method on both semi-synthetic and real datasets based on different sequencing technologies and metagenomic contexts. The results indicate that our method achieves robust performance compared to other baseline and state-of-the-art classifiers and provides a significant reduction in prediction time.

Living Donor Intestinal Transplantation: Recipient Outcomes.

OBJECTIVE: To examine outcomes of living-donor intestinal transplant (LDITx) recipients. BACKGROUND: LDITx is not routinely performed because of surgical risks to the donor and the potential inferior physiologic performance of the segmental graft. However, data on the effectiveness of LDITx are scarce. DESIGN: This retrospective cohort study included patients undergoing LDITx between May 1999 and December 2021 in intestinal transplant programs in 2 university-affiliated hospitals in China. RESULTS: Actuarial survival rates were 80%, 72.7%, 66.7% for patient and 72.4%, 63.6%, 60% for graft at 1, 3, and 5 years, respectively. Recipients with >3/6 HLA-matched grafts had superior patient and graft survival rates than those with ≤3/6 HLA-matched grafts ( P <0.05). There were 12 deaths among the recipients, with infection being the leading cause (41.7%), followed by rejection (33.3%), surgical complications (16.7%), and others (8.3%). There were 16 graft losses among the recipients, with acute cellular rejection being the predominant cause (37.5%), followed by infection (25%), technical failure (12.5%), chronic rejection (12.5%), and others (12.5%). With an average follow-up of 3.7 (range, 0.6-23) years, the rates of acute and chronic rejection were 35% and 5%, and the rate of cytomegalovirus disease and post-transplant lymphoproliferative disease were 5% and 2.5%, respectively. Of the 40 patients, 28 (70%) are currently alive and have achieved enteral autonomy. CONCLUSIONS: LDITx is a valuable treatment option for patients with end-stage intestinal failure. Improved immunosuppression, better HLA matching, and shorter cold ischemia times were associated with reduced rates of rejection, viral-mediated infection and improved graft survival.

Noninvasive early diagnosis of intestinal diseases based on artificial intelligence in genomics and microbiome.

The maturing development in artificial intelligence (AI) and genomics has propelled the advances in intestinal diseases including intestinal cancer, inflammatory bowel disease (IBD), and irritable bowel syndrome (IBS). On the other hand, colorectal cancer is the second most deadly and the third most common type of cancer in the world according to GLOBOCAN 2020 data. The mechanisms behind IBD and IBS are still speculative. The conventional methods to identify colorectal cancer, IBD, and IBS are based on endoscopy or colonoscopy to identify lesions. However, it is invasive, demanding, and time-consuming for early-stage intestinal diseases. To address those problems, new strategies based on blood and/or human microbiome in gut, colon, or even feces were developed; those methods took advantage of high-throughput sequencing and machine learning approaches. In this review, we summarize the recent research and methods to diagnose intestinal diseases with machine learning technologies based on cell-free DNA and microbiome data generated by amplicon sequencing or whole-genome sequencing. Those methods play an important role in not only intestinal disease diagnosis but also therapy development in the near future.

Circ_DOCK1 regulates USP11 through miR-132-3p to control colorectal cancer progression.

