Identification of Benign and Malignant Lung Nodules in CT Images Based on Ensemble Learning Method.

BACKGROUND AND OBJECTIVE: Under the background of urgent need for computer-aided technology to provide physicians with objective decision support, aiming at reducing the false positive rate of nodule CT detection in pulmonary nodules detection and improving the accuracy of lung nodule recognition, this paper puts forward a method based on ensemble learning to distinguish between malignant and benign pulmonary nodules. METHODS: Firstly, trained on a public data set, a multi-layer feature fusion YOLOv3 network is used to detect lung nodules. Secondly, a CNN was trained to differentiate benign from malignant pulmonary nodules. Then, based on the idea of ensemble learning, the confidence probability of the above two models and the label of the training set are taken as data features to build a Logistic regression model. Finally, two test sets (public data set and private data set) were tested, and the confidence probability output by the two models was fused into the established logistic regression model to determine benign and malignant pulmonary nodules. RESULTS: The YOLOv3 network was trained to detect chest CT images of the test set. The number of pulmonary nodules detected in the public and private test sets was 356 and 314, respectively. The accuracy, sensitivity and specificity of the two test sets were 80.97%, 81.63%, 78.75% and 79.69%, 86.59%, 72.16%, respectively. With CNN training pulmonary nodules benign and malignant discriminant model analysis of two kinds of test set, the result of accuracy, sensitivity and specificity were 90.12%, 90.66%, 89.47% and 88.57%, 85.62%, 90.87%, respectively. Fused model based on YOLOv3 network and CNN is tested on two test sets, and the result of accuracy, sensitivity and specificity were 93.82%, 94.85%, 92.59% and 92.31%, 92.68%, 91.89%, respectively. CONCLUSION: The ensemble learning model is more effective than YOLOv3 network and CNN in removing false positives, and the accuracy of the ensemble. Learning model is higher than the other two networks in identifying pulmonary nodules.

Establishment of a Lung Cancer Discriminative Model Based on an Optimized Support Vector Machine Algorithm and Study of Key Targets of Wogonin in Lung Cancer.

An optimized support vector machine model was used to construct a lung cancer diagnosis model based on serological indicators, and a molecular regulation model of Wogonin, a component of Scutellaria baicalensis, was established. Serological indexes of patients were collected, the grid search method was used to identify the optimal penalty coefficient C and parameter g of the support vector machine model, and the benign and malignant auxiliary diagnosis model of isolated pulmonary nodules based on serological indicators was established. The regulatory network and key targets of Wogonin in lung cancer were analyzed by network pharmacology, and key targets were detected by western blot. The relationship between serological susceptibility genes and key targets of Wogonin was established, and the signaling pathway of Wogonin regulating lung cancer was constructed. After support vector machine parameter optimization (C = 90.597, g = 32), the accuracy of the model was 90.8333%, with nine false positives and two false negative cases. Ontology functional analysis of 67 common genes between Wogonin targets and lung cancer-related genes showed that the targets were associated with biological processes involved in peptidye-serine modification and regulation of protein kinase B signaling; cell components in the membrane raft and chromosomal region; and molecular function in protein serine/threonine kinase activity and heme binding. Kyoto Encyclopedia of Genes and Genomes analysis showed that the regulation pathways involved the PI3K-Akt signaling pathway, ERBB signaling pathway, and EGFR tyrosine kinase inhibitor resistance. In vitro analyses using lung cancer cells showed that Wogonin led to significantly increased levels of cleaved caspase-3 and Bad and significantly decreased Bcl-2 expression in a concentration-dependent manner. ErbB4 expression also significantly decreased in lung cancer cells after treatment with Wogonin. A regulatory network of Wogonin regulating lung cancer cell apoptosis was constructed, including the participation of serological susceptibility genes. There is a certain regulatory effect between the serological indexes that can be used in the diagnosis of lung cancer and the key targets of Chinese herbal medicine treatment of lung cancer, which provides a new idea for the diagnosis, treatment and prognosis of clinical lung cancer.

The Key Target and Molecular Mechanism of the Volatile Component of Scutellaria baicalensis Georgi in Acute Lung Injury Based on Network Pharmacology.

