MolMVC: Enhancing molecular representations for drug-related tasks through multi-view contrastive learning.

MOTIVATION: Effective molecular representation is critical in drug development. The complex nature of molecules demands comprehensive multi-view representations, considering 1D, 2D, and 3D aspects, to capture diverse perspectives. Obtaining representations that encompass these varied structures is crucial for a holistic understanding of molecules in drug-related contexts. RESULTS: In this study, we introduce an innovative multi-view contrastive learning framework for molecular representation, denoted as MolMVC. Initially, we use a Transformer encoder to capture 1D sequence information and a Graph Transformer to encode the intricate 2D and 3D structural details of molecules. Our approach incorporates a novel attention-guided augmentation scheme, leveraging prior knowledge to create positive samples tailored to different molecular data views. To align multi-view molecular positive samples effectively in latent space, we introduce an adaptive multi-view contrastive loss (AMCLoss). In particular, we calculate AMCLoss at various levels within the model to effectively capture the hierarchical nature of the molecular information. Eventually, we pre-train the encoders via minimizing AMCLoss to obtain the molecular representation, which can be used for various down-stream tasks. In our experiments, we evaluate the performance of our MolMVC on multiple tasks, including molecular property prediction (MPP), drug-target binding affinity (DTA) prediction and cancer drug response (CDR) prediction. The results demonstrate that the molecular representation learned by our MolMVC can enhance the predictive accuracy on these tasks and also reduce the computational costs. Furthermore, we showcase MolMVC's efficacy in drug repositioning across a spectrum of drug-related applications. AVAILABILITY AND IMPLEMENTATION: The code and pre-trained model are publicly available at https://github.com/Hhhzj-7/MolMVC.

Incorporating clinician insight and care plans into an audit and feedback initiative for antipsychotic prescribing to Medicaid-enrolled youth in Philadelphia.

BACKGROUND: Audit and feedback (A/F), which include initiatives like report cards, have an inconsistent impact on clinicians' prescribing behavior. This may be attributable to their focus on aggregate prescribing measures, a one-size-fits-all approach, and the fact that A/F initiatives rarely engage with the clinicians they target. METHODS: In this study, we describe the development and delivery of a report card that summarized antipsychotic prescribing to publicly-insured youth in Philadelphia, which was introduced by a Medicaid managed care organization in 2020. In addition to measuring aggregate prescribing behavior, the report card included different elements of care plans, including whether youth were receiving polypharmacy, proper medication management, and the concurrent use of behavioral health outpatient services. The A/F initiative elicited feedback from clinicians, which we refer to as an "audit and feedback loop." We also evaluate the impact of the report card by comparing pre-post differences in prescribing measures for clinicians who received the report card with a group of clinicians who did not receive the report card. RESULTS: Report cards indicated that many youth who were prescribed antipsychotics were not receiving proper medication management or using behavioral health outpatient services alongside the antipsychotic prescription, but that polypharmacy was rare. In their feedback, clinicians who received report cards cited several challenges related to antipsychotic prescribing, such as the logistical difficulties of entering lab orders and family members' hesitancy to change care plans. The impact of the report card was mixed: there was a modest reduction in the share of youth receiving polypharmacy following the receipt of the report card, while other measures did not change. However, we documented a large reduction in the number of youth with one or more antipsychotic prescription fill among clinicians who received a report card. CONCLUSIONS: A/F initiatives are a common approach to improving the quality of care, and often target specific practices such as antipsychotic prescribing. Report cards are a low-cost and feasible intervention but there is room for quality improvement, such as adding measures that track medication management or eliciting feedback from clinicians who receive report cards. To ensure that the benefits of antipsychotic prescribing outweigh its risks, it is important to promote quality and safety of antipsychotic prescribing within a broader care plan.

Exploration of the Polymorphism Distribution of Bovine HMGA2 Gene in Worldwide Breeds and Its Associations with Ovarian Traits.