BACKGROUND: Circular RNAs (circRNAs) take part in colorectal cancer malignancies. CircRNA dedicator of cytokinesis 1 (circ_DOCK1) is involved in colorectal cancer progression, but the mechanism underlying this circRNA that takes part in colorectal cancer development remains largely undetermined. METHODS: Tumor and normal para-cancerous tissues were collected from 42 colorectal cancer patients. Human colorectal cancer cell lines (HCT116 and SW480) were used for the experiments in vitro. Circ_DOCK1, microRNA (miR)-132-3p, and ubiquitin-specific protease 11 (USP11) levels were measured through quantitative real-time polymerase chain reaction and Western blotting. Cell growth, metastasis, and apoptosis were investigated via colony formation, 5-ethynyl-2'-deoxyuridine (EdU) staining, MTT, flow cytometry, Western blotting, and transwell analyses. The target association was evaluated via dual-luciferase reporter analysis, RNA pull-down, and immunoprecipitation (RIP). Xenograft assay was performed using HCT116 cells. USP11 and Ki67 levels in tumor tissues were detected via immunohistochemistry. RESULTS: Circ_DOCK1 expression was enhanced in colorectal cancer tissues and cells. Silencing circ_DOCK1 repressed cell growth, migration, and invasion, and facilitated apoptosis. Circ_DOCK1 sponged miR-132-3p, and miR-132-3p silence mitigated the effect of circ_DOCK1 interference on cell growth, metastasis, and apoptosis. MiR-132-3p targeted USP11, and circ_DOCK1 could regulate USP11 level by miR-132-3p. MiR-132-3p suppressed cell growth, metastasis, and apoptosis, and USP11 attenuated these effects. Knockdown of circ_DOCK1 decreased colorectal cancer cell xenograft tumor growth. CONCLUSION: Circ_DOCK1 interference suppressed cell growth and metastasis, and increased apoptosis of colorectal cancer via decreasing USP11 by increasing miR-132-3p.

Pseudomonas syringae dual-function protein Lon switches between virulence and metabolism by acting as both DNA-binding transcriptional regulator and protease in different environments.

Lon, a member of the AAA+ protease family, plays vital roles in Type III secretion systems (T3SS), agglutination and colony shape in the model plant pathogen Pseudomonas syringae. Lon also functions as a transcriptional regulator in other bacterial species such as Escherichia coli and Brevibacillus thermoruber. To reveal the molecular mechanisms of Lon as a dual-function protein in P. syringae, we studied Lon-regulated genes by using RNA sequencing (RNA-seq), chromatin immunoprecipitation sequencing (ChIP-seq) and liquid chromatography-tandem mass spectrometry. As a transcriptional regulator, Lon directly regulated a group of genes (PSPPH_4788, gacA, fur, gntR, clpS, lon and glyA) and consequently regulated their functions, such as 1-dodecanol oxidation activity, motility, pyoverdine production, glucokinase activity, N-end rule pathway, lon expression and serine hydroxymethyltransferase activity. Mass spectrometry results revealed that the expression levels of five T3SS proteins (such as HrcV, HrpW1) were higher in the ∆lon strain than the wild-type (WT) strain in KB. In MM, 12 metabolic proteins (such as AcdS and NuoI) showed lower levels in the ∆lon strain than the WT strain. Taken together, these data demonstrate that the dual-function protein Lon sophisticatedly regulates virulence and metabolism in P. syringae.

Pleiotropic Effects of c-di-GMP Content in Pseudomonas syringae.

Although the ubiquitous bacterial secondary messenger cyclic diguanylate (c-di-GMP) has important cellular functions in a wide range of bacteria, its function in the model plant pathogen Pseudomonas syringae remains largely elusive. To this end, we overexpressed Escherichia coli diguanylate cyclase (YedQ) and phosphodiesterase (YhjH) in P. syringae, resulting in high and low in vivo levels of c-di-GMP, respectively. Via genome-wide RNA sequencing of these two strains, we found that c-di-GMP regulates (i) fliN, fliE, and flhA, which are associated with flagellar assembly; (ii) alg8 and alg44, which are related to the exopolysaccharide biosynthesis pathway; (iii) pvdE, pvdP, and pvsA, which are associated with the siderophore biosynthesis pathway; and (iv) sodA, which encodes a superoxide dismutase. In particular, we identified three promoters that are sensitive to elevated levels of c-di-GMP and inserted them into luciferase-based reporters that respond effectively to the c-di-GMP levels in P. syringae; these promoters could be useful in the measurement of in vivo levels of c-di-GMP in real time. Further phenotypic assays validated the RNA sequencing (RNA-seq) results and confirmed the effect on c-di-GMP-associated pathways, such as repressing the type III secretion system (T3SS) and motility while inducing biofilm production, siderophore production, and oxidative stress resistance. Taken together, these results demonstrate that c-di-GMP regulates the virulence and stress response in P. syringae, which suggests that tuning its level could be a new strategy to protect plants from attacks by this pathogen.IMPORTANCE The present work comprehensively analyzed the transcriptome and phenotypes that were regulated by c-di-GMP in P. syringae Given that the majority of diguanylate cyclases and phosphodiesterases have not been characterized in P. syringae, this work provided a very useful database for the future study on regulatory mechanism (especially its relationship with T3SS) of c-di-GMP in P. syringae In particular, we identified three promoters that were sensitive to elevated c-di-GMP levels and inserted them into luciferase-based reporters that effectively respond to intracellular levels of c-di-GMP in P. syringae, which could be used as an economic and efficient way to measure relative c-di-GMP levels in vivo in the future.

