Fang Liang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Fang Liang has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Buyao Zhang
Department of Critical Care Medicine, Xiangya Hospital of Central South University, Changsha, China
2 shared publications - Chenghuan Hu
Department of Critical Care Medicine, Xiangya Hospital of Central South University, Changsha, China
2 shared publications Kai Zhao
Department of Hematology and Critical Care Medicine, The 3rd Xiangya Hospital, Central South University, Changsha, China
2 shared publicationsMingwu Duan
Department of Hematology and Critical Care Medicine, The 3rd Xiangya Hospital, Central South University, Changsha, China
2 shared publications- Mingxia Li
Department of Critical Care Medicine, Zhujiang Hospital, Southern Medical University, Guangzhou, Guangdong, People’s Republic of China
2 shared publications - Qinlan Hou
Department of Critical Care Medicine, Xiangya Hospital of Central South University, Changsha, China
2 shared publications
Latest Publications
- Top Cited PapersOpen Access
- 1 May 2024
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 26 (1) , e51354
Machine Learning for Predicting Risk and Prognosis of Acute Kidney Disease in Critically Ill Elderly Patients During Hospitalization: Internet-Based and Interpretable Model Study
Background: Acute kidney disease (AKD) affects more than half of critically ill elderly patients with acute kidney injury (AKI), which leads to worse short-term outcomes. Objective: We aimed to establish 2 machine learning models to predict the risk and prognosis of AKD in the elderly and to deploy the models as online apps. Methods: Data on elderly patients with AKI (n=3542) and AKD (n=2661) from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database were used to develop 2 models for predicting the AKD risk and in-hospital mortality, respectively. Data collected from Xiangya Hospital of Central South University were for external validation. A bootstrap method was used for internal validation to obtain relatively stable results. We extracted the indicators within 24 hours of the first diagnosis of AKI and the fluctuation range of some indicators, namely delta (day 3 after AKI minus day 1), as features. Six machine learning algorithms were used for modeling; the area under the receiver operating characteristic curve (AUROC), decision curve analysis, and calibration curve for evaluating; Shapley additive explanation (SHAP) analysis for visually interpreting; and the Heroku platform for deploying the best-performing models as web-based apps. Results: For the model of predicting the risk of AKD in elderly patients with AKI during hospitalization, the Light Gradient Boosting Machine (LightGBM) showed the best overall performance in the training (AUROC=0.844, 95% CI 0.831-0.857), internal validation (AUROC=0.853, 95% CI 0.841-0.865), and external (AUROC=0.755, 95% CI 0.699–0.811) cohorts. In addition, LightGBM performed well for the AKD prognostic prediction in the training (AUROC=0.861, 95% CI 0.843-0.878), internal validation (AUROC=0.868, 95% CI 0.851-0.885), and external (AUROC=0.746, 95% CI 0.673-0.820) cohorts. The models deployed as online prediction apps allowed users to predict and provide feedback to submit new data for model iteration. In the importance ranking and correlation visualization of the model’s top 10 influencing factors conducted based on the SHAP value, partial dependence plots revealed the optimal cutoff of some interventionable indicators. The top 5 factors predicting the risk of AKD were creatinine on day 3, sepsis, delta blood urea nitrogen (BUN), diastolic blood pressure (DBP), and heart rate, while the top 5 factors determining in-hospital mortality were age, BUN on day 1, vasopressor use, BUN on day 3, and partial pressure of carbon dioxide (PaCO2). Conclusions: We developed and validated 2 online apps for predicting the risk of AKD and its prognostic mortality in elderly patients, respectively. The top 10 factors that influenced the AKD risk and mortality during hospitalization were identified and explained visually, which might provide useful applications for intelligent management and suggestions for future prospective research.