The high-mobility group AT-hook 2(HMGA2) gene has been widely studied in the context of cancer and animal growth. However, recently, several studies have uncovered its critical role in cell proliferation. A genome-wide association study (GWAS) further suggests that the HMGA2 gene is a candidate gene in fertility, indicating its connection not only to growth traits but also to reproduction, specifically ovarian traits. Thus, this study aimed to analyze the distribution of the HMGA2 gene in 54 bovine breeds worldwide, identify important short fragment variants (indels), and investigate the relationship between HMGA2 and ovarian development. The dataset included genotypic information from a bovine population of 634 individuals (n = 634). After genotyping and analyzing four selected loci, we found that one out of four loci, rs133750033 (P4-D22-bp), was polymorphic. Our results also reveal that this indel of HMGA2 is significantly associated with certain ovarian traits (p < 0.05). Specifically, it has connection with ovarian length (p = 0.004) and ovarian height (p = 0.026) during diestrus. Additionally, we discovered a higher expression of the HMGA2 gene in Asian cattle breeds. In summary, this study suggests that HMGA2 has the potential to serve as an animal fertility testing marker gene. Moreover, these findings contribute to a more promising outlook for the bovine industry.

Ligand recognition and G-protein coupling of trace amine receptor TAAR1.

Trace-amine-associated receptors (TAARs), a group of biogenic amine receptors, have essential roles in neurological and metabolic homeostasis1. They recognize diverse endogenous trace amines and subsequently activate a range of G-protein-subtype signalling pathways2,3. Notably, TAAR1 has emerged as a promising therapeutic target for treating psychiatric disorders4,5. However, the molecular mechanisms underlying its ability to recognize different ligands remain largely unclear. Here we present nine cryo-electron microscopy structures, with eight showing human and mouse TAAR1 in a complex with an array of ligands, including the endogenous 3-iodothyronamine, two antipsychotic agents, the psychoactive drug amphetamine and two identified catecholamine agonists, and one showing 5-HT1AR in a complex with an antipsychotic agent. These structures reveal a rigid consensus binding motif in TAAR1 that binds to endogenous trace amine stimuli and two extended binding pockets that accommodate diverse chemotypes. Combined with mutational analysis, functional assays and molecular dynamic simulations, we elucidate the structural basis of drug polypharmacology and identify the species-specific differences between human and mouse TAAR1. Our study provides insights into the mechanism of ligand recognition and G-protein selectivity by TAAR1, which may help in the discovery of ligands or therapeutic strategies for neurological and metabolic disorders.

LINC00426, a novel m6A-regulated long non-coding RNA, induces EMT in cervical cancer by binding to ZEB1.

PURPOSE: To explore the function and molecular mechanism of LINC00426 in Cervical Cancer (CC), and to explore the clinical treatment strategy of LINC00426 for CC. METHODS: Bioinformatics analysis was used to explore the expression of LINC00426 and patient prognosis of CC. Cell function experiments were conducted to explore the potential effect of LINC00426 on CC malignant phenotypes. The difference in m6A modification level between the high and low expression groups of LINC00426 was analyzed by detecting the total m6A level. The luciferase reporter assay was used to confirm the binding of miR-200a-3p to LINC00426. The RIP assay was used to confirm the binding of LINC00426 to ZEB1. Cell viability assay was performed to detect the effect of LINC00426 on cellular drug resistance. RESULTS: LINC00426 is up-regulated in CC, which can enhance the proliferation, migration and invasion of CC cells. METTL3 promotes the expression of LINC00426 by m6A methylation modification. In addition, the LINC00426/miR-200a-3p/ZEB1 axis affects the proliferation, migration, and invasion of CC by regulating the expression of EMT markers. Through the detection of cell viability, we observed that overexpression LINC00426 in cells resulted in resistance to cisplatin and bleomycin, and more sensitive to imatinib. CONCLUSION: LINC00426 is a cancer-promoting lncRNA related to m6A modification. The process of EMT in CC is regulated by the LINC00426/miR-200a/3p/ZEB1 axis. LINC00426 can affect the sensitivity of CC cells to chemotherapy drugs, and is expected to become a therapeutic target for CC.