Arabidopsis VAC14 Is Critical for Pollen Development through Mediating Vacuolar Organization.

Pollen viability depends on dynamic vacuolar changes during pollen development involving increases and decreases of vacuolar volume through water and osmolite accumulation and vacuolar fission. Mutations in FAB1A to FAB1D, the genes encoding phosphatidylinositol 3,5-bisphosphate [PI(3,5)P2]-converting kinases, are male gametophyte lethal in Arabidopsis (Arabidopsis thaliana) due to defective vacuolar fission after pollen mitosis I, suggesting a key role of the phospholipid in dynamic vacuolar organization. However, other genetic components that regulate the production of PI(3,5)P2 and its involvement in pollen germination and tube growth are unknown. Here, we identified and characterized Arabidopsis VAC14, a homolog of the yeast and metazoan VAC14s that are crucial for the production of PI(3,5)P2VAC14 is constitutively expressed and highly present in developing pollen. Loss of function of VAC14 was male gametophyte lethal due to defective pollen development. Ultrastructural studies showed that vacuolar fission after pollen mitosis I was compromised in vac14 mutant microspores, which led to pollen abortion. We further showed that inhibiting the production of PI(3,5)P2 or exogenous application of PI(3,5)P2 mimicked or rescued the pollen developmental defect of the vac14 mutant, respectively. Genetic interference and pharmacological approaches suggested a role of PI(3,5)P2 in pollen germination and tube growth. Our results provide insights into the function of VAC14 and, by inference, that of PI(3,5)P2 in plant cells.

Multilabel Feature Selection via Shared Latent Sublabel Structure and Simultaneous Orthogonal Basis Clustering.

Multilabel feature selection solves the dimension distress of high-dimensional multilabel data by selecting the optimal subset of features. Noisy and incomplete labels of raw multilabel data hinder the acquisition of label-guided information. In existing approaches, mapping the label space to a low-dimensional latent space by semantic decomposition to mitigate label noise is considered an effective strategy. However, the decomposed latent label space contains redundant label information, which misleads the capture of potential label relevance. To eliminate the effect of redundant information on the extraction of latent label correlations, a novel method named SLOFS via shared latent sublabel structure and simultaneous orthogonal basis clustering for multilabel feature selection is proposed. First, a latent orthogonal base structure shared (LOBSS) term is engineered to guide the construction of a redundancy-free latent sublabel space via the separated latent clustering center structure. The LOBSS term simultaneously retains latent sublabel information and latent clustering center structure. Moreover, the structure and relevance information of nonredundant latent sublabels are fully explored. The introduction of graph regularization ensures structural consistency in the data space and latent sublabels, thus helping the feature selection process. SLOFS employs a dynamic sublabel graph to obtain a high-quality sublabel space and uses regularization to constrain label correlations on dynamic sublabel projections. Finally, an effective convergence provable optimization scheme is proposed to solve the SLOFS method. The experimental studies on the 18 datasets demonstrate that the presented method performs consistently better than previous feature selection methods.

Identification of mIDH1 R132C/S280F Inhibitors from Natural Products by Integrated Molecular Docking, Pharmacophore Modeling and Molecular Dynamics Simulations.

