Support vector machine with Dirichlet feature mapping

Bioinformatics identification and validation of maternal blood biomarkers and immune cell infiltration in preeclampsia: An observational study

Preeclampsia (PE) is a pregnancy complication characterized by placental dysfunction. However, the relationship between maternal blood markers and PE is unclear. It is helpful to improve the diagnosis and treatment of PE using new biomarkers related to PE in the blood. Three PE-related microarray datasets were obtained from the Gene Expression Synthesis database. The limma software package was used to identify differentially expressed genes (DEGs) between PE and control groups. Least absolute shrinkage and selection operator regression, support vector machine, random forest, and multivariate logistic regression analyses were used to determine key diagnostic biomarkers, which were verified using clinical samples. Subsequently, functional enrichment analysis was performed. In addition, the datasets were combined for immune cell infiltration analysis and to determine their relationships with core diagnostic biomarkers. The diagnostic performance of key genes was evaluated using the receiver operating characteristic (ROC) curve, C-index, and GiViTi calibration band. Genes with potential clinical applications were evaluated using decision curve analysis (DCA). Seventeen DEGs were identified, and 6 key genes (FN1, MYADM, CA6, PADI4, SLC4A10, and PPP4R1L) were obtained using 3 types of machine learning methods and logistic regression. High diagnostic performance was found for PE through evaluation of the ROC, C-index, GiViti calibration band, and DCA. The 2 types of immune cells (M0 macrophages and activated mast cells) were significantly different between patients with PE and controls. All of these genes except SLC4A10 showed significant differences in expression levels between the 2 groups using quantitative reverse transcription-polymerase chain reaction. This model used 6 maternal blood markers to predict the occurrence of PE. The findings may stimulate ideas for the treatment and prevention of PE.

Identification of MACC1 as a potential biomarker for pulmonary arterial hypertension based on bioinformatics and machine learning

BACKGROUND

Pulmonary arterial hypertension (PAH) is a life-threatening disease characterized by abnormal early activation of pulmonary arterial smooth muscle cells (PASMCs), yet the underlying mechanisms remain to be elucidated.

METHODS

Normal and PAH gene expression profiles were obtained from the Gene Expression Omnibus (GEO) database and analyzed using gene set enrichment analysis (GSEA) to uncover the underlying mechanisms. Weighted gene co-expression network analysis (WGCNA) and machine learning methods were deployed to further filter hub genes. A number of immune infiltration analysis methods were applied to explore the immune landscape of PAH. Enzyme-linked immunosorbent assay (ELISA) was employed to compare MACC1 levels between PAH and normal subjects. The important role of MACC1 in the progression of PAH was verified through Western blot and real-time qPCR, among others.

RESULTS

39 up-regulated and 7 down-regulated genes were identified by 'limma' and 'RRA' packages. WGCNA and machine learning further narrowed down the list to 4 hub genes, with MACC1 showing strong diagnostic capacity. In vivo and in vitro experiments revealed that MACC1 was highsly associated with malignant features of PASMCs in PAH.

CONCLUSIONS

These findings suggest that targeting MACC1 may offer a promising therapeutic strategy for treating PAH, and further clinical studies are warranted to evaluate its efficacy.

Identification of sudden cardiac death from human blood using ATR-FTIR spectroscopy and machine learning

OBJECTIVE

The aim of this study is to identify a rapid, sensitive, and non-destructive auxiliary approach for postmortem diagnosis of SCD, addressing the challenges faced in forensic practice.

METHODS

ATR-FTIR spectroscopy was employed to collect spectral features of blood samples from different cases, combined with pathological changes. Mixed datasets were analyzed using ANN, KNN, RF, and SVM algorithms. Evaluation metrics such as accuracy, precision, recall, F1-score and confusion matrix were used to select the optimal algorithm and construct the postmortem diagnosis model for SCD.

RESULTS

A total of 77 cases were collected, including 43 cases in the SCD group and 34 cases in the non-SCD group. A total of 693 spectrogram were obtained. Compared to other algorithms, the SVM algorithm demonstrated the highest accuracy, reaching 95.83% based on spectral biomarkers. Furthermore, by combing spectral biomarkers with age, gender, and cardiac histopathological changes, the accuracy of the SVM model could get 100%.

CONCLUSION

Integrating artificial intelligence technology, pathology, and physical chemistry analysis of blood components can serve as an effective auxiliary method for postmortem diagnosis of SCD.

Identification of hub genes in calcific aortic valve disease

Calcific aortic valve disease (CAVD) is a heart valve disorder characterized primarily by calcification of the aortic valve, resulting in stiffness and dysfunction of the valve. CAVD is prevalent among aging populations and is linked to factors such as hypertension, dyslipidemia, tobacco use, and genetic predisposition, and can result in becoming a growing economic and health burden. Once aortic valve calcification occurs, it will inevitably progress to aortic stenosis. At present, there are no medications available that have demonstrated effectiveness in managing or delaying the progression of the disease. In this study, we mined four publicly available microarray datasets (GSE12644 GSE51472, GSE77287, GSE233819) associated with CAVD from the GEO database with the aim of identifying hub genes associated with the occurrence of CAVD and searching for possible biological targets for the early prevention and diagnosis of CAVD. This study provides preliminary evidence for therapeutic and preventive targets for CAVD and may provide a solid foundation for subsequent biological studies.

Predictive Value of Machine Learning for Platinum Chemotherapy Responses in Ovarian Cancer: Systematic Review and Meta-Analysis

BACKGROUND

Machine learning is a potentially effective method for predicting the response to platinum-based treatment for ovarian cancer. However, the predictive performance of various machine learning methods and variables is still a matter of controversy and debate.

OBJECTIVE

This study aims to systematically review relevant literature on the predictive value of machine learning for platinum-based chemotherapy responses in patients with ovarian cancer.

METHODS

Following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, we systematically searched the PubMed, Embase, Web of Science, and Cochrane databases for relevant studies on predictive models for platinum-based therapies for the treatment of ovarian cancer published before April 26, 2023. The Prediction Model Risk of Bias Assessment tool was used to evaluate the risk of bias in the included articles. Concordance index (C-index), sensitivity, and specificity were used to evaluate the performance of the prediction models to investigate the predictive value of machine learning for platinum chemotherapy responses in patients with ovarian cancer.

