SubtypeDrug: a software package for prioritization of candidate cancer subtype-specific drugs

Knowledge-Based Artificial Intelligence System for Drug Prioritization

In silico drug prioritization may be a promising and time-saving strategy to identify potential drugs, standing as a faster and more cost-effective approach than de novo approaches. In recent years, artificial intelligence has greatly evolved the drug development process. Here, we present a novel computational framework for drug prioritization, labyrinth, designed to simulate human knowledge retrieval and inference to identify potential drug candidates for each disease. With the integration of up-to-date clinical trials, literature co-occurrences, drug-target interactions, and disease similarities, our framework achieves over 90% predictive accuracy across clinical trial phases and strong alignment with clinical practice in TCGA cohorts. We have demonstrated effectiveness across 20 different disease categories with robust ROC-AUC metrics and the balance between predictive accuracy and model interpretability. We further demonstrate its effectiveness at both the population and the individual levels. This study not only demonstrates the capacity for its drug prioritization but underscores the importance of aligning computational models with intuitive human reasoning. We have wrapped the core function into an R package named labyrinth, which is freely available on GitHub under the GPL-v2 license (https://github.com/hanjunwei-lab/labyrinth).

NetSDR: Drug repurposing for cancers based on subtype-specific network modularization and perturbation analysis

Cancer, a heterogeneous disease, presents significant challenges for drug development due to its complex etiology. Drug repurposing, particularly through network medicine approaches, offers a promising avenue for cancer treatment by analyzing how drugs influence cellular networks on a systemic scale. The advent of large-scale proteomics data provides new opportunities to elucidate regulatory mechanisms specific to cancer subtypes. Herein, we present NetSDR, a Network-based Subtype-specific Drug Repurposing framework for prioritizing repurposed drugs specific to certain cancer subtypes, guided by subtype-specific proteomic signatures and network perturbations. First, by integrating cancer subtype information into a network-based method, we developed a pipeline to recognize subtype-specific functional modules. Next, we conducted drug response analysis for each module to identify the "therapeutic module" and then used deep learning to construct weighted drug response network for the particular subtype. Finally, we employed a perturbation response scanning-based drug repurposing method, which incorporates dynamic information, to facilitate the prioritization of candidate drugs. Applying the framework to gastric cancer, we attested the significance of the extracellular matrix module in treatment strategies and discovered a promising potential drug target, LAMB2, as well as a series of possible repurposed drugs. This study demonstrates a systems biology framework for precise drug repurposing in cancer and other complex diseases.

Strategies for robust, accurate, and generalizable benchmarking of drug discovery platforms

Benchmarking is an important step in the improvement, assessment, and comparison of the performance of drug discovery platforms and technologies. We revised the existing benchmarking protocols in our Computational Analysis of Novel Drug Opportunities (CANDO) multiscale therapeutic discovery platform to improve utility and performance. We optimized multiple parameters used in drug candidate prediction and assessment with these updated benchmarking protocols. CANDO ranked 7.4% of known drugs in the top 10 compounds for their respective diseases/indications based on drug-indication associations/mappings obtained from the Comparative Toxicogenomics Database (CTD) using these optimized parameters. This increased to 12.1% when drug-indication mappings were obtained from the Therapeutic Targets Database. Performance on an indication was weakly correlated (Spearman correlation coefficient >0.3) with indication size (number of drugs associated with an indication) and moderately correlated (correlation coefficient >0.5) with compound chemical similarity. There was also moderate correlation between our new and original benchmarking protocols when assessing performance per indication using each protocol. Benchmarking results were also dependent on the source of the drug-indication mapping used: a higher proportion of indication-associated drugs were recalled in the top 100 compounds when using the Therapeutic Targets Database (TTD), which only includes FDA-approved drug-indication associations (in contrast to the CTD, which includes associations drawn from the literature). We also created compbench, a publicly available head-to-head benchmarking protocol that allows consistent assessment and comparison of different drug discovery platforms. Using this protocol, we compared two pipelines for drug repurposing within CANDO; our primary pipeline outperformed another similarity-based pipeline still in development that clusters signatures based on their associated Gene Ontology terms. Our study sets a precedent for the complete, comprehensive, and comparable benchmarking of drug discovery platforms, resulting in more accurate drug candidate predictions.

Integrated edge information and pathway topology for drug-disease associations

Drug repurposing is a promising approach to find new therapeutic indications for approved drugs. Many computational approaches have been proposed to prioritize candidate anticancer drugs by gene or pathway level. However, these methods neglect the changes in gene interactions at the edge level. To address the limitation, we develop a computational drug repurposing method (iEdgePathDDA) based on edge information and pathway topology. First, we identify drug-induced and disease-related edges (the changes in gene interactions) within pathways by using the Pearson correlation coefficient. Next, we calculate the inhibition score between drug-induced edges and disease-related edges. Finally, we prioritize drug candidates according to the inhibition score on all disease-related edges. Case studies show that our approach successfully identifies new drug-disease pairs based on CTD database. Compared to the state-of-the-art approaches, the results demonstrate our method has the superior performance in terms of five metrics across colorectal, breast, and lung cancer datasets.