Post-transcriptional regulation of tumor suppressor gene lncRNA CARMN via m6A modification and miRNA regulation in cervical cancer.

PURPOSE: The abnormal regulation of lncRNA CARMN has been proved to be a tumor suppressor gene of cervical cancer (CC). However, its role in CC is still elusive. The regulation of CARMN post-transcriptional level by m6A modification and miRNA has not been studied. This study aims to analyze the molecular mechanism of m6A modification and miRNA on the abnormal expression of CARMN in CC cells, so as to provide a new theoretical basis for the diagnosis and treatment of CC. METHODS: MeRIP-seq was used to identify the differential m6A-modified genes between tumor and normal cervical tissues. RT-qPCR assay was used to detect gene expression levels in tissues or cells. The m6A modification sites of CARMN was predicted by bioinformatics, and the modification of m6A and its regulatory effect on CARMN were analyzed by MeRIP-qPCR, Actinomycin D assay and RIP assay. RIP-microarray combined with bioinformatics methods to screen miRNAs that may target CARMN. The regulation mechanism between miRNA and CARMN was verified by RT-qPCR, nucleo-plasmic separation assay, mRNA stability assay, dual-luciferase reporter assay, and in vivo experiments. RESULTS: MeRIP-seq found that CARMN is a significant different gene in the abundance of m6A in CC, and the modification level of m6A in CC tissues was higher than that in normal cervical tissues. Further, this study verified that m6A reader YTHDF2 could recognize m6A-modified CARMN and promote its degradation in CC cells. miR-21-5p was proved to be the downstream target gene of CARMN, and miR-21-5p could negatively regulate the expression of CARMN. Further experiments showed that miR-21-5p could directly bind to CARMN and lead to the degradation of CARMN. The in vivo experimental results indicated that the level of miR-21-5p in the overexpressed CARMN group was significantly lower than that in the control group. CONCLUSION: m6A modification and miR-21-5p play important roles in promoting the occurrence and development of tumors by regulating CARMN, provide new potential targets for the treatment of CC.

Immune-Related Genes' Prognostic, Therapeutic and Diagnostic Value in Ovarian Cancer Immune-Related Gene Biomarker in Ovarian Cancer.

OBJECTIVES: The abnormal expression of immune-related genes (IRGs) plays an important role in the occurrence and progression of ovarian cancer (OC), which is the main cause of mortality among gynecological cancer patients. This study aims to establish a prognostic risk model and comprehensively analyze the relationship between OC risk score and prognosis, immune cell infiltration (ICI) and therapeutic sensitivity in OC. METHODS: We retrospectively evaluated the clinicopathological characteristics of consecutive OC patients in the Cancer Genome Atlas (TCGA) database. First, the prognostic risk model was constructed by bioinformatics methods. And then, we systematically assessed model robustness, and correlations between risk score and prognosis, and immune cell infiltration. The ICGC cohort was used to verify the prognostic risk model. Finally, we evaluated their value in the treatment of OC immunotherapy and chemotherapy. RESULTS: A total of 10 IRGs were identified to construct the prognostic risk model. Survival analysis revealed that patients in the low-risk group had a better prognosis (P < .01), and the risk score might be considered an independent predictor for predicting the prognosis. In addition, risk scores and patient clinical information were used to construct clinical nomograms, improving the prediction's precision. We also explored the relationship between the risk score and ICI, immunotherapy and drug sensitivity. CONCLUSIONS: Collectively, we identified a novel ten IRGs signature that may be applied as a prognostic predictor of OC, thereby benefiting clinical decision-making and personalized treatment of patients.

Lacticaseibacillus rhamnosus Probio-M9 enhanced the antitumor response to anti-PD-1 therapy by modulating intestinal metabolites.