RESULTS

A total of 1749 articles were examined, and 19 of them involving 39 models were eligible for this study. The most commonly used modeling methods were logistic regression (16/39, 41%), Extreme Gradient Boosting (4/39, 10%), and support vector machine (4/39, 10%). The training cohort reported C-index in 39 predictive models, with a pooled value of 0.806; the validation cohort reported C-index in 12 predictive models, with a pooled value of 0.831. Support vector machine performed well in both the training and validation cohorts, with a C-index of 0.942 and 0.879, respectively. The pooled sensitivity was 0.890, and the pooled specificity was 0.790 in the training cohort.

CONCLUSIONS

Machine learning can effectively predict how patients with ovarian cancer respond to platinum-based chemotherapy and may provide a reference for the development or updating of subsequent scoring systems.

New techniques to identify the tissue of origin for cancer of unknown primary in the era of precision medicine: progress and challenges

Despite a standardized diagnostic examination, cancer of unknown primary (CUP) is a rare metastatic malignancy with an unidentified tissue of origin (TOO). Patients diagnosed with CUP are typically treated with empiric chemotherapy, although their prognosis is worse than those with metastatic cancer of a known origin. TOO identification of CUP has been employed in precision medicine, and subsequent site-specific therapy is clinically helpful. For example, molecular profiling, including genomic profiling, gene expression profiling, epigenetics and proteins, has facilitated TOO identification. Moreover, machine learning has improved identification accuracy, and non-invasive methods, such as liquid biopsy and image omics, are gaining momentum. However, the heterogeneity in prediction accuracy, sample requirements and technical fundamentals among the various techniques is noteworthy. Accordingly, we systematically reviewed the development and limitations of novel TOO identification methods, compared their pros and cons and assessed their potential clinical usefulness. Our study may help patients shift from empirical to customized care and improve their prognoses.

Multifactor assessment of ovarian cancer reveals immunologically interpretable molecular subtypes with distinct prognoses

Background

Ovarian cancer (OC) is a highly heterogeneous and malignant gynecological cancer, thereby leading to poor clinical outcomes. The study aims to identify and characterize clinically relevant subtypes in OC and develop a diagnostic model that can precisely stratify OC patients, providing more diagnostic clues for OC patients to access focused therapeutic and preventative strategies.

Methods

Gene expression datasets of OC were retrieved from TCGA and GEO databases. To evaluate immune cell infiltration, the ESTIMATE algorithm was applied. A univariate Cox analysis and the two-sided log-rank test were used to screen OC risk factors. We adopted the ConsensusClusterPlus algorithm to determine OC subtypes. Enrichment analysis based on KEGG and GO was performed to determine enriched pathways of signature genes for each subtype. The machine learning algorithm, support vector machine (SVM) was used to select the feature gene and develop a diagnostic model. A ROC curve was depicted to evaluate the model performance.

Results

A total of 1,273 survival-related genes (SRGs) were firstly determined and used to clarify OC samples into different subtypes based on their different molecular pattern. SRGs were successfully stratified in OC patients into three robust subtypes, designated S-I (Immunoreactive and DNA Damage repair), S-II (Mixed), and S-III (Proliferative and Invasive). S-I had more favorable OS and DFS, whereas S-III had the worst prognosis and was enriched with OC patients at advanced stages. Meanwhile, comprehensive functional analysis highlighted differences in biological pathways: genes associated with immune function and DNA damage repair including CXCL9, CXCL10, CXCL11, APEX, APEX2, and RBX1 were enriched in S-I; S-II combined multiple gene signatures including genes associated with metabolism and transcription; and the gene signature of S-III was extensively involved in pathways reflecting malignancies, including many core kinases and transcription factors involved in cancer such as CDK6, ERBB2, JAK1, DAPK1, FOXO1, and RXRA. The SVM model showed superior diagnostic performance with AUC values of 0.922 and 0.901, respectively. Furthermore, a new dataset of the independent cohort could be automatically analyzed by this innovative pipeline and yield similar results.

Conclusion

This study exploited an innovative approach to construct previously unexplored robust subtypes significantly related to different clinical and molecular features for OC and a diagnostic model using SVM to aid in clinical diagnosis and treatment. This investigation also illustrated the importance of targeting innate immune suppression together with DNA damage in OC, offering novel insights for further experimental exploration and clinical trial.

Development of genomic phenotype and immunophenotype of acute respiratory distress syndrome using autophagy and metabolism-related genes

Background

Distinguishing ARDS phenotypes is of great importance for its precise treatment. In the study, we attempted to ascertain its phenotypes based on metabolic and autophagy-related genes and infiltrated immune cells.

Methods

Transcription datasets of ARDS patients were obtained from Gene expression omnibus (GEO), autophagy and metabolic-related genes were from the Human Autophagy Database and the GeneCards Database, respectively. Autophagy and metabolism-related differentially expressed genes (AMRDEGs) were further identified by machine learning and processed for constructing the nomogram and the risk prediction model. Functional enrichment analyses of differentially expressed genes were performed between high- and low-risk groups. According to the protein-protein interaction network, these hub genes closely linked to increased risk of ARDS were identified with CytoHubba. ssGSEA and CIBERSORT was applied to analyze the infiltration pattern of immune cells in ARDS. Afterwards, immunologically characterized and molecular phenotypes were constructed according to infiltrated immune cells and hub genes.

Results

A total of 26 AMRDEGs were obtained, and CTSB and EEF2 were identified as crucial AMRDEGs. The predictive capability of the risk score, calculated based on the expression levels of CTSB and EEF2, was robust for ARDS in both the discovery cohort (AUC = 1) and the validation cohort (AUC = 0.826). The mean risk score was determined to be 2.231332, and based on this score, patients were classified into high-risk and low-risk groups. 371 differential genes in high- and low-risk groups were analyzed. ITGAM, TYROBP, ITGB2, SPI1, PLEK, FGR, MPO, S100A12, HCK, and MYC were identified as hub genes. A total of 12 infiltrated immune cells were differentially expressed and have correlations with hub genes. According to hub genes and implanted immune cells, ARDS patients were divided into two different molecular phenotypes (Group 1: n = 38; Group 2: n = 19) and two immune phenotypes (Cluster1: n = 22; Cluster2: n = 35), respectively.

Conclusion

This study picked up hub genes of ARDS related to autophagy and metabolism and clustered ARDS patients into different molecular phenotypes and immunophenotypes, providing insights into the precision medicine of treating patients with ARDS.