DrugRepoBank: a comprehensive database and discovery platform for accelerating drug repositioning

In recent years, drug repositioning has emerged as a promising alternative to the time-consuming, expensive and risky process of developing new drugs for diseases. However, the current database for drug repositioning faces several issues, including insufficient data volume, restricted data types, algorithm inaccuracies resulting from the neglect of multidimensional or heterogeneous data, a lack of systematic organization of literature data associated with drug repositioning, limited analytical capabilities and user-unfriendly webpage interfaces. Hence, we have established the first all-encompassing database called DrugRepoBank, consisting of two main modules: the 'Literature' module and the 'Prediction' module. The 'Literature' module serves as the largest repository of literature-supported drug repositioning data with experimental evidence, encompassing 169 repositioned drugs from 134 articles from 1 January 2000 to 1 July 2023. The 'Prediction' module employs 18 efficient algorithms, including similarity-based, artificial-intelligence-based, signature-based and network-based methods to predict repositioned drug candidates. The DrugRepoBank features an interactive and user-friendly web interface and offers comprehensive functionalities such as bioinformatics analysis of disease signatures. When users provide information about a drug, target or disease of interest, DrugRepoBank offers new indications and targets for the drug, proposes new drugs that bind to the target or suggests potential drugs for the queried disease. Additionally, it provides basic information about drugs, targets or diseases, along with supporting literature. We utilize three case studies to demonstrate the feasibility and effectiveness of predictively repositioned drugs within DrugRepoBank. The establishment of the DrugRepoBank database will significantly accelerate the pace of drug repositioning. Database URL:  https://awi.cuhk.edu.cn/DrugRepoBank.

PathwayTMB: A pathway-based tumor mutational burden analysis method for predicting the clinical outcome of cancer immunotherapy

Immunotherapy has become one of the most promising therapy methods for cancer, but only a small number of patients are responsive to it, indicating that more effective biomarkers are urgently needed. This study developed a pathway analysis method, named PathwayTMB, to identify genomic mutation pathways that serve as potential biomarkers for predicting the clinical outcome of immunotherapy. PathwayTMB first calculates the patient-specific pathway-based tumor mutational burden (PTMB) to reflect the cumulative extent of mutations for each pathway. It then screens mutated survival benefit-related pathways to construct an immune-related prognostic signature based on PTMB (IPSP). In a melanoma training set, IPSP-high patients presented a longer overall survival and a higher response rate than IPSP-low patients. Moreover, the IPSP showed a superior predictive effect compared with TMB. In addition, the prognostic and predictive value of the IPSP was consistently validated in two independent validation sets. Finally, in a multi-cancer dataset, PathwayTMB also exhibited good performance. Our results indicate that PathwayTMB could identify the mutation pathways for predicting immunotherapeutic survival, and their combination may serve as a potential predictive biomarker for immune checkpoint inhibitor therapy.

A drug repurposing method based on inhibition effect on gene regulatory network

Numerous computational drug repurposing methods have emerged as efficient alternatives to costly and time-consuming traditional drug discovery approaches. Some of these methods are based on the assumption that the candidate drug should have a reversal effect on disease-associated genes. However, such methods are not applicable in the case that there is limited overlap between disease-related genes and drug-perturbed genes. In this study, we proposed a novel Drug Repurposing method based on the Inhibition Effect on gene regulatory network (DRIE) to identify potential drugs for cancer treatment. DRIE integrated gene expression profile and gene regulatory network to calculate inhibition score by using the shortest path in the disease-specific network. The results on eleven datasets indicated the superior performance of DRIE when compared to other state-of-the-art methods. Case studies showed that our method effectively discovered novel drug-disease associations. Our findings demonstrated that the top-ranked drug candidates had been already validated by CTD database. Additionally, it clearly identified potential agents for three cancers (colorectal, breast, and lung cancer), which was beneficial when annotating drug-disease relationships in the CTD. This study proposed a novel framework for drug repurposing, which would be helpful for drug discovery and development.