- 1 May 2024
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 26 (1) , e51354
Top Cited PapersOpen AccessMachine Learning for Predicting Risk and Prognosis of Acute Kidney Disease in Critically Ill Elderly Patients During Hospitalization: Internet-Based and Interpretable Model Study
Background: Acute kidney disease (AKD) affects more than half of critically ill elderly patients with acute kidney injury (AKI), which leads to worse short-term outcomes. Objective: We aimed to establish 2 machine learning models to predict the risk and prognosis of AKD in the elderly and to deploy the models as online apps. Methods: Data on elderly patients with AKI (n=3542) and AKD (n=2661) from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database were used to develop 2 models for predicting the AKD risk and in-hospital mortality, respectively. Data collected from Xiangya Hospital of Central South University were for external validation. A bootstrap method was used for internal validation to obtain relatively stable results. We extracted the indicators within 24 hours of the first diagnosis of AKI and the fluctuation range of some indicators, namely delta (day 3 after AKI minus day 1), as features. Six machine learning algorithms were used for modeling; the area under the receiver operating characteristic curve (AUROC), decision curve analysis, and calibration curve for evaluating; Shapley additive explanation (SHAP) analysis for visually interpreting; and the Heroku platform for deploying the best-performing models as web-based apps. Results: For the model of predicting the risk of AKD in elderly patients with AKI during hospitalization, the Light Gradient Boosting Machine (LightGBM) showed the best overall performance in the training (AUROC=0.844, 95% CI 0.831-0.857), internal validation (AUROC=0.853, 95% CI 0.841-0.865), and external (AUROC=0.755, 95% CI 0.699–0.811) cohorts. In addition, LightGBM performed well for the AKD prognostic prediction in the training (AUROC=0.861, 95% CI 0.843-0.878), internal validation (AUROC=0.868, 95% CI 0.851-0.885), and external (AUROC=0.746, 95% CI 0.673-0.820) cohorts. The models deployed as online prediction apps allowed users to predict and provide feedback to submit new data for model iteration. In the importance ranking and correlation visualization of the model’s top 10 influencing factors conducted based on the SHAP value, partial dependence plots revealed the optimal cutoff of some interventionable indicators. The top 5 factors predicting the risk of AKD were creatinine on day 3, sepsis, delta blood urea nitrogen (BUN), diastolic blood pressure (DBP), and heart rate, while the top 5 factors determining in-hospital mortality were age, BUN on day 1, vasopressor use, BUN on day 3, and partial pressure of carbon dioxide (PaCO2). Conclusions: We developed and validated 2 online apps for predicting the risk of AKD and its prognostic mortality in elderly patients, respectively. The top 10 factors that influenced the AKD risk and mortality during hospitalization were identified and explained visually, which might provide useful applications for intelligent management and suggestions for future prospective research.
- Open Access
- 28 July 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Machine Learning for Predicting Risk and Prognosis of Acute Kidney Disease in Critically Ill Elderly Patients During Hospitalization: Internet-Based and Interpretable Model Study (Preprint)
BACKGROUND Acute kidney disease (AKD) affects more than half of critically ill elderly patients with acute kidney injury (AKI), which leads to worse short-term outcomes. OBJECTIVE We aimed to establish 2 machine learning models to predict the risk and prognosis of AKD in the elderly and to deploy the models as online apps. METHODS Data on elderly patients with AKI (n=3542) and AKD (n=2661) from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database were used to develop 2 models for predicting the AKD risk and in-hospital mortality, respectively. Data collected from Xiangya Hospital of Central South University were for external validation. A bootstrap method was used for internal validation to obtain relatively stable results. We extracted the indicators within 24 hours of the first diagnosis of AKI and the fluctuation range of some indicators, namely delta (day 3 after AKI minus day 1), as features. Six machine learning algorithms were used for modeling; the area under the receiver operating characteristic curve (AUROC), decision curve analysis, and calibration curve for evaluating; Shapley additive explanation (SHAP) analysis for visually interpreting; and the Heroku platform for deploying the best-performing models as web-based apps. RESULTS For the model of predicting the risk of AKD in elderly patients with AKI during hospitalization, the Light Gradient Boosting Machine (LightGBM) showed the best overall performance in the training (AUROC=0.844, 95% CI 0.831-0.857), internal validation (AUROC=0.853, 95% CI 0.841-0.865), and external (AUROC=0.755, 95% CI 0.699–0.811) cohorts. In addition, LightGBM performed well for the AKD prognostic prediction in the training (AUROC=0.861, 95% CI 0.843-0.878), internal validation (AUROC=0.868, 95% CI 0.851-0.885), and external (AUROC=0.746, 95% CI 0.673-0.820) cohorts. The models deployed as online prediction apps allowed users to predict and provide feedback to submit new data for model iteration. In the importance ranking and correlation visualization of the model’s top 10 influencing factors conducted based on the SHAP value, partial dependence plots revealed the optimal cutoff of some interventionable indicators. The top 5 factors predicting the risk of AKD were creatinine on day 3, sepsis, delta blood urea nitrogen (BUN), diastolic blood pressure (DBP), and heart rate, while the top 5 factors determining in-hospital mortality were age, BUN on day 1, vasopressor use, BUN on day 3, and partial pressure of carbon dioxide (PaCO2). CONCLUSIONS We developed and validated 2 online apps for predicting the risk of AKD and its prognostic mortality in elderly patients, respectively. The top 10 factors that influenced the AKD risk and mortality during hospitalization were identified and explained visually, which might provide useful applications for intelligent management and suggestions for future prospective research.