Machine learning-based analysis of risk factors for chronic total occlusion in an Asian population

OBJECTIVES

Chronic total occlusion (CTO) is a form of coronary artery disease (CAD) requiring percutaneous coronary intervention. There has been minimal research regarding CTO-specific risk factors and predictive models. We developed machine learning predictive models based on clinical characteristics to identify patients with CTO before coronary angiography.

METHODS

Data from 1473 patients with CAD, including 317 patients with and 1156 patients without CTO, were retrospectively analyzed. Partial least squares discriminant analysis (PLS-DA), random forest (RF), and support vector machine (SVM) models were used to identify CTO-specific risk factors and predict CTO development. Receiver operating characteristic (ROC) curve analysis was performed for model validation.

RESULTS

For CTO prediction, the PLS-DA model included 10 variables; the ROC value was 0.706. The RF model included 42 variables; the ROC value was 0.702. The SVM model included 20 variables; the ROC value was 0.696. DeLong's test showed no difference among the three models. Four variables were present in all models: sex, neutrophil percentage, creatinine, and brain natriuretic peptide (BNP).

CONCLUSIONS

Validation of machine learning prediction models for CTO revealed that the PLS-DA model had the best prediction performance. Sex, neutrophil percentage, creatinine, and BNP may be important risk factors for CTO development.

Machine Learning Algorithms Applied to Predict Autism Spectrum Disorder Based on Gut Microbiome Composition

The application of machine learning (ML) techniques stands as a reliable method for aiding in the diagnosis of complex diseases. Recent studies have related the composition of the gut microbiota to the presence of autism spectrum disorder (ASD), but until now, the results have been mostly contradictory. This work proposes using machine learning to study the gut microbiome composition and its role in the early diagnosis of ASD. We applied support vector machines (SVMs), artificial neural networks (ANNs), and random forest (RF) algorithms to classify subjects as neurotypical (NT) or having ASD, using published data on gut microbiome composition. Naive Bayes, k-nearest neighbors, ensemble learning, logistic regression, linear regression, and decision trees were also trained and validated; however, the ones presented showed the best performance and interpretability. All the ML methods were developed using the SAS Viya software platform. The microbiome's composition was determined using 16S rRNA sequencing technology. The application of ML yielded a classification accuracy as high as 90%, with a sensitivity of 96.97% and specificity reaching 85.29%. In the case of the ANN model, no errors occurred when classifying NT subjects from the first dataset, indicating a significant classification outcome compared to traditional tests and data-based approaches. This approach was repeated with two datasets, one from the USA and the other from China, resulting in similar findings. The main predictors in the obtained models differ between the analyzed datasets. The most important predictors identified from the analyzed datasets are Bacteroides, Lachnospira, Anaerobutyricum, and Ruminococcus torques. Notably, among the predictors in each model, there is the presence of bacteria that are usually considered insignificant in the microbiome's composition due to their low relative abundance. This outcome reinforces the conventional understanding of the microbiome's influence on ASD development, where an imbalance in the composition of the microbiota can lead to disrupted host-microbiota homeostasis. Considering that several previous studies focused on the most abundant genera and neglected smaller (and frequently not statistically significant) microbial communities, the impact of such communities has been poorly analyzed. The ML-based models suggest that more research should focus on these less abundant microbes. A novel hypothesis explains the contradictory results in this field and advocates for more in-depth research to be conducted on variables that may not exhibit statistical significance. The obtained results seem to contribute to an explanation of the contradictory findings regarding ASD and its relation with gut microbiota composition. While some research correlates higher ratios of Bacillota/Bacteroidota, others find the opposite. These discrepancies are closely linked to the minority organisms in the microbiome's composition, which may differ between populations but share similar metabolic functions. Therefore, the ratios of Bacillota/Bacteroidota regarding ASD may not be determinants in the manifestation of ASD.

Combining WGCNA and machine learning to construct immune-related EMT patterns to predict HCC prognosis and immune microenvironment

Hepatocellular carcinoma (HCC) is a malignancy with a very high mortality rate. Because of its high heterogeneity, there is an urgent need to find biomarkers that accurately predict prognosis. Epithelial-mesenchymal transition (EMT) is closely associated with frequent recurrence and high mortality of HCC. Therefore, it is necessary to comprehensively analyze the prognostic value and immunological properties of EMT gene in HCC. In our study, we performed bioinformatics analysis of the TCGA and ICGC liver cancer cohorts and identified the module genes of immune-associated EMTs (iEMT) by Weighted Gene Co-Expression Network Analysis (WGCNA). Further we used machine learning (support vector machines-recursive feature elimination and Lasso) to identify three central iEMT genes (ARMC9, ADAM15 and STC2) and construct iEMT_score. Subsequently, in the training and validation cohorts, it was demonstrated that the overall survival (OS) of patients in the high iEMT_score group was worse than that of patients in the low iEMT_score group. Based on this, we have constructed a nomogram that is easy for clinicians to use. In addition, our study explored differences in pathway enrichment, immunological properties, and sensitivity to common chemotherapy and targeted drugs in different subgroups of iEMT_score. Finally, we showed through in vitro experiments that knockdown of ARMC9 could significantly inhibit the proliferation, migration and invasion of HCC cells BEL7402. Taken together, our findings suggest that iEMT_score is an excellent biomarker for predicting prognosis and provide some new insights for personalized treatment of HCC patients.

Identification of aging-related biomarkers and immune infiltration characteristics in osteoarthritis based on bioinformatics analysis and machine learning

Background

Osteoarthritis (OA) is a degenerative disease closely related to aging. Nevertheless, the role and mechanisms of aging in osteoarthritis remain unclear. This study aims to identify potential aging-related biomarkers in OA and to explore the role and mechanisms of aging-related genes and the immune microenvironment in OA synovial tissue.

Methods

Normal and OA synovial gene expression profile microarrays were obtained from the Gene Expression Omnibus (GEO) database and aging-related genes (ARGs) from the Human Aging Genomic Resources database (HAGR). Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), Disease Ontology (DO), and Gene set variation analysis (GSVA) enrichment analysis were used to uncover the underlying mechanisms. To identify Hub ARDEGs with highly correlated OA features (Hub OA-ARDEGs), Weighted Gene Co-expression Network Analysis (WGCNA) and machine learning methods were used. Furthermore, we created diagnostic nomograms and receiver operating characteristic curves (ROC) to assess Hub OA-ARDEGs' ability to diagnose OA and predict which miRNAs and TFs they might act on. The Single sample gene set enrichment analysis (ssGSEA) algorithm was applied to look at the immune infiltration characteristics of OA and their relationship with Hub OA-ARDEGs.