DTSEA: A network-based drug target set enrichment analysis method for drug repurposing against COVID-19

The Coronavirus Disease 2019 (COVID-19) pandemic is still wreaking havoc worldwide. Therefore, the urgent need for efficient treatments pushes researchers and clinicians into screening effective drugs. Drug repurposing may be a promising and time-saving strategy to identify potential drugs against this disease. Here, we developed a novel computational approach, named Drug Target Set Enrichment Analysis (DTSEA), to identify potent drugs against COVID-19. DTSEA first mapped the disease-related genes into a gene functional interaction network, and then it used a network propagation algorithm to rank all genes in the network by calculating the network proximity of genes to disease-related genes. Finally, an enrichment analysis was performed on drug target sets to prioritize disease-candidate drugs. It was shown that the top three drugs predicted by DTSEA, including Ataluren, Carfilzomib, and Aripiprazole, were significantly enriched in the immune response pathways indicating the potential for use as promising COVID-19 inhibitors. In addition to these drugs, DTSEA also identified several drugs (such as Remdesivir and Olumiant), which have obtained emergency use authorization (EUA) for COVID-19. These results indicated that DTSEA could effectively identify the candidate drugs for COVID-19, which will help to accelerate the development of drugs for COVID-19. We then performed several validations to ensure the reliability and validity of DTSEA, including topological analysis, robustness analysis, and prediction consistency. Collectively, DTSEA successfully predicted candidate drugs against COVID-19 with high accuracy and reliability, thus making it a formidable tool to identify potential drugs for a specific disease and facilitate further investigation.

Drug mechanism enrichment analysis improves prioritization of therapeutics for repurposing

BACKGROUND

There is a pressing need for improved methods to identify effective therapeutics for diseases. Many computational approaches have been developed to repurpose existing drugs to meet this need. However, these tools often output long lists of candidate drugs that are difficult to interpret, and individual drug candidates may suffer from unknown off-target effects. We reasoned that an approach which aggregates information from multiple drugs that share a common mechanism of action (MOA) would increase on-target signal compared to evaluating drugs on an individual basis. In this study, we present drug mechanism enrichment analysis (DMEA), an adaptation of gene set enrichment analysis (GSEA), which groups drugs with shared MOAs to improve the prioritization of drug repurposing candidates.

RESULTS

First, we tested DMEA on simulated data and showed that it can sensitively and robustly identify an enriched drug MOA. Next, we used DMEA on three types of rank-ordered drug lists: (1) perturbagen signatures based on gene expression data, (2) drug sensitivity scores based on high-throughput cancer cell line screening, and (3) molecular classification scores of intrinsic and acquired drug resistance. In each case, DMEA detected the expected MOA as well as other relevant MOAs. Furthermore, the rankings of MOAs generated by DMEA were better than the original single-drug rankings in all tested data sets. Finally, in a drug discovery experiment, we identified potential senescence-inducing and senolytic drug MOAs for primary human mammary epithelial cells and then experimentally validated the senolytic effects of EGFR inhibitors.

CONCLUSIONS

DMEA is a versatile bioinformatic tool that can improve the prioritization of candidates for drug repurposing. By grouping drugs with a shared MOA, DMEA increases on-target signal and reduces off-target effects compared to analysis of individual drugs. DMEA is publicly available as both a web application and an R package at https://belindabgarana.github.io/DMEA .

A pathway-based mutation signature to predict the clinical outcomes and response to CTLA-4 inhibitors in melanoma

Immune checkpoint inhibitor (ICI) therapy has become a powerful clinical strategy for treating melanoma. The relationship between somatic mutations and the clinical benefits of immunotherapy has been widely recognized. However, the gene-based predictive biomarkers are less stable due to the heterogeneity of cancer at the individual gene level. Recent studies have suggested that the accumulation of gene mutations in biological pathways may activate antitumor immune responses. Herein, a novel pathway mutation signature (PMS) was constructed to predict the survival and efficacy of ICI therapy. In a dataset of melanoma patients treated with anti-CTLA-4, we mapped the mutated genes into the pathways and then identified seven significant mutation pathways associated with survival and immunotherapy response, which were used to construct the PMS model. According to the PMS model, the patients in the PMS-high group showed better overall survival (hazard ratio (HR) = 0.37; log-rank test, p < 0.0001) and progression-free survival (HR = 0.52; log-rank test, p = 0.014) than those in the PMS-low group. The PMS-high patients also showed a significantly higher objective response rate to anti-CTLA-4 therapy than the PMS-low patients (Fisher's exact test, p = 0.0055), and the predictive power of the PMS model was superior to that of TMB. Finally, the prognostic and predictive value of the PMS model was validated in two independent validation sets. Our study demonstrated that the PMS model can be considered a potential biomarker to predict the clinical outcomes and response to anti-CTLA-4 therapy in melanoma patients.

DrugSim2DR: systematic prediction of drug functional similarities in the context of specific disease for drug repurposing

BACKGROUND

Traditional approaches to drug development are costly and involve high risks. The drug repurposing approach can be a valuable alternative to traditional approaches and has therefore received considerable attention in recent years.