BACKGROUND: Probiotics have been increasingly proposed for enhancing immune checkpoint blockade (ICB) treatments against cancer. However, its causal relationship with immunotherapeutic efficacy remains unclear, which promoted us to explore if and how probiotic Lacticaseibacillus rhamnosus Probio-M9 manipulates gut microbiome for expected outcomes. METHODS: We evaluated the effects of Probio-M9 on the anti-PD-1 treatment against colorectal cancer in mice via a multi-omics approach. We defined the mechanisms of Probio-M9-mediated antitumor immunity by comprehensive analyses of metagenome and metabolites of commensal gut microbes as well as the immunologic factors and serum metabolome of the host. FINDINGS: The results indicated that Probio-M9 intervention strengthened the anti-PD-1-based tumor inhibition. Both prophylactic and therapeutic administration of Probio-M9 showed conspicuous performance in controlling tumor growth with ICB treatment. The supplement of Probio-M9 modulated enhanced immunotherapy response through promoting beneficial microbes (e.g., Lactobacillus and Bifidobacterium animalis), producing beneficial metabolites including butyric acids in the gut, and accumulating blood-derived α-ketoglutaric acid, N-acetyl-l-glutamic acid and pyridoxine in particular, which promoted the infiltration and activation of cytotoxic T lymphocytes (CTLs) and suppressing the function of regulatory T cells (Tregs) in the tumor microenvironment (TME). Subsequently, we found that enhanced immunotherapeutic response was transmissible by transplanting either post-probiotic-treatment gut microbes or intestinal metabolites to new tumor-bearing mice. INTERPRETATION: This study offered valuable insight into the causal role of Probio-M9 in correcting the defects in gut microbiota that compromised anti-PD-1 therapeutic efficacy, which can be used as an alternative synergetic agent with ICB for clinical cancer treatment. FUNDING: This research was supported by Research Fund for the National Key R&D Program of China (2022YFD2100702), Inner Mongolia Science and Technology Major Projects (2021ZD0014), and China Agriculture Research System of MOF and MARA.

XGBCDA: a multiple heterogeneous networks-based method for predicting circRNA-disease associations.

BACKGROUND: Biological experiments have demonstrated that circRNA plays an essential role in various biological processes and human diseases. However, it is time-consuming and costly to merely conduct biological experiments to detect the association between circRNA and diseases. Accordingly, developing an efficient computational model to predict circRNA-disease associations is urgent. METHODS: In this research, we propose a multiple heterogeneous networks-based method, named XGBCDA, to predict circRNA-disease associations. The method first extracts original features, namely statistical features and graph theory features, from integrated circRNA similarity network, disease similarity network and circRNA-disease association network, and then sends these original features to the XGBoost classifier for training latent features. The method utilizes the tree learned by the XGBoost model, the index of leaf that instance finally falls into, and the 1 of K coding to represent the latent features. Finally, the method combines the latent features from the XGBoost with the original features to train the final model for predicting the association between the circRNA and diseases. RESULTS: The tenfold cross-validation results of the XGBCDA method illustrate that the area under the ROC curve reaches 0.9860. In addition, the method presents a striking performance in the case studies of colorectal cancer, gastric cancer and cervical cancer. CONCLUSION: With fabulous performance in predicting potential circRNA-disease associations, the XGBCDA method has the promising ability to assist biomedical researchers in terms of circRNA-disease association prediction.

Gene signature of m6A RNA regulators in diagnosis, prognosis, treatment, and immune microenvironment for cervical cancer.