Mutant isocitrate dehydrogenase 1 (mIDH1) is a common driving factor in acute myeloid leukemia (AML), with the R132 mutation accounting for a high proportion. The U.S. Food and Drug Administration (FDA) approved Ivosidenib, a molecular entity that targets IDH1 with R132 mutations, as a promising therapeutic option for AML with mIDH1 in 2018. It was of concern that the occurrence of disease resistance or recurrence, attributed to the IDH1 R132C/S280F second site mutation, was observed in certain patients treated with Ivosidenib within the same year. Furthermore, it should be noted that most mIDH1 inhibitors demonstrated limited efficacy against mutations at this specific site. Therefore, there is an urgent need to investigate novel inhibitors targeting mIDH1 for combating resistance caused by IDH1 R132C/S280F mutations in AML. This study aimed to identify novel mIDH1 R132C/S280F inhibitors through an integrated strategy of combining virtual screening and dynamics simulations. First, 2000 hits were obtained through structure-based virtual screening of the COCONUT database, and hits with better scores than -10.67 kcal/mol were obtained through molecular docking. A total of 12 potential small molecule inhibitors were identified through pharmacophore modeling screening and Prime MM-GBSA. Dynamics simulations were used to study the binding modes between the positive drug and the first three hits and IDH1 carrying the R132C/S280F mutation. RMSD showed that the four dynamics simulation systems remained stable, and RMSF and Rg showed that the screened molecules have similar local flexibility and tightness to the positive drug. Finally, the lowest energy conformation, hydrogen bond analysis, and free energy decomposition results indicate that in the entire system the key residues LEU120, TRP124, TRP267, and VAL281 mainly contribute van der Waals forces to the interaction, while the key residues VAL276 and CYS379 mainly contribute electrostatic forces.

Autonomous closed-loop mechanistic investigation of molecular electrochemistry via automation.

Electrochemical research often requires stringent combinations of experimental parameters that are demanding to manually locate. Recent advances in automated instrumentation and machine-learning algorithms unlock the possibility for accelerated studies of electrochemical fundamentals via high-throughput, online decision-making. Here we report an autonomous electrochemical platform that implements an adaptive, closed-loop workflow for mechanistic investigation of molecular electrochemistry. As a proof-of-concept, this platform autonomously identifies and investigates an EC mechanism, an interfacial electron transfer (E step) followed by a solution reaction (C step), for cobalt tetraphenylporphyrin exposed to a library of organohalide electrophiles. The generally applicable workflow accurately discerns the EC mechanism's presence amid negative controls and outliers, adaptively designs desired experimental conditions, and quantitatively extracts kinetic information of the C step spanning over 7 orders of magnitude, from which mechanistic insights into oxidative addition pathways are gained. This work opens opportunities for autonomous mechanistic discoveries in self-driving electrochemistry laboratories without manual intervention.

Government subsidies and innovation in new energy vehicle companies: An empirical study of new energy vehicle listed companies based on Shanghai and Shenzhen A-shares.

The panel data of 50 new energy vehicle enterprises in Shanghai and Shenzhen A-shares from 2012 to 2021 are selected to empirically analyze the impact of government subsidies on the innovation of new energy vehicle enterprises and to further discuss the differences between such an impact in different forms and regions. The study finds that, first, government subsidies have a certain promotion effect on the innovation of new energy vehicle enterprises, and an inverted U-shaped relationship exists thereof. Second, at the enterprise level, government subsidies have a significant effect on the innovation of non-state enterprises, downstream vehicle enterprises, and enterprises with lower establishment years, and the inverted-U trend is evident. Third, at the regional level, government subsidies have a more significant effect on the innovation of enterprises in non-eastern regions and low-environmental regulation regions, and the inverted-U-shaped relationship trend is more apparent. The study establishes the nonlinear relationship between government subsidies and the innovation of new energy vehicle enterprises through empirical research, which expands the theory of enterprise innovation and has a certain guiding significance for improving the innovation capability of new energy vehicle enterprises in the future.

Local Community Detection Algorithm Based on Alternating Strategy of Strong Fusion and Weak Fusion.