- 28 July 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessMachine Learning for Predicting Risk and Prognosis of Acute Kidney Disease in Critically Ill Elderly Patients During Hospitalization: Internet-Based and Interpretable Model Study (Preprint)
BACKGROUND Acute kidney disease (AKD) affects more than half of critically ill elderly patients with acute kidney injury (AKI), which leads to worse short-term outcomes. OBJECTIVE We aimed to establish 2 machine learning models to predict the risk and prognosis of AKD in the elderly and to deploy the models as online apps. METHODS Data on elderly patients with AKI (n=3542) and AKD (n=2661) from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database were used to develop 2 models for predicting the AKD risk and in-hospital mortality, respectively. Data collected from Xiangya Hospital of Central South University were for external validation. A bootstrap method was used for internal validation to obtain relatively stable results. We extracted the indicators within 24 hours of the first diagnosis of AKI and the fluctuation range of some indicators, namely delta (day 3 after AKI minus day 1), as features. Six machine learning algorithms were used for modeling; the area under the receiver operating characteristic curve (AUROC), decision curve analysis, and calibration curve for evaluating; Shapley additive explanation (SHAP) analysis for visually interpreting; and the Heroku platform for deploying the best-performing models as web-based apps. RESULTS For the model of predicting the risk of AKD in elderly patients with AKI during hospitalization, the Light Gradient Boosting Machine (LightGBM) showed the best overall performance in the training (AUROC=0.844, 95% CI 0.831-0.857), internal validation (AUROC=0.853, 95% CI 0.841-0.865), and external (AUROC=0.755, 95% CI 0.699–0.811) cohorts. In addition, LightGBM performed well for the AKD prognostic prediction in the training (AUROC=0.861, 95% CI 0.843-0.878), internal validation (AUROC=0.868, 95% CI 0.851-0.885), and external (AUROC=0.746, 95% CI 0.673-0.820) cohorts. The models deployed as online prediction apps allowed users to predict and provide feedback to submit new data for model iteration. In the importance ranking and correlation visualization of the model’s top 10 influencing factors conducted based on the SHAP value, partial dependence plots revealed the optimal cutoff of some interventionable indicators. The top 5 factors predicting the risk of AKD were creatinine on day 3, sepsis, delta blood urea nitrogen (BUN), diastolic blood pressure (DBP), and heart rate, while the top 5 factors determining in-hospital mortality were age, BUN on day 1, vasopressor use, BUN on day 3, and partial pressure of carbon dioxide (PaCO2). CONCLUSIONS We developed and validated 2 online apps for predicting the risk of AKD and its prognostic mortality in elderly patients, respectively. The top 10 factors that influenced the AKD risk and mortality during hospitalization were identified and explained visually, which might provide useful applications for intelligent management and suggestions for future prospective research.