Continuing studies imply that m6A RNA modification is involved in the development of cervical cancer (CC), but lack strong support on recurrence and diagnosis prediction. In this research, a comprehensive analysis of 33 m6A regulators was performed to fulfill them. Here, we performed diagnostic and prognosis models and identified key regulators, respectively. Then the CC patients were separated into two clusters in accordance with 33 regulators, and participants in the cluster 1 had a worse prognosis. Subsequently, the m6AScore was calculated to quantify the m6A modification pattern based on regulators and we found that patients in cluster 1 had higher m6AScore. Afterwards, immune microenvironment, cell infiltration, escape analyses and tumor burden mutation analyses were executed, and results showed that m6AScore was correlated with them, but to a limited extent. Interestingly, HLAs and immune checkpoint expression, and immunophenoscore in patients with high-m6AScores were significantly lower than those in the low-m6AScore group. These suggested the m6AScores might be used to predict the feasibility of immunotherapy in patients. Results provided a distinctive perspective on m6A modification and theoretical basis for CC diagnosis, prognosis, clinical treatment strategies, and potential mechanism exploration.

Comprehensive analysis of prognosis-related alternative splicing events in ovarian cancer.

Ovarian cancer (OV) is characterized by high incidence and poor prognosis. Increasing evidence indicates that aberrant alternative splicing (AS) events are associated with the pathogenesis of cancer. We examined prognosis-related alternative splicing events and constructed a clinically applicable model to predict patients' outcomes. Public database including The Cancer Genome Atlas (TCGA), TCGA SpliceSeq, and the Genomics of Drug Sensitivity in Cancer databases were used to detect the AS expression, immune cell infiltration and IC50. The prognosis-related AS model was constructed and validated by using Cox regression, LASSO regression, C-index, calibration plots, and ROC curves. A total of eight AS events (including FLT3LG|50942|AP) were selected to establish the prognosis-related AS model. Compared with high-risk group, low-risk group had a better outcome (P = 1.794e-06), was more sensitive to paclitaxel (P = 0.022), and higher proportions of plasma cells. We explored the upstream regulatory mechanisms of prognosis-related AS and found that two splicing factor and 156 tag single nucleotide polymorphisms may be involved in the regulation of prognosis-related AS. In order to assess patient prognosis more comprehensively, we constructed a clinically applicable model combining risk score and clinicopathological features, and the 1 -, and 3-year AUCs of the clinically applicable model were 0.812, and 0.726, which were 7.5% and 3.3% higher than that of the risk score. We constructed a prognostic signature for OV patients and comprehensively analysed the regulatory characteristics of the prognostic AS events in OV.

Overexpression of SHARPIN promotes tumor progression in ovarian cancer.

SHARPIN (Shank-associated RH domain interacting protein) plays an important role in tumorigenesis. However, its role in ovarian cancer remains largely unknown. To investigate this issue, we systematically analyzed the amplification and expression of the SHARPIN in the TCGA database. From the database, we found that SHARPIN was amplified in ovarian cancer compared to normal ovarian tissue, and the mRNA level of SHARPIN was significantly elevated in ovarian cancer compared to non-tumorigenic ovarian tissue. In addition, we observed similar results from ovarian cancer cell lines and clinical samples from ovarian cancer patients, which indicated that increased SHARPIN expression is associated with tumorigenesis in ovarian cancer. SHARPIN knockdown inhibited the migration and invasion of ovarian cancer cells, also inhibited cell cycle and promoted apoptosis, thereby suppressing cell proliferation. RNA-seq results showed that SHARPIN significantly increased the expression of P53 and P21 and decreased the expression of Cyclin D1 and c-Myc, all of which are involved in the regulation of cell proliferation. Subsequent mechanistic exploration revealed that SHARPIN knockdown increased the expression of caspases 3 and 9, leading to apoptosis of ovarian cancer cells. We also found that high expression of SHARPIN was associated with poor prognosis of ovarian cancer patients. Collectively, we demonstrated a positive correlation between SHARPIN and ovarian cancer progression and provide a basis for combined targeted therapy strategies for future ovarian cancer treatment.

Construction of a hypoxia-immune-related prognostic model and targeted therapeutic strategies for cervical cancer.