Results

We discovered 87 ARDEGs in normal and OA synovium samples. According to functional enrichment, ARDEGs are primarily associated with inflammatory regulation, cellular stress response, cell cycle regulation, and transcriptional regulation. Hub OA-ARDEGs with excellent OA diagnostic ability were identified as MCL1, SIK1, JUND, NFKBIA, and JUN. Wilcox test showed that Hub OA-ARDEGs were all significantly downregulated in OA and were validated in the validation set and by qRT-PCR. Using the ssGSEA algorithm, we discovered that 15 types of immune cell infiltration and six types of immune cell activation were significantly increased in OA synovial samples and well correlated with Hub OA-ARDEGs.

Conclusion

Synovial aging may promote the progression of OA by inducing immune inflammation. MCL1, SIK1, JUND, NFKBIA, and JUN can be used as novel diagnostic biomolecular markers and potential therapeutic targets for OA.

Machine learning and bioinformatics framework integration reveal potential characteristic genes related to immune cell infiltration in preeclampsia

Introduction: Preeclampsia is a disease that affects both the mother and child, with serious consequences. Screening the characteristic genes of preeclampsia and studying the placental immune microenvironment are expected to explore specific methods for the treatment of preeclampsia and gain an in-depth understanding of the pathological mechanism of preeclampsia. Methods: We screened for differential genes in preeclampsia by using limma package. Gene Ontology, Kyoto Encyclopedia of Genes and Genomes, disease ontology enrichment, and gene set enrichment analyses were performed. Analysis and identification of preeclampsia biomarkers were performed by using the least absolute shrinkage and selection operator regression model, support vector machine recursive feature elimination, and random forest algorithm. The CIBERSORT algorithm was used to analyze immune cell infiltration. The characteristic genes were verified by RT-qPCR. Results: We identified 73 differential genes, which mainly involved in reproductive structure and system development, hormone transport, etc. KEGG analysis revealed emphasis on cytokine-cytokine receptor interactions and interleukin-17 signaling pathways. Differentially expressed genes were dominantly concentrated in endocrine system diseases and reproductive system diseases. Our findings suggest that LEP, SASH1, RAB6C, and FLT1 can be used as placental markers for preeclampsia and they are associated with various immune cells. Conclusion: The differentially expressed genes in preeclampsia are related to inflammatory response and other pathways. Characteristic genes, LEP, SASH1, RAB6C, and FLT1 can be used as diagnostic and therapeutic targets for preeclampsia, and they are associated with immune cell infiltration. Our findings contribute to the pathophysiological mechanism exploration of preeclampsia. In the future, the sample size needs to be expanded for data analysis and validation, and the immune cells need to be further validated.

Machine learning in TCM with natural products and molecules: current status and future perspectives

Traditional Chinese medicine (TCM) has been practiced for thousands of years with clinical efficacy. Natural products and their effective agents such as artemisinin and paclitaxel have saved millions of lives worldwide. Artificial intelligence is being increasingly deployed in TCM. By summarizing the principles and processes of deep learning and traditional machine learning algorithms, analyzing the application of machine learning in TCM, reviewing the results of previous studies, this study proposed a promising future perspective based on the combination of machine learning, TCM theory, chemical compositions of natural products, and computational simulations based on molecules and chemical compositions. In the first place, machine learning will be utilized in the effective chemical components of natural products to target the pathological molecules of the disease which could achieve the purpose of screening the natural products on the basis of the pathological mechanisms they target. In this approach, computational simulations will be used for processing the data for effective chemical components, generating datasets for analyzing features. In the next step, machine learning will be used to analyze the datasets on the basis of TCM theories such as the superposition of syndrome elements. Finally, interdisciplinary natural product-syndrome research will be established by unifying the results of the two steps outlined above, potentially realizing an intelligent artificial intelligence diagnosis and treatment model based on the effective chemical components of natural products under the guidance of TCM theory. This perspective outlines an innovative application of machine learning in the clinical practice of TCM based on the investigation of chemical molecules under the guidance of TCM theory.

Machine learning with neuroimaging biomarkers: Application in the diagnosis and prediction of drug addiction

Drug abuse is a serious problem worldwide. Owing to intermittent intake of certain substances and the early inconspicuous clinical symptoms, this brings huge challenges for timely diagnosing addiction status and preventing substance use disorders (SUDs). As a non-invasive technique, neuroimaging can capture neurobiological signatures of abnormality in multiple brain regions caused by drug consumption in each clinical stage, like parenchymal morphology alteration as well as aberrant functional activity and connectivity of cerebral areas, making it realizable to diagnosis, prediction and even preemptive therapy of addiction. Machine learning (ML) algorithms primarily used for classification have been extensively applied in analysing medical imaging datasets. Significant neurobiological characteristics employed and revealed by classifiers were used to diagnose addictive states and predict initiation and vulnerability to drug usage, treatment abstinence, relapse and resilience of addicts and the risk of SUD. In this review, we summarize application of ML methods in neuroimaging focusing on addicts' diagnosis of clinical status and risk prediction and elucidate the discriminative neurobiological features from brain electrophysiological, morphological and functional perspectives that contribute most to the classifier, finally highlighting the auxiliary role of ML in addiction treatment.

Application of machine learning in the prediction of deficient mismatch repair in patients with colorectal cancer based on routine preoperative characterization

Simple summary

Detecting deficient mismatch repair (dMMR) in patients with colorectal cancer is essential for clinical decision-making, including evaluation of prognosis, guidance of adjuvant chemotherapy and immunotherapy, and primary screening for Lynch syndrome. However, outside of tertiary care centers, existing detection methods are not widely disseminated and highly depend on the experienced pathologist. Therefore, it is of great clinical significance to develop a broadly accessible and low-cost tool for dMMR prediction, particularly prior to surgery. In this study, we developed a convenient and reliable model for predicting dMMR status in CRC patients on routine preoperative characterization utilizing multiple machine learning algorithms. This model will work as an automated screening tool for identifying patients suitable for mismatch repair testing and consequently for improving the detection rate of dMMR, while reducing unnecessary labor and cost in patients with proficient mismatch repair.

Background

Deficient mismatch repair (dMMR) indicates a sustained anti-tumor immune response and has a favorable prognosis in patients with colorectal cancer (CRC). Although all CRC patients are recommended to undergo dMMR testing after surgery, current diagnostic approaches are not available for all country hospitals and patients. Therefore, efficient and low-cost predictive models for dMMR, especially for preoperative evaluations, are warranted.