FINDINGS

Herein, we develop a previously undescribed computational approach, called DrugSim2DR, which uses a network diffusion algorithm to identify candidate anticancer drugs based on a drug functional similarity network. The innovation of the approach lies in the drug-drug functional similarity network constructed in a manner that implicitly links drugs through their common biological functions in the context of a specific disease state, as the similarity relationships based on general states (e.g., network proximity or Jaccard index of drug targets) ignore disease-specific molecular characteristics. The drug functional similarity network may provide a reference for prediction of drug combinations. We describe and validate the DrugSim2DR approach through analysis of data on breast cancer and lung cancer. DrugSim2DR identified some US Food and Drug Administration-approved anticancer drugs, as well as some candidate drugs validated by previous studies in the literature. Moreover, DrugSim2DR showed excellent predictive performance, as evidenced by receiver operating characteristic analysis and multiapproach comparisons in various cancer datasets.

CONCLUSIONS

DrugSim2DR could accurately assess drug-drug functional similarity within a specific disease context and may more effectively prioritize disease candidate drugs. To increase the usability of our approach, we have developed an R-based software package, DrugSim2DR, which is freely available on CRAN (https://CRAN.R-project.org/package=DrugSim2DR).

Identification of comutation in signaling pathways to predict the clinical outcomes of immunotherapy

BACKGROUND

Immune checkpoint blockades (ICBs) have emerged as a promising treatment for cancer. Recently, tumour mutational burden (TMB) and neoantigen load (NAL) have been proposed to be potential biomarkers to predict the efficacy of ICB; however, they were limited by difficulties in defining the cut-off values and inconsistent detection platforms. Therefore, it is critical to identify more effective predictive biomarkers for screening patients who will potentially benefit from immunotherapy. In this study, we aimed to identify comutated signaling pathways to predict the clinical outcomes of immunotherapy.

METHODS

Here, we comprehensively analysed the signaling pathway mutation status of 9763 samples across 33 different cancer types from The Cancer Genome Atlas (TCGA) by mapping the somatic mutations to the pathways. We then explored the comutated pathways that were associated with increased TMB and NAL by using receiver operating characteristic (ROC) curve analysis and multiple linear regressions.

RESULTS

Our results revealed that comutation of the Spliceosome (Sp) pathway and Hedgehog (He) signaling pathway (defined as SpHe-comut⁺) could be used as a predictor of increased TMB and NAL and was associated with increased levels of immune-related signatures. In seven independent immunotherapy cohorts, we validated that SpHe-comut⁺ patients exhibited a longer overall survival (OS) or progression-free survival (PFS) and a higher objective response rate (ORR) than SpHe-comut^- patients. Moreover, a combination of SpHe-comut status with PD-L1 expression further improved the predictive value for ICB therapy.

CONCLUSION

Overall, SpHe-comut⁺ was demonstrated to be an effective predictor of immunotherapeutic benefit in seven independent immunotherapy cohorts and may serve as a potential and convenient biomarker for the clinical application of ICB therapy.

Tumor Microenvironment Characterization in Breast Cancer Identifies Prognostic Pathway Signatures

Breast cancer is one of the most common female malignancies worldwide. Due to its early metastases formation and a high degree of malignancy, the 10 year-survival rate of metastatic breast cancer does not exceed 30%. Thus, more precise biomarkers are urgently needed. In our study, we first estimated the tumor microenvironment (TME) infiltration using the xCell algorithm. Based on TME infiltration, the three main TME clusters were identified using consensus clustering. Our results showed that the three main TME clusters cause significant differences in survival rates and TME infiltration patterns (log-rank test, p = 0.006). Then, multiple machine learning algorithms were used to develop a nine-pathway-based TME-related risk model to predict the prognosis of breast cancer (BRCA) patients (the immune-related pathway-based risk score, defined as IPRS). Based on the IPRS, BRCA patients were divided into two subgroups, and patients in the IPRS-low group presented significantly better overall survival (OS) rates than the IPRS-high group (log-rank test, p < 0.0001). Correlation analysis revealed that the IPRS-low group was characterized by increases in immune-related scores (cytolytic activity (CYT), major histocompatibility complex (MHC), T cell-inflamed immune gene expression profile (GEP), ESTIMATE, immune, and stromal scores) while exhibiting decreases in tumor purity, suggesting IPRS-low patients may have a strong immune response. Additionally, the gene-set enrichment analysis (GSEA) result confirmed that the IPRS-low patients were significantly enriched in several immune-associated signaling pathways. Furthermore, multivariate Cox analysis revealed that the IPRS was an independent prognostic biomarker after adjustment by clinicopathologic characteristics. The prognostic value of the IPRS model was further validated in three external validation cohorts. Altogether, our findings demonstrated that the IPRS was a powerful predictor to screen out certain populations with better prognosis in breast cancer and may serve as a potential biomarker guiding clinical treatment decisions.