Existing fusion-based local community detection algorithms have achieved good results. However, when assigning a node to a community, similarity functions are sometimes used, which only use node information, while ignoring connection information within the community. These algorithms sometimes fail to find influential nodes, which eventually leads to the failure to find a complete local community. To address these problems, a new local community detection algorithm is proposed in this article. Two strategies, of strong fusion followed by weak fusion, are used alternately to fuse nodes. Compared with using two fusion strategies alone, the alternating loop method can improve the solution of the algorithm in each stage. In strong fusion, we propose a new membership function that considers both node information and connection information in the local community. This improves the quality of the fused node while preserving the structure of the current community. In weak fusion, we propose a parameter-based similarity measure, which can detect influential nodes for a local community. We also propose a local community evaluation metric, which does not require true division to determine the optimal local community under different parameters. Experiments, compared to six state-of-the-art algorithms, show that the proposed algorithm improves accuracy and stability, and also demonstrate the effectiveness of the new local community evaluation metrics in parameter selection.

[Di-2-ethylhexyl phthalate and its metabolite single-ethylhexyl phthalate affect TGF-beta 1 expression and telomerase activity in the testis of young male rats].

OBJECTIVE: To investigate the influences of di-2-ethylhexyl phthalate (DEHP) and its metabolite single-ethylhexyl phthalate (MEHP) on the expression of transforming growth factor-beta 1 (TGF-beta1) and telomerase activity in young male Wistar METHODS: Ninety-six 2-week-old male Wistar rats were equally randomized into a normal control (NC) group, a positive control (PC) group, and six experimental groups. Those of the NC group were intragastrically administered 0.9% normal saline at a dose of 0.2 ml per kg per d for 3 weeks, those in the PC group cyclophosphamide (CTX) at 100 mg per kg per d for 1 week, and those of the experimental groups DEHP and MEHP, respectively, at a low dose (100 mg per kg per d) for 3 weeks, a moderate dose (200 mg per kg per d) for 2 weeks, and a high dose (300 mg per kg per d) for 1 week. Then we observed the morphological changes of the testicular sperm and counted the sperm heads and their abnormity rate at different doses and times. We detected the expression of TGF-beta1 in the testis tissue using immunohistochemical SABC and RT-PCR, measured the area density, and determined telomerase activity by ELISA. RESULTS: Compared with the NC group, the experimental groups showed an obvious reduction in the total sperm count and number of sperm heads (P < 0.05) and a significant increase in the rate of teratosperm (P < 0.05), such as decapitated, hookless, and double-tailed sperm. And there were no significant differences between the high-dose short-term and low-dose long-term medication groups (P > 0.05). The expression of TGF-beta1 was low in the NC group, high in the PC group, and obviously increased in the membrane and cytoplasm of spermatogenic cells of the experimental groups. The area density and TGF-beta1 mRNA expression were 0.156 0 +/- 0.003 5 and 1.51 +/- 0.20 in the NC group, 0.534 0 +/- 0.003 1 and 8.43 +/- 1.75 in the PC group, 0.289 0 +/- 0.003 6 and 3.83 +/- 1.57 in the DEHP groups, and 0.284 0 +/- 0.003 1 and 3.51 +/- 1.41 in the MEHP groups. There were significant differences between the experimental and the other two groups (P < 0.01), but not between the high-dose short-term and low-dose long-term medication groups (P > 0.05). Telomerase activity was remarkably reduced in the experimental groups as compared with the NC group (P < 0.05), but with no significant difference between the high-dose short-term and low-dose long-term medication groups (P > 0.05). CONCLUSION: DEHP and its metabolite MEHP can evidently induce spermatogenic injury in young male rats, which may be associated with their induction of increased TGF-beta1 expression and decreased telomerase activity in the rat testis.

Dynamic Immunization Node Model for Complex Networks Based on Community Structure and Threshold.