Methods

A large scale of 5596 CRC patients who underwent surgical resection and mismatch repair testing were enrolled and randomly divided into training and validation cohorts. The clinical features exploited for predicting dMMR comprised the demographic characteristics, preoperative laboratory data, and tumor burden information. Machine learning (ML) methods involving eight basic algorithms, ensemble learning methods, and fusion algorithms were adopted with 10-fold cross-validation, and their performance was evaluated based on the area under the receiver operating characteristic curve (AUC) and calibration curves. The clinical net benefits were assessed using a decision curve analysis (DCA), and a nomogram was developed to facilitate model clinical practicality.

Results

All models achieved an AUC of nearly 0.80 in the validation cohort, with the stacking model exhibiting the best performance (AUC = 0.832). Logistical DCA revealed that the stacking model yielded more clinical net benefits than the conventional regression models. In the subgroup analysis, the stacking model also predicted dMMR regardless of the clinical stage. The nomogram showed a favorable consistence with the actual outcome in the calibration curve.

Conclusion

With the aid of ML algorithms, we developed a novel and robust model for predicting dMMR in CRC patients with satisfactory discriminative performance and designed a user-friendly and convenient nomogram.

Common molecular mechanism and immune infiltration patterns of thoracic and abdominal aortic aneurysms

BACKGROUND

Aortic disease (aortic aneurysm (AA), dissection (AD)) is a serious threat to patient lives. Little is currently known about the molecular mechanisms and immune infiltration patterns underlying the development and progression of thoracic and abdominal aortic aneurysms (TAA and AAA), warranting further research.

METHODS

We downloaded AA (includes TAA and AAA) datasets from the GEO database. The potential biomarkers in TAA and AAA were identified using differential expression analysis and two machine-learning algorithms. The discrimination power of the potential biomarkers and their diagnostic accuracy was assessed in validation datasets using ROC curve analysis. Then, GSEA, KEGG, GO and DO analyses were conducted. Furthermore, two immuno-infiltration analysis algorithms were utilized to analyze the common immune infiltration patterns in TAA and AAA. Finally, a retrospective clinical study was performed on 78 patients with AD, and the serum from 6 patients was used for whole exome sequencing (WES).

RESULTS

The intersection of TAA and AAA datasets yielded 82 differentially expressed genes (DEGs). Subsequently, the biomarkers (CX3CR1 and HBB) were acquired by screening using two machine-learning algorithms and ROC curve analysis. The functional analysis of DEGs showed significant enrichment in inflammation and regulation of angiogenic pathways. Immune cell infiltration analysis revealed that adaptive and innate immune responses were closely linked to AA progression. However, neither CX3CR1 nor HBB was associated with B cell-mediated humoral immunity. CX3CR1 expression was correlated with macrophages and HBB with eosinophils. Finally, our retrospective clinical study revealed a hyperinflammatory environment in aortic disease. The WES study identified disease biomarkers and gene variants, some of which may be druggable.

CONCLUSION

The genes CX3CR1 and HBB can be used as common biomarkers in TAA and AAA. Large numbers of innate and adaptive immune cells are infiltrated in AA and are closely linked to the development and progression of AA. Moreover, CX3CR1 and HBB are highly correlated with the infiltration of immune cells and may be potential targets of immunotherapeutic drugs. Gene mutation research is a promising direction for the treatment of aortic disease.

Development of a Novel Burned-Area Subpixel Mapping (BASM) Workflow for Fire Scar Detection at Subpixel Level

The accurate detection of burned forest area is essential for post-fire management and assessment, and for quantifying carbon budgets. Therefore, it is imperative to map burned areas accurately. Currently, there are few burned-area products around the world. Researchers have mapped burned areas directly at the pixel level that is usually a mixture of burned area and other land cover types. In order to improve the burned area mapping at subpixel level, we proposed a Burned Area Subpixel Mapping (BASM) workflow to map burned areas at the subpixel level. We then applied the workflow to Sentinel 2 data sets to obtain burned area mapping at subpixel level. In this study, the information of true fire scar was provided by the Department of Emergency Management of Hunan Province, China. To validate the accuracy of the BASM workflow for detecting burned areas at the subpixel level, we applied the workflow to the Sentinel 2 image data and then compared the detected burned area at subpixel level with in situ measurements at fifteen fire-scar reference sites located in Hunan Province, China. Results show the proposed method generated successfully burned area at the subpixel level. The methods, especially the BASM-Feature Extraction Rule Based (BASM-FERB) method, could minimize misclassification and effects due to noise more effectively compared with the BASM-Random Forest (BASM-RF), BASM-Backpropagation Neural Net (BASM-BPNN), BASM-Support Vector Machine (BASM-SVM), and BASM-notra methods. We conducted a comparison study among BASM-FERB, BASM-RF, BASM-BPNN, BASM-SVM, and BASM-notra using five accuracy evaluation indices, i.e., overall accuracy (OA), user’s accuracy (UA), producer’s accuracy (PA), intersection over union (IoU), and Kappa coefficient (Kappa). The detection accuracy of burned area at the subpixel level by BASM-FERB’s OA, UA, IoU, and Kappa is 98.11%, 81.72%, 74.32%, and 83.98%, respectively, better than BASM-RF’s, BASM-BPNN’s, BASM-SVM’s, and BASM-notra’s, even though BASM-RF’s and BASM-notra’s average PA is higher than BASM-FERB’s, with 89.97%, 91.36%, and 89.52%, respectively. We conclude that the newly proposed BASM workflow can map burned areas at the subpixel level, providing greater accuracy in regards to the burned area for post-forest fire management and assessment.

Artificial intelligence and machine-learning approaches in structure and ligand-based discovery of drugs affecting central nervous system

CNS disorders are indications with a very high unmet medical needs, relatively smaller number of available drugs, and a subpar satisfaction level among patients and caregiver. Discovery of CNS drugs is extremely expensive affair with its own unique challenges leading to extremely high attrition rates and low efficiency. With explosion of data in information age, there is hardly any aspect of life that has not been touched by data driven technologies such as artificial intelligence (AI) and machine learning (ML). Drug discovery is no exception, emergence of big data via genomic, proteomic, biological, and chemical technologies has driven pharmaceutical giants to collaborate with AI oriented companies to revolutionise drug discovery, with the goal of increasing the efficiency of the process. In recent years many examples of innovative applications of AI and ML techniques in CNS drug discovery has been reported. Research on therapeutics for diseases such as schizophrenia, Alzheimer's and Parkinsonism has been provided with a new direction and thrust from these developments. AI and ML has been applied to both ligand-based and structure-based drug discovery and design of CNS therapeutics. In this review, we have summarised the general aspects of AI and ML from the perspective of drug discovery followed by a comprehensive coverage of the recent developments in the applications of AI/ML techniques in CNS drug discovery.