Identification of DNA-binding proteins via Multi-view LSSVM with independence criterion

DNA-binding proteins actively participate in life activities such as DNA replication, recombination, gene expression and regulation and play a prominent role in these processes. As DNA-binding proteins continue to be discovered and increase, it is imperative to design an efficient and accurate identification tool. Considering the time-consuming and expensive traditional experimental technology and the insufficient number of samples in the biological computing method based on structural information, we proposed a machine learning algorithm based on sequence information to identify DNA binding proteins, named multi-view Least Squares Support Vector Machine via Hilbert-Schmidt Independence Criterion (multi-view LSSVM via HSIC). This method took 6 feature sets as multi-view input and trains a single view through the LSSVM algorithm. Then, we integrated HSIC into LSSVM as a regular term to reduce the dependence between views and explored the complementary information of multiple views. Subsequently, we trained and coordinated the submodels and finally combined the submodels in the form of weights to obtain the final prediction model. On training set PDB1075, the prediction results of our model were better than those of most existing methods. Independent tests are conducted on the datasets PDB186 and PDB2272. The accuracy of the prediction results was 85.5% and 79.36%, respectively. This result exceeded the current state-of-the-art methods, which showed that the multi-view LSSVM via HSIC can be used as an efficient predictor.

DRviaSPCN: a software package for drug repurposing in cancer via a subpathway crosstalk network

SUMMARY

Drug repurposing is an approach used to discover new indications for existing drugs. Recently, several computational approaches have been developed for drug repurposing in cancer. Nevertheless, no approaches have reported a systematic analysis of pathway crosstalk. Pathway crosstalk, which refers to the phenomenon of interaction or cooperation between pathways, is a critical aspect of tumor pathways that allows cancer cells to survive and acquire resistance to drug therapy. Here, we innovatively developed a system biology R-based software package, DRviaSPCN, to repurpose drugs for cancer via a subpathway (SP) crosstalk network. This package provides a novel approach to prioritize cancer candidate drugs by considering drug-induced SPs and their crosstalk effects. The operation modes mainly include construction of the SP network and calculation of the centrality scores of SPs to reflect the influence of SP crosstalk, calculation of enrichment scores of drug- and disease-induced dysfunctional SPs and weighted them by the centrality scores of SPs, evaluation of the drug-disease reverse association at the weighted SP level, identification of cancer candidate drugs, and visualization of the results. Its capabilities enable DRviaSPCN to find cancer candidate drugs, which will complement the recent tools which did not consider crosstalk among pathways/SPs. DRviaSPCN may help to facilitate the development of drug discovery.

AVAILABILITY AND IMPLEMENTATION

The package is implemented in R and available under GPL-2 license from the CRAN website (https://CRAN.R-project.org/package=DRviaSPCN).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A novel pathway mutation perturbation score predicts the clinical outcomes of immunotherapy

The link between tumor genetic variations and immunotherapy benefits has been widely recognized. Recent studies suggested that the key biological pathways activated by accumulated genetic mutations may act as an effective biomarker for predicting the efficacy of immune checkpoint inhibitor (ICI) therapy. Here, we developed a novel individual Pathway Mutation Perturbation (iPMP) method that measures the pathway mutation perturbation level by combining evidence of the cumulative effect of mutated genes with the position of mutated genes in the pathways. In iPMP, somatic mutations on a single sample were first mapped to genes in a single pathway to infer the pathway mutation perturbation score (PMPscore), and then, an integrated PMPscore profile was produced, which can be used in place of the original mutation dataset to identify associations with clinical outcomes. To illustrate the effect of iPMP, we applied it to a melanoma cohort treated with ICIs and identified seven significant perturbation pathways, which jointly constructed a pathway-based signature. With the signature, patients were classified into two subgroups with significant distinctive overall survival and objective response rate to immunotherapy. Moreover, the pathway-based signature was consistently validated in two independent melanoma cohorts. We further applied iPMP to two non-small cell lung cancer cohorts and also obtained good performance. Altogether, the iPMP method could be used to identify the significant mutation perturbation pathways for constructing the pathway-based biomarker to predict the clinical outcomes of immunotherapy. The iPMP method has been implemented as a freely available R-based package (https://CRAN.R-project.org/package=PMAPscore).

Discussion of tumor mutation burden as an indicator to predict efficacy of immune checkpoint inhibitors: A case report

There are many treatment options for advanced lung cancer, among which immunotherapy has developed rapidly and benefited a lot of patients. However, immunotherapy can only benefit a subgroup of patients, and how to select patients suitable for this therapy is critical. Tumor mutation burden (TMB) is one of the important reference indicators for immune checkpoint inhibitors (ICIs). However, there are many factors influencing the usage of this indicator, which will lead to considerable consequences if not treated well. In this study, we performed a case study on a male advanced lung squamous cell carcinoma patient of age 83. The patient suffered from “cough and sputum”, and did chest CT scans on 24 October 2018, which showed “a mass-like mass in the anterior segment of the right lung upper lobe, about 38mm×28mm”. He was treated with systemic chemotherapy; however, the tumor was still under progression. Although PD-L1 was not tested in gene testing, he had a TMB value of 10.26 mutations/Mb with a quantile value 88.63%. Thus, “toripalimab injection” was added as immunotherapy and the size of the lesion decreased. In summary, we adopted a clinical case as the basis to explore the value and significance of TMB in immunotherapy in this study. We hope that more predictive molecular markers will be discovered, which will bring more treatment methods for advanced lung cancer.