Ethnopharmacological relevance: Scutellaria baicalensis georgi is one of the most widely studied TCMs; its effects in ALI have been studied in a large number of experiments, and the efficacy of volatile oil from TCM remains to be studied. Aim: The volatile component of Scutellaria baicalensis georgi was selected to act on the key target of acute lung injury and was preliminarily studied for its specific molecular mechanism. Methods: The volatile active substances of Scutellaria baicalensis georgi were extracted by GC-MS, and the active ingredients related with the occurrence and development of acute lung injury were searched and matched by the TCMSP database. The pharmacologic data and analysis platform of TCM were used to retrieve and screen for the volatile active components and the possible therapeutic targets of Scutellaria baicalensis georgi. In addition, acute lung injury was searched in the disease target database to identify the corresponding disease target proteins, thereby establishing a protein-protein interaction network. Finally, the effects of wogonin on the apoptotic and inflammatory factors in the acute lung injury cell model were analyzed experimentally. Results: We identified 100 candidate targets and successfully constructed a complex target network. The targets identified by the above gene enrichment analysis played important roles in the autoimmune disease cell cycle apoptosis and related signaling pathways. The KEGG pathway analysis showed that most of the target genes were involved in the inflammatory response regulation of the TRP, PI3K-Akt, and IL-17 signaling pathways. The participation of wogonin in the specific regulatory pathways of PI3K-Akt signaling and IL-17 signaling was verified through experiments. In the lung-injured cell model, the results showed that wogonin inhibited the apoptosis of injured lung cells by inhibiting the expression of BAD gene and the activation of cleaved caspase-3 gene while increasing Bcl-2 expression. In addition, wogonin inhibited the expression of the abovementioned inflammatory factors and further inhibited the inflammatory response in the lung injury cells. Conclusion: The results of pharmacological network analysis can predict and explain the regulation mechanism of multi-target and multi-pathway of TCM components. This study identified the potential target and important pathway of wogonin in regulating acute lung injury. At the same time, the accuracy of network pharmacological prediction is also preliminarily verified by molecular biology experiment.

Network Pharmacology for Analyzing the Key Targets and Potential Mechanism of Wogonin in Gliomas.

Objective: To analyze the key targets and potential mechanisms underlying the volatile components of Scutellaria baicalensis Georgi acting on gliomas through network pharmacology combined with biological experiments. Methods: We have extracted the volatile components of Scutellaria baicalensis by gas chromatography-mass spectrometry (GC-MS) and determined the active components related to the onset and development of gliomas by combining the results with the data from the Traditional Chinese Medicine Systems Pharmacology database. We screened the same targets for the extracted active components and gliomas through network pharmacology and then constructed a protein-protein interaction network. Using a Gene Ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis, we analyzed the protein effects and regulatory pathways of the common targets. Lastly, we employed ELISA and Western blot in verifying the key targets in the regulatory pathway. Results: We ultimately determined that the active component in S. baicalensis Georgi related to the onset and development of gliomas was Wogonin. The results of the network pharmacology revealed 85 targets for glioma and Wogonin. We used gene ontology to analyze these target genes and found that they involved 30 functions, such as phosphatidylinositol phosphokinase activation, while the KEGG analysis showed that there were 10 regulatory pathways involved. Through the following analysis, we found that most of the key target genes are distributed in the PI3K-Akt and interleukin 17 signaling pathways. We then cultured U251 glioma cells for the experiments. Compared with the control group, no significant change was noted in the caspase-3 expression; however, cleaved caspase-3 expression increased significantly and was dose-dependent on Wogonin. The expression of Bad and Bcl-2 with 25 µM of Wogonin has remained unchanged, but when the Wogonin dose was increased to 100 µM, the expression of Bad and Bcl-2 was noted to change significantly (Bad was significantly upregulated, while Bcl-2 was significantly downregulated) and was dose-dependent on Wogonin. The ELISA results showed that, compared with the control group, the secretion of tumor necrosis factor alpha, IL-1ß, and IL-6 decreased as the Wogonin concentration increased. Tumor necrosis factor alpha downregulation had no significant dose-dependent effect on Wogonin, the inhibitory effect of 25 µM of Wogonin on IL-6 was not significant, and IL-1ß downregulation had a significant dose-dependent effect on Wogonin. Conclusion: Wogonin might promote the apoptosis of glioma cells by upregulating proapoptotic factors, downregulating antiapoptotic factors, and inhibiting the inflammatory response, thereby inhibiting glioma progression.