Diagnostic classification of cancers using DNA methylation of paracancerous tissues

The potential role of DNA methylation from paracancerous tissues in cancer diagnosis has not been explored until now. In this study, we built classification models using well-known machine learning models based on DNA methylation profiles of paracancerous tissues. We evaluated our methods on nine cancer datasets collected from The Cancer Genome Atlas (TCGA) and utilized fivefold cross-validation to assess the performance of models. Additionally, we performed gene ontology (GO) enrichment analysis on the basis of the significant CpG sites selected by feature importance scores of XGBoost model, aiming to identify biological pathways involved in cancer progression. We also exploited the XGBoost algorithm to classify cancer types using DNA methylation profiles of paracancerous tissues in external validation datasets. Comparative experiments suggested that XGBoost achieved better predictive performance than the other four machine learning methods in predicting cancer stage. GO enrichment analysis revealed key pathways involved, highlighting the importance of paracancerous tissues in cancer progression. Furthermore, XGBoost model can accurately classify nine different cancers from TCGA, and the feature sets selected by XGBoost can also effectively predict seven cancer types on independent GEO datasets. This study provided new insights into cancer diagnosis from an epigenetic perspective and may facilitate the development of personalized diagnosis and treatment strategies.

Prediction of Bronchopneumonia Inpatients' Total Hospitalization Expenses Based on BP Neural Network and Support Vector Machine Models

Objective

BP neural network (BPNN) model and support vector machine (SVM) model were used to predict the total hospitalization expenses of patients with bronchopneumonia.

Methods

A total of 355 patients with bronchopneumonia from January 2018 to December 2020 were collected and sorted out. The data set was randomly divided into a training set (n = 249) and a test set (n = 106) according to 7 : 3. The BPNN model and SVM model were constructed to analyze the predictors of total hospitalization expenses. The effectiveness was compared between these two prediction models.

Results

The top three influencing factors and their importance for predicting total hospitalization cost by the BPNN model were hospitalization days (0.477), age (0.154), and discharge department (0.083). The top 3 factors predicted by the SVM model were hospitalization days (0.215), age (0.196), and marital status (0.172). The area under the curve of these two models is 0.838 (95% CI: 0.755~0.921) and 0.889 (95% CI: 0.819~0.959), respectively.

Conclusion

Both the BPNN model and SVM model can predict the total hospitalization expenses of patients with bronchopneumonia, but the prediction effect of the SVM model is better than the BPNN model.

A Machine-Learning Approach to Developing a Predictive Signature Based on Transcriptome Profiling of Ground-Glass Opacities for Accurate Classification and Exploring the Immune Microenvironment of Early-Stage LUAD

Screening for early-stage lung cancer with low-dose computed tomography is recommended for high-risk populations; consequently, the incidence of pure ground-glass opacity (pGGO) is increasing. Ground-glass opacity (GGO) is considered the appearance of early lung cancer, and there remains an unmet clinical need to understand the pathology of small GGO (<1 cm in diameter). The objective of this study was to use the transcriptome profiling of pGGO specimens <1 cm in diameter to construct a pGGO-related gene risk signature to predict the prognosis of early-stage lung adenocarcinoma (LUAD) and explore the immune microenvironment of GGO. pGGO-related differentially expressed genes (DEGs) were screened to identify prognostic marker genes with two machine learning algorithms. A 15-gene risk signature was constructed from the DEGs that were shared between the algorithms. Risk scores were calculated using the regression coefficients for the pGGO-related DEGs. Patients with Stage I/II LUAD or Stage IA LUAD and high-risk scores had a worse prognosis than patients with low-risk scores. The prognosis of high-risk patients with Stage IA LUAD was almost identical to that of patients with Stage II LUAD, suggesting that treatment strategies for patients with Stage II LUAD may be beneficial in high-risk patients with Stage IA LUAD. pGGO-related DEGs were mainly enriched in immune-related pathways. Patients with high-risk scores and high tumor mutation burden had a worse prognosis and may benefit from immunotherapy. A nomogram was constructed to facilitate the clinical application of the 15-gene risk signature. Receiver operating characteristic curves and decision curve analysis validated the predictive ability of the nomogram in patients with Stage I LUAD in the TCGA-LUAD cohort and GEO datasets.

Machine Learning Based on Diffusion Kurtosis Imaging Histogram Parameters for Glioma Grading

Glioma grading plays an important role in surgical resection. We investigated the ability of different feature reduction methods in support vector machine (SVM)-based diffusion kurtosis imaging (DKI) histogram parameters to distinguish glioma grades. A total of 161 glioma patients who underwent magnetic resonance imaging (MRI) from January 2017 to January 2020 were included retrospectively. The patients were divided into low-grade (n = 61) and high-grade (n = 100) groups. Parametric DKI maps were derived, and 45 features from the DKI maps were extracted semi-automatically for analysis. Three feature selection methods [principal component analysis (PCA), recursive feature elimination (RFE) and least absolute shrinkage and selection operator (LASSO)] were used to establish the glioma grading model with an SVM classifier. To evaluate the performance of SVM models, the receiver operating characteristic (ROC) curves of SVM models for distinguishing glioma grades were compared with those of conventional statistical methods. The conventional ROC analysis showed that mean diffusivity (MD) variance, MD skewness and mean kurtosis (MK) C50 could effectively distinguish glioma grades, particularly MD variance. The highest classification distinguishing AUC was found using LASSO at 0.904 ± 0.069. In comparison, classification AUC by PCA was 0.866 ± 0.061, and 0.899 ± 0.079 by RFE. The SVM-PCA model with the lowest AUC among the SVM models was significantly better than the conventional ROC analysis (z = 1.947, p = 0.013). These findings demonstrate the superiority of DKI histogram parameters by LASSO analysis and SVM for distinguishing glioma grades.