A Machine Learning Method to Trace Cancer Primary Lesion Using Microarray-Based Gene Expression Data

Cancer of unknown primary site (CUP) is a heterogeneous group of cancers whose tissue of origin remains unknown after detailed investigation by conventional clinical methods. The number of CUP accounts for roughly 3%–5% of all human malignancies. CUP patients are usually treated with broad-spectrum chemotherapy, which often leads to a poor prognosis. Recent studies suggest that the treatment targeting the primary lesion of CUP will significantly improve the prognosis of the patient. Therefore, it is urgent to develop an efficient method to accurately detect tissue of origin of CUP in clinical cancer research. In this work, we developed a novel framework that uses Extreme Gradient Boosting (XGBoost) to trace the primary site of CUP based on microarray-based gene expression data. First, we downloaded the microarray-based gene expression profiles of 59,385 genes for 57,08 samples from The Cancer Genome Atlas (TCGA) and 6,364 genes for 3,101 samples from the Gene Expression Omnibus (GEO). Both data were divided into training and independent testing data with a ratio of 4:1. Then, we obtained in the training data 200 and 290 genes from TCGA and the GEO datasets, respectively, to train XGBoost models for the identification of the primary site of CUP. The overall 5-fold cross-validation accuracies of our methods were 96.9% and 95.3% on TCGA and GEO training datasets, respectively. Meanwhile, the macro-precision for the independent dataset reached 96.75% and 98.8% on, respectively, TCGA and GEO. Experimental results demonstrated that the XGBoost framework not only can reduce the cost of clinical cancer traceability but also has high efficiency, which might be useful in clinical usage.

Development and Validation of a Three-Gene Prognostic Signature Based on Tumor Microenvironment for Gastric Cancer

Gastric cancer (GC), which has high morbidity and low survival rate, is one of the most common malignant tumors in the world. The increasing evidences show that the tumor microenvironment (TME) is related to the occurrence and progression of tumors and the prognosis of patients. In this study, we aimed to develop a TME-based prognostic signature for GC. We first identified the differentially expressed genes (DEGs) related to the TME using the Wilcoxon rank-sum test in a training set of GC. Univariate Cox regression analysis was used to identify prognostic-related DEGs. To decrease the overfitting, we performed the least absolute shrinkage and selection operator (LASSO) regression to reduce the number of signature genes and obtained three genes (LPPR4, ADAM12, NOX4). Next, the multivariate Cox regression was performed to construct the risk score model, and a three-gene prognostic signature was developed. According to the signature, patients were classified into high-risk and low-risk groups with significantly different survival. The signature was then applied to three independent validated sets and obtained the same results. We conducted the time-dependent Receiver Operating Characteristic (ROC) curve analysis to evaluate our signature. We further evaluated the differential immune characters between high-risk and low-risk patients to reveal the potential immune mechanism of the impact on the prognosis of the model. Overall, we identified a three-gene prognostic signature based on TME to predict the prognosis of patients with GC and facilitate the development of a precise treatment strategy.

A SNARE Protein Identification Method Based on iLearnPlus to Efficiently Solve the Data Imbalance Problem

Machine learning has been widely used to solve complex problems in engineering applications and scientific fields, and many machine learning-based methods have achieved good results in different fields. SNAREs are key elements of membrane fusion and required for the fusion process of stable intermediates. They are also associated with the formation of some psychiatric disorders. This study processes the original sequence data with the synthetic minority oversampling technique (SMOTE) to solve the problem of data imbalance and produces the most suitable machine learning model with the iLearnPlus platform for the identification of SNARE proteins. Ultimately, a sensitivity of 66.67%, specificity of 93.63%, accuracy of 91.33%, and MCC of 0.528 were obtained in the cross-validation dataset, and a sensitivity of 66.67%, specificity of 93.63%, accuracy of 91.33%, and MCC of 0.528 were obtained in the independent dataset (the adaptive skip dipeptide composition descriptor was used for feature extraction, and LightGBM with proper parameters was used as the classifier). These results demonstrate that this combination can perform well in the classification of SNARE proteins and is superior to other methods.