Network-based method for mining novel HPV infection related genes using random walk with restart algorithm.

The human papillomavirus (HPV), a common virus that infects the reproductive tract, may lead to malignant changes within the infection area in certain cases and is directly associated with such cancers as cervical cancer, anal cancer, and vaginal cancer. Identification of novel HPV infection related genes can lead to a better understanding of the specific signal pathways and cellular processes related to HPV infection, providing information for the development of more efficient therapies. In this study, several novel HPV infection related genes were predicted by a computation method based on the known genes involved in HPV infection from HPVbase. This method applied the algorithm of random walk with restart (RWR) to a protein-protein interaction (PPI) network. The candidate genes were further filtered by the permutation and association tests. These steps eliminated genes occupying special positions in the PPI network and selected key genes with strong associations to known HPV infection related genes based on the interaction confidence and functional similarity obtained from published databases, such as STRING, gene ontology (GO) terms and KEGG pathways. Our study identified 104 novel HPV infection related genes, a number of which were confirmed to relate to the infection processes and complications of HPV infection, as reported in the literature. These results demonstrate the reliability of our method in identifying HPV infection related genes. This article is part of a Special Issue entitled: Accelerating Precision Medicine through Genetic and Genomic Big Data Analysis edited by Yudong Cai & Tao Huang.

Melanoma long non-coding RNA signature predicts prognostic survival and directs clinical risk-specific treatments.

BACKGROUND: Various studies have demonstrated that the Breslow thickness, tumor ulceration and mitotic index could serve as prognostic markers in patients with cutaneous melanoma. Recently, however, as these clinicopathological biomarkers lack efficient interpretation of endogenous mechanism of melanoma, the emphasis on the prognosis of melanoma has transformed to molecular tumor markers. OBJECTIVE: This study was designed to identify survival-related long non-coding RNAs (lncRNAs), and based on the different expressions of these lncRNAs, clinical risk-specific diagnosis and adjuvant therapy could be employed on melanoma patients, especially patients in the early course of disease or patients with a Breslow thickness no more than 2mm. METHODS: The clinical information and corresponding RNA expression data were obtained from The Cancer Genome Atlas dataset and Gene Expression Omnibus dataset (GSE65904). All samples were categorized into one training dataset and two validation datasets. Cox proportional hazard regression analysis was then used to identify survival-related lncRNAs and risk assessment signature was constructed in training dataset. Kaplan-Meier method was used to estimate the utility of this signature in predicting the duration of survival of patients both in the training dataset and two validation datasets. Meanwhile receiver operating characteristic analyses were used to evaluate the predictive effectiveness of this signature in two validation datasets. RESULTS: It was found that the signature was effective while used for risk stratification, and Kaplan-Meier analyses indicated that the duration of survival of patients in high-risk groups were significantly shorter than that of low-risk groups. Moreover, areas under the receiver operating characteristic curve were 0.711 (95% confidence interval: 0.618-0.804) and 0.698 (95% confidence interval: 0.614-0.782) when this signature was used to predict the patients' duration of survival in two validation datasets respectively, indicating the superior specificity and sensitivity of this signature. CONCLUSION: We identified a four-lncRNA prognostic signature with the ability of risk stratification for melanoma patients. Risk score acquired from this signature, combining with differential diagnosis and differential adjuvant therapy, could potentially improve the prognosis quality of life for patients, especially patients in the early course of disease or patients with a Breslow thickness no more than 2mm.

A Shortest-Path-Based Method for the Analysis and Prediction of Fruit-Related Genes in Arabidopsis thaliana.

Biologically, fruits are defined as seed-bearing reproductive structures in angiosperms that develop from the ovary. The fertilization, development and maturation of fruits are crucial for plant reproduction and are precisely regulated by intrinsic genetic regulatory factors. In this study, we used Arabidopsis thaliana as a model organism and attempted to identify novel genes related to fruit-associated biological processes. Specifically, using validated genes, we applied a shortest-path-based method to identify several novel genes in a large network constructed using the protein-protein interactions observed in Arabidopsis thaliana. The described analyses indicate that several of the discovered genes are associated with fruit fertilization, development and maturation in Arabidopsis thaliana.