Revealing Potential Diagnostic Gene Biomarkers Associated with Immune Infiltration in Patients with Renal Fibrosis Based on Machine Learning Analysis

Chronic kidney disease is characterized by the development of renal fibrosis. The basic mechanisms of renal fibrosis have not yet been fully investigated despite significant progress in understanding the etiology of the disease. In this work, the researchers sought to identify potential diagnostic indicators for renal fibrosis. From the GEO database, we were able to acquire two gene expression profiles with publically available data (GSE22459 and GSE76882, respectively) from human renal fibrosis and control samples. 215 renal fibrosis specimens and 124 normal specimens were examined for differentially expressed genes (DEGs). The SVM-RFE and LASSO regression models were used to discover potential markers. CIBERSORT was applied to estimate the combined cohorts' immune cell fraction compositional trends in renal fibrosis. RT-PCR was used to examine the expression of ISG20 in renal fibrosis and healthy samples. In vitro experiments were applied to examine the function of ISG20 knockdown on the progression of renal fibrosis. In this study, we identified 24 DEGs. The result of LASSO and SVM-RFE identified nine critical genes. ROC assays confirmed the diagnostic value of the above nine genes for renal fibrosis. Immune cell infiltration analysis revealed that ISG20 and SERPINA3 were both found to be correlated with T cell follicular helper, neutrophils, T cell CD4 memory activated, eosinophils, T cell CD8, dendritic cell activated, B cell memory, monocytes, macrophage M2, plasma cells, T cell CD4 naïve, mast cell resting, B cell naïve, T cell regulatory, and NK cell activated. Finally, we observed that the expression of ISG20 and SERPINA3 was distinctly increased in renal fibrosis samples compared with normal samples. ISG20 siRNA significantly suppressed the progression of renal fibrosis in vitro. Overall, this study identified nine diagnostic biomarkers for renal fibrosis. ISG20 may be a novel therapeutic target of renal fibrosis.

Exhaled volatile organic compounds for diagnosis of hepatocellular carcinoma

Volatile organic compounds (VOCs) profile for diagnosis and monitoring therapeutic response of hepatocellular carcinoma (HCC) has not been well studied. We determined VOCs profile in exhaled breath of 97 HCC patients and 111 controls using gas chromatography–mass spectrometry and Support Vector Machine algorithm. The combination of acetone, 1,4-pentadiene, methylene chloride, benzene, phenol and allyl methyl sulfide provided the highest accuracy of 79.6%, with 76.5% sensitivity and 82.7% specificity in the training set; and 55.4% accuracy, 44.0% sensitivity, and 75.0% specificity in the test set. This combination was correlated with the HCC stages demonstrating by the increased distance from the classification boundary when the stage advanced. For early HCC detection, d-limonene provided a 62.8% sensitivity, 51.8% specificity and 54.9% accuracy. The levels of acetone, butane and dimethyl sulfide were significantly altered after treatment. Patients with complete response had a greater decreased acetone level than those with remaining tumor post-treatment (73.38 ± 56.76 vs. 17.11 ± 58.86 (× 10⁶ AU, p = 0.006). Using a cutoff of 35.9 × 10⁶ AU, the reduction in acetone level predicted treatment response with 77.3% sensitivity, 83.3% specificity, 79.4%, accuracy, and AUC of 0.784. This study demonstrates the feasibility of exhaled VOCs as a non-invasive tool for diagnosis, monitoring of HCC progression and treatment response.

Classifiers for Predicting Coronary Artery Disease Based on Gene Expression Profiles in Peripheral Blood Mononuclear Cells

Objective

Coronary artery disease (CAD) is a serious global health concern. Current diagnostic methods for CAD involve risk to the patient and are costly, so better diagnostic tools are needed. We defined four classifiers based on gene expression profiles in peripheral blood mononuclear cells and determined their potential for CAD detection.

Methods

We downloaded a CAD-related data set (GSE113079) from the Gene Expression Omnibus (GEO) database. We identified differentially expressed genes (DEGs) in peripheral blood mononuclear cells between CAD samples and healthy controls. DEGs were analyzed for functional enrichment. To create a robust CAD classifier, DEGs were identified by feature selection using the principal component analysis. Then, least absolute shrinkage and selection operator (LASSO) logistic regression, random forest, and support vector machine (SVM) models were created. Gene set variation analysis (GSVA) score and gene set enrichment analysis (GSEA) were also conducted. The performance of the models was evaluated in terms of the area under receiver operating characteristic curves (AUC).

Results

In the training set, we found 135 up-regulated genes and 104 down-regulated genes in CAD patients compared with controls. The DEGs were involved in some pathways associated with CAD, such as pathways involving calcium and interleukin-17 signaling. Twenty genes were identified as optimal features and used to generate the logistic classifier based on LASSO. The AUC for the classifier was 1.00 in the training set and 0.997 in the test set. Using the 20 DEGs, SVM and random forest classifiers were also generated and showed high diagnostic efficacy, with respective AUCs of 0.997 and 1.00 against the training set. A GSVA score was also established using the top 20 significant DEGs, which showed an AUC of 0.971 in the training set and 0.989 in the test set. Furthermore, GSEA showed autophagy and the proteasome to be major pathways involving the DEGs.

Conclusion

We identified a set of genes specific for CAD whose expression can be measured non-invasively. Using these genes, we defined four diagnostic classifiers using multiple methods.

Development and Validation of a Prediction Model for Elevated Arterial Stiffness in Chinese Patients With Diabetes Using Machine Learning

Background

Arterial stiffness assessed by pulse wave velocity is a major risk factor for cardiovascular diseases. The incidence of cardiovascular events remains high in diabetics. However, a clinical prediction model for elevated arterial stiffness using machine learning to identify subjects consequently at higher risk remains to be developed.

Methods

Least absolute shrinkage and selection operator and support vector machine-recursive feature elimination were used for feature selection. Four machine learning algorithms were used to construct a prediction model, and their performance was compared based on the area under the receiver operating characteristic curve metric in a discovery dataset (n = 760). The model with the best performance was selected and validated in an independent dataset (n = 912) from the Dryad Digital Repository (https://doi.org/10.5061/dryad.m484p). To apply our model to clinical practice, we built a free and user-friendly web online tool.

Results

The predictive model includes the predictors: age, systolic blood pressure, diastolic blood pressure, and body mass index. In the discovery cohort, the gradient boosting-based model outperformed other methods in the elevated arterial stiffness prediction. In the validation cohort, the gradient boosting model showed a good discrimination capacity. A cutoff value of 0.46 for the elevated arterial stiffness risk score in the gradient boosting model resulted in a good specificity (0.813 in the discovery data and 0.761 in the validation data) and sensitivity (0.875 and 0.738, respectively) trade-off points.