In the information age of big data, and increasingly large and complex networks, there is a growing challenge of understanding how best to restrain the spread of harmful information, for example, a computer virus. Establishing models of propagation and node immunity are important parts of this problem. In this article, a dynamic node immune model, based on the community structure and threshold (NICT), is proposed. First, a network model is established, which regards nodes carrying harmful information as new nodes in the network. The method of establishing the edge between the new node and the original node can be changed according to the needs of different networks. The propagation probability between nodes is determined by using community structure information and a similarity function between nodes. Second, an improved immune gain, based on the propagation probability of the community structure and node similarity, is proposed. The improved immune gain value is calculated for neighbors of the infected node at each time step, and the node is immunized according to the hand-coded parameter: immune threshold. This can effectively prevent invalid or insufficient immunization at each time step. Finally, an evaluation index, considering both the number of immune nodes and the number of infected nodes at each time step, is proposed. The immune effect of nodes can be evaluated more effectively. The results of network immunization experiments, on eight real networks, suggest that the proposed method can deliver better network immunization than several other well-known methods from the literature.

Diagnosis of an extremely rare case of malignant adenomyoepithelioma in pleomorphic adenoma: A case report.

BACKGROUND: Pleomorphic adenoma (PA) is the most common type of salivary gland tumor, and its common sites are parotid gland, sinus, nasal septum and cleft palate. PA is an uncommon benign type of tumor occurring in the breast, and there are few reports of cases in Asia. CASE SUMMARY: An 84-year-old woman found a mass in the upper outer quadrant of the right breast > 1 year ago. The patient underwent a right breast lumpectomy and sentinel lymph node biopsy. The pathological diagnosis was PA in the upper outer quadrant of the right breast, and the malignant component was malignant adenomyoepithelioma. The postoperative course was uneventful, and no chemotherapy was administered. At 18 mo of follow-up, the patient is alive and well, with no evidence of recurrent disease. CONCLUSION: Patients with breast PA should first undergo extended excision of breast masses followed by pathological examination. If malignancy is confirmed or the surgical margin is positive, modified radical mastectomy should be performed.

SRG-Vote: Predicting Mirna-Gene Relationships via Embedding and LSTM Ensemble.

Targeted therapy for one for a set of genes has made it possible to apply precision medicine for different patients due to the existence of tumor heterogeneity. However, how to regulate those genes are still problematic. One of the natural regulators of genes is microRNAs. Thus, a better understanding of the miRNA-gene interaction mechanism might contribute to future diagnosis, prevention, and cancer therapy. The interactions between microRNA and genes play an essential role in molecular genetics. The in-vivo experiments validating the relationships between them are time-consuming, money-costly, and labor-intensive. With the development of high-throughput technology, we dealt with tons of biological data. However, extracting features from tremendous raw data and making a mathematical model is still a challenging topic. Machine learning and deep learning algorithms have become powerful tools in dealing with biological data. Inspired by this, in this paper, we propose a model that combines features/embedding extraction methods, deep learning algorithms, and a voting system. We leverage doc2vec to generate sequential embedding from molecular sequences. The role2vec, GCN, and GMM for geometrical embedding were generated from the complex network from similarity and pair-wise datasets. For the deep learning algorithms, we leveraged LSTM and Bi-LSTM according to different embedding and features. Finally, we adopted a voting system to balance results from different data sources. The results have shown that our voting system could achieve a higher AUC than the existing benchmark. The case studies demonstrate that our model could reveal potential relationships between miRNAs and genes. The source code, features, and predictive results can be downloaded at https://github.com/Xshelton/SRG-vote.

Human disease prediction from microbiome data by multiple feature fusion and deep learning.

Human disease prediction from microbiome data has broad implications in metagenomics. It is rare for the existing methods to consider abundance profiles from both known and unknown microbial organisms, or capture the taxonomic relationships among microbial taxa, leading to significant information loss. On the other hand, deep learning has shown unprecedented advantages in classification tasks for its feature-learning ability. However, it encounters the opposite situation in metagenome-based disease prediction since high-dimensional low-sample-size metagenomic datasets can lead to severe overfitting; and black-box model fails in providing biological explanations. To circumvent the related problems, we developed MetaDR, a comprehensive machine learning-based framework that integrates various information and deep learning to predict human diseases. Experimental results indicate that MetaDR achieves competitive prediction performance with a reduction in running time, and effectively discovers the informative features with biological insights.