Emerging evidence indicates that hypoxia and immunity play important roles in tumorigenesis and development. However, the hypoxia-immune-related prognostic risk model has not been established in cervical cancer (CC). We aimed to construct a hypoxia-immune-related prognostic risk model, which has potential application in predicting the prognosis of CC patients and the response to targeted therapy. The RNA-seq data and corresponding clinical information were retrieved from The Cancer Genome Atlas (TCGA) database. The hypoxia status and immune status of CC patients were evaluated using the Consensus Clustering method and single-sample gene set enrichment analysis (ssGSEA), respectively. The univariate Cox regression, least absolute shrinkage and selection operator (LASSO) and multivariate Cox regression were applied to establish the prognostic risk model of CC. The chemotherapy response for six chemotherapeutic agents of each CC patient was calculated according to the Genomics of Drug Sensitivity in Cancer (GDSC). And the Connectivity Map (CMap) database was performed to screen candidate small-molecule drugs. In this study, we identified seven gene signatures (P4HA2, MSMO1, EGLN1, ZNF316, IKZF3, ISCU and MYO1B) with prognostic values. And the survival time of patients with low risk was significantly longer than those with high risk. Meanwhile, CC patients in the high-risk group yielded higher sensitivity to five chemotherapeutic agents. And we listed 10 candidate small-molecule drugs that exhibited a high correlation with the prognosis of CC. Thus, the prognostic model can accurately predict the prognosis of patients with CC and may be helpful for the development of new hypoxia-immune prognostic markers and therapeutic strategies for CC.

Effect of glycation on physicochemical properties and volatile flavor characteristics of silver carp mince.

The purpose of this study was to explore the effect of glycation on physicochemical properties and volatile flavor characteristics of silver carp mince (SCM). The changes in the degree of grafting, chemical composition, pH, color, total amino acid composition, and volatile flavor compounds of SCM with or without glucose were studied at different heating times. The results showed that the addition of glucose could promote the glycation reaction rate of SCM. Lysine and cysteine were the main amino acids involved in glycation. Glycation enhanced the overall aroma of SCM by accelerating lipid oxidation and Strecker degradation. In conclusion, these results suggest that glycation can enhance the volatile flavor of SCM during thermal processing and can be used as a volatile flavor enhancement technology for the development of protein nutrition food with good flavor from low-value fish.

Regulation of PTEN and ovarian cancer progression by an E3 ubiquitin ligase RBCK1.

Ovarian cancer is one of the most lethal gynecologic malignancies worldwide, with the 5-year survival is less than 50%. Although some clinical achievements have been achieved, the overall survival rate has remained unchanged over the past 20 years. Therefore, it is necessary and urgent to develop the potential modifiers and therapeutic approach to improve the overall survival rate in ovarian cancer patients. RBCK1 is an RING protein E3 ubiquitin ligase, which was revealed to involve in the progression of several cancers through its ubiquitination function. In this research, we report that RBCK1 expression is significantly elevated in human ovarian cancer and strongly associated with poor patients' prognosis. RBCK1 deficiency induces cell apoptosis and inhibits cell proliferation and migration in ovarian cancer cells. In terms of molecular mechanism, we report that RBCK1 interacts with PTEN and promotes PTEN degradation in K48-linked ubiquitination. Our study suggests a new and interesting regulatory mechanism that RBCK1 facilitates PTEN degradation, which could be a new potential therapeutic target for ovarian cancer treatment.

DeepDDS: deep graph neural network with attention mechanism to predict synergistic drug combinations.