- Open Access
- 8 June 2023
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 14, 1188482
The landscape of mitophagy in sepsis reveals PHB1 as an NLRP3 inflammasome inhibitor
Mitophagy is a selective autophagy targeting damaged and potential cytotoxic mitochondria, which can effectively prevent excessive cytotoxic production from damaged mitochondria and alleviate the inflammatory response. However, the potential role of mitophagy in sepsis remains poorly explored. Here, we studied the role of mitophagy in sepsis and its immune heterogeneity. By performing mitophagy-related typing on 348 sepsis samples, three clusters (A, B, and C) were obtained. Cluster A had the highest degree of mitophagy accompanied by lowest disease severity, while cluster C had the lowest degree of mitophagy with the highest disease severity. The three clusters had unique immune characteristics. We further revealed that the expression of PHB1 in these three clusters was significantly different and negatively correlated with the severity of sepsis, suggesting that PHB1 was involved in the development of sepsis. It has been reported that impaired mitophagy leads to the over-activation of inflammasomes, which promotes sepsis development. Further analysis showed that the expressions of NLRP3 inflammasomes core genes in cluster C were significantly up-regulated and negatively correlated with PHB1. Next, we verified whether PHB1 downregulation caused the activation of inflammasomes and found that the PHB1 knockdown increased the levels of mtDNA in the cytoplasm and enhanced the activation of NLRP3 inflammasomes. In addition, mitophagy inhibitor treatment abolished PHB1 knockdown-mediated activation of NLRP3 inflammasomes, suggesting that PHB1 inhibited the activation of inflammasomes through mitophagy. In conclusion, this study reveals that a high degree of mitophagy may predict a good outcome of sepsis, and PHB1 is a key NLRP3 inflammasome regulator via mitophagy in inflammatory diseases such as sepsis.
- 8 June 2023
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 14, 1188482
Open AccessThe landscape of mitophagy in sepsis reveals PHB1 as an NLRP3 inflammasome inhibitor
Mitophagy is a selective autophagy targeting damaged and potential cytotoxic mitochondria, which can effectively prevent excessive cytotoxic production from damaged mitochondria and alleviate the inflammatory response. However, the potential role of mitophagy in sepsis remains poorly explored. Here, we studied the role of mitophagy in sepsis and its immune heterogeneity. By performing mitophagy-related typing on 348 sepsis samples, three clusters (A, B, and C) were obtained. Cluster A had the highest degree of mitophagy accompanied by lowest disease severity, while cluster C had the lowest degree of mitophagy with the highest disease severity. The three clusters had unique immune characteristics. We further revealed that the expression of PHB1 in these three clusters was significantly different and negatively correlated with the severity of sepsis, suggesting that PHB1 was involved in the development of sepsis. It has been reported that impaired mitophagy leads to the over-activation of inflammasomes, which promotes sepsis development. Further analysis showed that the expressions of NLRP3 inflammasomes core genes in cluster C were significantly up-regulated and negatively correlated with PHB1. Next, we verified whether PHB1 downregulation caused the activation of inflammasomes and found that the PHB1 knockdown increased the levels of mtDNA in the cytoplasm and enhanced the activation of NLRP3 inflammasomes. In addition, mitophagy inhibitor treatment abolished PHB1 knockdown-mediated activation of NLRP3 inflammasomes, suggesting that PHB1 inhibited the activation of inflammasomes through mitophagy. In conclusion, this study reveals that a high degree of mitophagy may predict a good outcome of sepsis, and PHB1 is a key NLRP3 inflammasome regulator via mitophagy in inflammatory diseases such as sepsis.