Identify DNA-Binding Proteins Through the Extreme Gradient Boosting Algorithm

The exploration of DNA-binding proteins (DBPs) is an important aspect of studying biological life activities. Research on life activities requires the support of scientific research results on DBPs. The decline in many life activities is closely related to DBPs. Generally, the detection method for identifying DBPs is achieved through biochemical experiments. This method is inefficient and requires considerable manpower, material resources and time. At present, several computational approaches have been developed to detect DBPs, among which machine learning (ML) algorithm-based computational techniques have shown excellent performance. In our experiments, our method uses fewer features and simpler recognition methods than other methods and simultaneously obtains satisfactory results. First, we use six feature extraction methods to extract sequence features from the same group of DBPs. Then, this feature information is spliced together, and the data are standardized. Finally, the extreme gradient boosting (XGBoost) model is used to construct an effective predictive model. Compared with other excellent methods, our proposed method has achieved better results. The accuracy achieved by our method is 78.26% for PDB2272 and 85.48% for PDB186. The accuracy of the experimental results achieved by our strategy is similar to that of previous detection methods.

Identification of Helicobacter pylori Membrane Proteins Using Sequence-Based Features

Helicobacter pylori (H. pylori) is the most common risk factor for gastric cancer worldwide. The membrane proteins of the H. pylori are involved in bacterial adherence and play a vital role in the field of drug discovery. Thus, an accurate and cost-effective computational model is needed to predict the uncharacterized membrane proteins of H. pylori. In this study, a reliable benchmark dataset consisted of 114 membrane and 219 nonmembrane proteins was constructed based on UniProt. A support vector machine- (SVM-) based model was developed for discriminating H. pylori membrane proteins from nonmembrane proteins by using sequence information. Cross-validation showed that our method achieved good performance with an accuracy of 91.29%. It is anticipated that the proposed model will be useful for the annotation of H. pylori membrane proteins and the development of new anti-H. pylori agents.

Multiple Laplacian Regularized RBF Neural Network for Assessing Dry Weight of Patients With End-Stage Renal Disease

Dry weight (DW) is an important dialysis index for patients with end-stage renal disease. It can guide clinical hemodialysis. Brain natriuretic peptide, chest computed tomography image, ultrasound, and bioelectrical impedance analysis are key indicators (multisource information) for assessing DW. By these approaches, a trial-and-error method (traditional measurement method) is employed to assess DW. The assessment of clinician is time-consuming. In this study, we developed a method based on artificial intelligence technology to estimate patient DW. Based on the conventional radial basis function neural (RBFN) network, we propose a multiple Laplacian-regularized RBFN (MLapRBFN) model to predict DW of patient. Compared with other model and body composition monitor, our method achieves the lowest value (1.3226) of root mean square error. In Bland-Altman analysis of MLapRBFN, the number of out agreement interval is least (17 samples). MLapRBFN integrates multiple Laplace regularization terms, and employs an efficient iterative algorithm to solve the model. The ratio of out agreement interval is 3.57%, which is lower than 5%. Therefore, our method can be tentatively applied for clinical evaluation of DW in hemodialysis patients.

Graph Embedding Based Novel Gene Discovery Associated With Diabetes Mellitus

Diabetes mellitus is a group of complex metabolic disorders which has affected hundreds of millions of patients world-widely. The underlying pathogenesis of various types of diabetes is still unclear, which hinders the way of developing more efficient therapies. Although many genes have been found associated with diabetes mellitus, more novel genes are still needed to be discovered towards a complete picture of the underlying mechanism. With the development of complex molecular networks, network-based disease-gene prediction methods have been widely proposed. However, most existing methods are based on the hypothesis of guilt-by-association and often handcraft node features based on local topological structures. Advances in graph embedding techniques have enabled automatically global feature extraction from molecular networks. Inspired by the successful applications of cutting-edge graph embedding methods on complex diseases, we proposed a computational framework to investigate novel genes associated with diabetes mellitus. There are three main steps in the framework: network feature extraction based on graph embedding methods; feature denoising and regeneration using stacked autoencoder; and disease-gene prediction based on machine learning classifiers. We compared the performance by using different graph embedding methods and machine learning classifiers and designed the best workflow for predicting genes associated with diabetes mellitus. Functional enrichment analysis based on Human Phenotype Ontology (HPO), KEGG, and GO biological process and publication search further evaluated the predicted novel genes.

iTTCA-RF: a random forest predictor for tumor T cell antigens

BACKGROUND

Cancer is one of the most serious diseases threatening human health. Cancer immunotherapy represents the most promising treatment strategy due to its high efficacy and selectivity and lower side effects compared with traditional treatment. The identification of tumor T cell antigens is one of the most important tasks for antitumor vaccines development and molecular function investigation. Although several machine learning predictors have been developed to identify tumor T cell antigen, more accurate tumor T cell antigen identification by existing methodology is still challenging.