MOTIVATION: Drug combination therapy has become an increasingly promising method in the treatment of cancer. However, the number of possible drug combinations is so huge that it is hard to screen synergistic drug combinations through wet-lab experiments. Therefore, computational screening has become an important way to prioritize drug combinations. Graph neural network has recently shown remarkable performance in the prediction of compound-protein interactions, but it has not been applied to the screening of drug combinations. RESULTS: In this paper, we proposed a deep learning model based on graph neural network and attention mechanism to identify drug combinations that can effectively inhibit the viability of specific cancer cells. The feature embeddings of drug molecule structure and gene expression profiles were taken as input to multilayer feedforward neural network to identify the synergistic drug combinations. We compared DeepDDS (Deep Learning for Drug-Drug Synergy prediction) with classical machine learning methods and other deep learning-based methods on benchmark data set, and the leave-one-out experimental results showed that DeepDDS achieved better performance than competitive methods. Also, on an independent test set released by well-known pharmaceutical enterprise AstraZeneca, DeepDDS was superior to competitive methods by more than 16% predictive precision. Furthermore, we explored the interpretability of the graph attention network and found the correlation matrix of atomic features revealed important chemical substructures of drugs. We believed that DeepDDS is an effective tool that prioritized synergistic drug combinations for further wet-lab experiment validation. AVAILABILITY AND IMPLEMENTATION: Source code and data are available at https://github.com/Sinwang404/DeepDDS/tree/master.

Expanding the Colorectal Cancer Biomarkers Based on the Human Gut Phageome.

With the increasing prevalence of colorectal cancer (CRC), extending the present biomarkers for the diagnosis of colorectal cancer is crucial. Previous studies have highlighted the importance of bacteriophages in gastrointestinal diseases, suggesting the potential value of gut phageome in early CRC diagnostic. Here, based on 317 metagenomic samples of three discovery cohorts collected from China (Hong Kong), Austria, and Japan, five intestinal bacteriophages, including Fusobacterium nucleatum, Peptacetobacter hiranonis, and Parvimonas micra phages were identified as potential CRC biomarkers. The five CRC enriched bacteriophagic markers classified patients from controls with an area under the receiver-operating characteristics curve (AUC) of 0.8616 across different populations. Subsequently, we used a total of 80 samples from China (Hainan) and Italy for validation. The AUC of the validation cohort is 0.8197. Moreover, to further explore the specificity of the five intestinal bacteriophage biomarkers in a broader background, we performed a confirmatory meta-analysis using two inflammatory bowel disease cohorts, ulcerative colitis (UC) and Crohn's disease (CD). Excitingly, we observed that the five CRC-enriched phage markers also exhibited high discrimination in UC (AUC = 78.02%). Unfortunately, the five CRC-rich phage markers did not show high resolution in CD (AUC = 48.00%). The present research expands the potential of microbial biomarkers in CRC diagnosis by building a more accurate classification model based on the human gut phageome, providing a new perspective for CRC gut phagotherapy. IMPORTANCE Worldwide, by 2020, colorectal cancer has become the third most common cancer after lung and breast cancer. Phages are strictly host-specific, and this specificity makes them more accurate as biomarkers, but phage biomarkers for colorectal cancer have not been thoroughly explored. Therefore, it is crucial to extend the existing phage biomarkers for the diagnosis of colorectal cancer. Here, we innovatively constructed a relatively accurate prediction model, including: three discovery cohorts, two additional validation cohorts and two cross-disease cohorts. A total of five possible biomarkers of intestinal bacteriophages were obtained. They are Peptacetobacter hiranonis Phage, Fusobacterium nucleatum animalis 7_1 Phage, Fusobacterium nucleatum polymorphum Phage, Fusobacterium nucleatum animalis 4_8 Phage, and Parvimonas micra Phage. This study aims at identifying fine-scale species-strain level phage biomarkers for colorectal cancer diseases, so as to expand the existing CRC biomarkers and provide a new perspective for intestinal phagocytosis therapy of colorectal cancer.

Fanconi Anemia Pathway Genes Advance Cervical Cancer via Immune Regulation and Cell Adhesion.