Conclusion

The gradient boosting-based prediction system presents a good classification in elevated arterial stiffness prediction. The web online tool makes our gradient boosting-based model easily accessible for further clinical studies and utilization.

Development and Validation of a Prediction Model for Elevated Arterial Stiffness in Chinese Patients With Diabetes Using Machine Learning

Background

Arterial stiffness assessed by pulse wave velocity is a major risk factor for cardiovascular diseases. The incidence of cardiovascular events remains high in diabetics. However, a clinical prediction model for elevated arterial stiffness using machine learning to identify subjects consequently at higher risk remains to be developed.

Methods

Least absolute shrinkage and selection operator and support vector machine-recursive feature elimination were used for feature selection. Four machine learning algorithms were used to construct a prediction model, and their performance was compared based on the area under the receiver operating characteristic curve metric in a discovery dataset (n = 760). The model with the best performance was selected and validated in an independent dataset (n = 912) from the Dryad Digital Repository (https://doi.org/10.5061/dryad.m484p). To apply our model to clinical practice, we built a free and user-friendly web online tool.

Results

The predictive model includes the predictors: age, systolic blood pressure, diastolic blood pressure, and body mass index. In the discovery cohort, the gradient boosting-based model outperformed other methods in the elevated arterial stiffness prediction. In the validation cohort, the gradient boosting model showed a good discrimination capacity. A cutoff value of 0.46 for the elevated arterial stiffness risk score in the gradient boosting model resulted in a good specificity (0.813 in the discovery data and 0.761 in the validation data) and sensitivity (0.875 and 0.738, respectively) trade-off points.

Conclusion

The gradient boosting-based prediction system presents a good classification in elevated arterial stiffness prediction. The web online tool makes our gradient boosting-based model easily accessible for further clinical studies and utilization.

Recent advances on the machine learning methods in predicting ncRNA-protein interactions

Recent transcriptomics and bioinformatics studies have shown that ncRNAs can affect chromosome structure and gene transcription, participate in the epigenetic regulation, and take part in diseases such as tumorigenesis. Biologists have found that most ncRNAs usually work by interacting with the corresponding RNA-binding proteins. Therefore, ncRNA-protein interaction is a very popular study in both the biological and medical fields. However, due to the limitations of manual experiments in the laboratory, machine-learning methods for predicting ncRNA-protein interactions are increasingly favored by the researchers. In this review, we summarize several machine learning predictive models of ncRNA-protein interactions over the past few years, and briefly describe the characteristics of these machine learning models. In order to optimize the performance of machine learning models to better predict ncRNA-protein interactions, we give some promising future computational directions at the end.

A novel artificial intelligence protocol to investigate potential leads for Parkinson's disease

Previous studies have shown that small molecule inhibitors of NLRP3 may be a potential treatment for Parkinson's disease (PD). NACHT, LRR and PYD domains-containing protein 3 (NLRP3), heat shock protein HSP 90-beta (HSP90AB1), caspase-1 (CASP1) and cellular tumor antigen p53 (TP53) have significant involvement in the pathogenesis pathway of PD. Molecular docking was used to screen the traditional Chinese medicine database TCM Database@Taiwan. Top traditional Chinese medicine (TCM) compounds with high affinities based on Dock Score were selected to form the drug-target interaction network to investigate potential candidates targeting NLRP3, HSP90AB1, CASP1, and TP53 proteins. Artificial intelligence model, 3D-Quantitative Structure-Activity Relationship (3D-QSAR) were constructed respectively utilizing training sets of inhibitors against the four proteins with known inhibitory activities (pIC50). The results showed that 2007_22057 (an indole derivative), 2007_22325 (a valine anhydride) and 2007_15317 (an indole derivative) might be a potential medicine formula for the treatment of PD. Then there are three candidate compounds identified by the result of molecular dynamics.

Diagnostic Value of Plasma MicroRNAs for Lung Cancer Using Support Vector Machine Model

Aim: Small single-stranded non-coding RNAs (miRNAs) play an important role in carcinogenesis through degrading target mRNAs. However, the diagnostic value of miRNAs was not explored in lung cancers. In this study, a support-vector-machine (SVM) model for diagnosis of lung cancer was established based on plasma miRNAs biomarkers, clinical symptoms and epidemiology material.

Methods: The expressions of plasma miRNA were examined with SYBR Green-based quantitative real-time PCR.

Results: We identified that the expressions of 10 plasma miRNAs (miR-21, miR-20a, miR-210, miR-145, miR-126, miR-223, miR-197, miR-30a, miR-30d, miR-25), smoking status, fever, cough, chest pain or tightness, bloody phlegm, haemoptysis, were significantly different between lung cancer and control groups (P<0.05). The accuracies of the combined SVM, miRNAs SVM, symptom SVM, combined Fisher, miRNAs Fisher and symptom Fisher were 96.34%, 80.49%, 84.15%, 84.15%, 75.61%, and 80.49%, respectively; AUC of these six model were 0.976, 0.841, 0.838, 0.865, 0.750, and 0.801, respectively. The accuracy and AUC of combined SVM were higher than the other 5 models (P<0.05).

Conclusions: Our findings indicate that SVM model based on plasma miRNAs biomarkers may serve as a novel, accurate, noninvasive method for auxiliary diagnosis of lung cancer.

Sparse support vector machines with L0 approximation for ultra-high dimensional omics data

Omics data usually have ultra-high dimension (p) and small sample size (n). Standard support vector machines (SVMs), which minimize the L₂ norm for the primal variables, only lead to sparse solutions for the dual variables. L₁ based SVMs, directly minimizing the L₁ norm, have been used for feature selection with omics data. However, most current methods directly solve the primal formulations of the problem, which are not computationally scalable. The computational complexity increases with the number of features. In addition, L₁ norm is known to be asymptotically biased and not consistent for feature selection. In this paper, we develop an efficient method for sparse support vector machines with L₀ norm approximation. The proposed method approximates the L₀ minimization through solving a series of L₂ optimization problems, which can be formulated with dual variables. It finds the optimal solution for p primal variables through estimating n dual variables, which is more efficient as long as the sample size is small. L₀ approximation leads to sparsity in both dual and primal variables, and can be used for both feature and sample selections. The proposed method identifies much less number of features and achieves similar performances in simulations. We apply the proposed method to feature selections with metagenomic sequencing and gene expression data. It can identify biologically important genes and taxa efficiently.