METHODS

In this study, we used a non-redundant dataset of 592 tumor T cell antigens (positive samples) and 393 tumor T cell antigens (negative samples). Four types feature encoding methods have been studied to build an efficient predictor, including amino acid composition, global protein sequence descriptors and grouped amino acid and peptide composition. To improve the feature representation ability of the hybrid features, we further employed a two-step feature selection technique to search for the optimal feature subset. The final prediction model was constructed using random forest algorithm.

RESULTS

Finally, the top 263 informative features were selected to train the random forest classifier for detecting tumor T cell antigen peptides. iTTCA-RF provides satisfactory performance, with balanced accuracy, specificity and sensitivity values of 83.71%, 78.73% and 88.69% over tenfold cross-validation as well as 73.14%, 62.67% and 83.61% over independent tests, respectively. The online prediction server was freely accessible at http://lab.malab.cn/~acy/iTTCA .

CONCLUSIONS

We have proven that the proposed predictor iTTCA-RF is superior to the other latest models, and will hopefully become an effective and useful tool for identifying tumor T cell antigens presented in the context of major histocompatibility complex class I.

Identification of Biomarkers for Cervical Cancer Radiotherapy Resistance Based on RNA Sequencing Data

Cervical cancer as a common gynecological malignancy threatens the health and lives of women. Resistance to radiotherapy is the primary cause of treatment failure and is mainly related to difference in the inherent vulnerability of tumors after radiotherapy. Here, we investigated signature genes associated with poor response to radiotherapy by analyzing an independent cervical cancer dataset from the Gene Expression Omnibus, including pre-irradiation and mid-irradiation information. A total of 316 differentially expressed genes were significantly identified. The correlations between these genes were investigated through the Pearson correlation analysis. Subsequently, random forest model was used in determining cancer-related genes, and all genes were ranked by random forest scoring. The top 30 candidate genes were selected for uncovering their biological functions. Functional enrichment analysis revealed that the biological functions chiefly enriched in tumor immune responses, such as cellular defense response, negative regulation of immune system process, T cell activation, neutrophil activation involved in immune response, regulation of antigen processing and presentation, and peptidyl-tyrosine autophosphorylation. Finally, the top 30 genes were screened and analyzed through literature verification. After validation, 10 genes (KLRK1, LCK, KIF20A, CD247, FASLG, CD163, ZAP70, CD8B, ZNF683, and F10) were to our objective. Overall, the present research confirmed that integrated bioinformatics methods can contribute to the understanding of the molecular mechanisms and potential therapeutic targets underlying radiotherapy resistance in cervical cancer.

Deep-4mCW2V: A sequence-based predictor to identify N4-methylcytosine sites in Escherichia coli

N4-methylcytosine (4mC) is a type of DNA modification which could regulate several biological progressions such as transcription regulation, replication and gene expressions. Precisely recognizing 4mC sites in genomic sequences can provide specific knowledge about their genetic roles. This study aimed to develop a deep learning-based model to predict 4mC sites in the Escherichia coli. In the model, DNA sequences were encoded by word embedding technique 'word2vec'. The obtained features were inputted into 1-D convolutional neural network (CNN) to discriminate 4mC sites from non-4mC sites in Escherichia coli genome. The examination on independent dataset showed that our model could yield the overall accuracy of 0.861, which was about 4.3% higher than the existing model. To provide convenience to scholars, we provided the data and source code of the model which can be freely download from https://github.com/linDing-groups/Deep-4mCW2V.

Identification of Somatic Mutation-Driven Immune Cells by Integrating Genomic and Transcriptome Data

Tumor somatic mutations in protein-coding regions may generate neoantigens which may trigger antitumor immune cell response. Increasing evidence supports that immune cell response may profoundly influence tumor progression. However, there are no calculated tools to systematically identify immune cells driven by specific somatic mutations. It is urgent to develop a calculated method to comprehensively detect tumor-infiltrating immune cells driven by the specific somatic mutations in cancer. We developed a novel software package (SMDIC) that enables the automated identification of somatic mutation-driven immune cell. SMDIC provides a novel pipeline to discover mutation-specific immune cells by integrating genomic and transcriptome data. The operation modes include inference of the relative abundance matrix of tumor-infiltrating immune cells, detection of differential abundance immune cells with respect to the gene mutation status, conversion of the abundance matrix of significantly dysregulated cells into two binary matrices (one for upregulated and one for downregulated cells), identification of somatic mutation-driven immune cells by comparing the gene mutation status with each immune cell in the binary matrices across all samples, and visualization of immune cell abundance of samples in different mutation status for each gene. SMDIC provides a user-friendly tool to identify somatic mutation-specific immune cell response. SMDIC may contribute to understand the mechanisms underlying anticancer immune response and find targets for cancer immunotherapy. The SMDIC was implemented as an R-based tool which was freely available from the CRAN website https://CRAN.R-project.org/package=SMDIC.