Fanconi anemia (FA) pathway is a typical and multienzyme-regulated DNA damage repairer that influences the occurrence and development of disease including cancers. Few comprehensive analyses were reported about the role of FA-related genes (FARGs) and their prognostic values in cancers. In this study, a comprehensive pan-cancer analysis on 79 FARGs was performed. According to the correlation analyses between HPV integration sites and FARGs, we found that FARGs played specific and critical roles in HPV-related cancers, especially in cervical cancer (CC). Based on this, a FARGs-associated prognostic risk score (FPS) model was constructed, and subsequently a nomogram model containing the FPS was developed with a good accuracy for CC overall survival (OS) and recurrence-free survival (RFS) outcome prediction. We also used the similar expression pattern of FARGs by consensus clustering analysis to separate the patients into three subgroups that exhibited significant differential OS but not RFS. Moreover, differential expressed genes (DEGs) between the two risk groups or three clusters were identified and immune pathways as well as cell adhesion processes were determined by functional enrichment analysis. Results indicated that FARGs might promote occurrence and development of CC by regulating the immune cells' infiltration and cell adhesion. In addition, through the machine learning models containing decision tree, random forest, naïve bayes, and support vector machine models, screening of important variables on CC prognosis, we finally determined that ZBTB32 and CENPS were the main elements affecting CC OS, while PALB2 and BRCA2 were for RFS. Kaplan-Meier analysis revealed that bivariate prediction of CC outcome was reliable. Our study systematically analyzed the prognostic prediction values of FARGs and demonstrated their potential mechanism in CC aggressiveness. Results provided perspective in FA pathway-associated modification and theoretical basis for CC clinical treatments.

Esophageal Cancer Associated Immune Genes as Biomarkers for Predicting Outcome in Upper Gastrointestinal Tumors.

Esophageal cancer (EC) is the seventh most common tumor in the world, ranking the sixth leading cause of cancer death, with a 5-year survival rate of 15-25%. Therefore, reliable prognostic biomarkers are needed to effectively predict the prognosis of EC. In this study, the gene profile information of the EC cohort served as a training set, which was derived from TCGA and Immport databases. GO and KEGG enrichment analysis was performed on the differential genes in normal and tumor groups of EC. The immune genes in differentially expressed genes (DEGs) were further obtained for univariate and multivariate Cox and Lasso regression analysis, and 6 independent immune genes (S100A3, STC2, HSPA6, CCL25, GPER1, and OSM) associated with prognosis were obtained to establish an immune risk score signature (IRSS). The signature was validated using head and neck cancers (HNSC) and gastric cancer (GC)in upper gastrointestinal malignancies as validation sets. The Kaplan-Meier results showed that the prognosis of the high-risk group was significantly favorable than that of the low-risk group in both the training set (P < 0.001; HR = 3.68, 95% CI = 2.14-6.35) and the validation set (P = 0.010; HR = 1.43, 95% CI = 1.09-1.88). A nomogram combining multiple clinical information and IRSS was more effective than a single independent prognostic factor in predicting outcome. This study explored the potential link between immunity and EC, and established and validated prognostic biomarkers that can effectively predict the prognosis of EC, HNSC and GC based on six immune genes.

Identification of autophagy-related risk signatures for the prognosis, diagnosis, and targeted therapy in cervical cancer.

BACKGROUND: To rummage autophagy-related prognostic, diagnostic, and therapeutic biomarkers in cervical cancer (CC). METHODS: The RNA-sequence and clinical information were from the TCGA and GTEx databases. We operated Cox regression to determine signatures related to overall survival (OS) and recurrence-free survival (RFS) respectively. The diagnostic and therapeutic effectiveness of prognostic biomarkers were further explored. RESULTS: We identified nine (VAMP7, MTMR14, ATG4D, KLHL24, TP73, NAMPT, CD46, HGS, ATG4C) and three risk signatures (SERPINA1, HSPB8, SUPT20H) with prognostic values for OS and RFS respectively. Six risk signatures (ATG4C, ATG4D, CD46, TP73, SERPINA1, HSPB8) were selected for qPCR. We screened five prognostic signatures(ATG4C, CD46, HSPB8, MTMR14, NAMPT) with diagnostic function through the GEO database. Correlation between our models and treatment targets certificated the prognostic score provided a reference for precision medicine. CONCLUSIONS: We constructed OS and RFS prognostic models in CC. Autophagy-related risk signatures might serve as diagnostic and therapeutic biomarkers.