- Open Access
- 21 December 2022
- Published by Frontiers Media SA in Frontiers in Cellular and Infection Microbiology
- Vol. 12, 1013526
Pulmonary fibrosis in patients with COVID-19: A retrospective study
Background: Coronavirus disease 2019 (COVID-19) pandemic has caused substantial threats to people’s physical health and lives, claiming the lives of over 6 million people worldwide. Although the mortality rate of COVID-19 is very low, many survivors may have different degrees and various sequelae. Previous studies have shown that pulmonary fibrosis (PF) were common on discharged COVID-19 patients, and PF itself is a poor prognostic factor.Methods: 227 COVID-19 hospitalized patients’ clinical and laboratory data from the first 15 days following admission were collected in this retrospective study. Groups were based on with or without PF of COVID-19. Categorical variables were compared with the chi-square test or Fisher’s exact test. Continuous variables were tested by Wilcoxon rank-sum test for the non-normal distribution. Spearman correlations were used to assess the correlations between PF with clinic parameters of multiple time points. Univariate and multivariate logistic regression were used to analyze for risk factors of COVID-19 patients with pulmonary fibrosis.Results: Sixty cases of COVID-19 patients were diagnosed with PF. Compared with 167 non-PF patients, those with PF were older and had higher proportions of fever, shortness of breath, hemoptysis, abdominal pain, hypertension, cardiovascular, diabetes, high flow nasal cannula (HFNC), severe disease, and virus shedding duration. Furthermore, the correlation analysis between PF and clinic parameters showed that PF were positively related to the C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), and negatively correlated with hemoglobin (HGB) and albumin (ALB) at all time points in the first 15 days after admission. Moreover, We found that PF were significantly correlated with coagulation indexes prothrombin time (PT), activated partial thromboplastin time (APTT), fibrinogen (Fib) and fibrinolysis index D-Dimer at some phases. In addition, Univariate logistic regression analyses showed that age, fever, shortness of breath, hemoptysis, hypertension, cardiovascular, diabetes, HFNC, severe disease were the risk factors of COVID-19 patients with PF. However, multivariate logistic regression showed that age was the risk factor of COVID-19 patients with PF.Conclusion: Combining various factors, advanced age is an independent risk factor of COVID-19 patients with PF. PF was significantly related with clinic parameter of inflammation/coagulopathy/fibrinolysis.
- 21 December 2022
- Published by Frontiers Media SA in Frontiers in Cellular and Infection Microbiology
- Vol. 12, 1013526
Open AccessPulmonary fibrosis in patients with COVID-19: A retrospective study
Background: Coronavirus disease 2019 (COVID-19) pandemic has caused substantial threats to people’s physical health and lives, claiming the lives of over 6 million people worldwide. Although the mortality rate of COVID-19 is very low, many survivors may have different degrees and various sequelae. Previous studies have shown that pulmonary fibrosis (PF) were common on discharged COVID-19 patients, and PF itself is a poor prognostic factor.Methods: 227 COVID-19 hospitalized patients’ clinical and laboratory data from the first 15 days following admission were collected in this retrospective study. Groups were based on with or without PF of COVID-19. Categorical variables were compared with the chi-square test or Fisher’s exact test. Continuous variables were tested by Wilcoxon rank-sum test for the non-normal distribution. Spearman correlations were used to assess the correlations between PF with clinic parameters of multiple time points. Univariate and multivariate logistic regression were used to analyze for risk factors of COVID-19 patients with pulmonary fibrosis.Results: Sixty cases of COVID-19 patients were diagnosed with PF. Compared with 167 non-PF patients, those with PF were older and had higher proportions of fever, shortness of breath, hemoptysis, abdominal pain, hypertension, cardiovascular, diabetes, high flow nasal cannula (HFNC), severe disease, and virus shedding duration. Furthermore, the correlation analysis between PF and clinic parameters showed that PF were positively related to the C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), and negatively correlated with hemoglobin (HGB) and albumin (ALB) at all time points in the first 15 days after admission. Moreover, We found that PF were significantly correlated with coagulation indexes prothrombin time (PT), activated partial thromboplastin time (APTT), fibrinogen (Fib) and fibrinolysis index D-Dimer at some phases. In addition, Univariate logistic regression analyses showed that age, fever, shortness of breath, hemoptysis, hypertension, cardiovascular, diabetes, HFNC, severe disease were the risk factors of COVID-19 patients with PF. However, multivariate logistic regression showed that age was the risk factor of COVID-19 patients with PF.Conclusion: Combining various factors, advanced age is an independent risk factor of COVID-19 patients with PF. PF was significantly related with clinic parameter of inflammation/coagulopathy/fibrinolysis.
- Open Access
- 19 October 2022
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 13, 986214
The role of cuproptosis-related gene in the classification and prognosis of melanoma
Background: Melanoma, as one of the most aggressive and malignant cancers, ranks first in the lethality rate of skin cancers. Cuproptosis has been shown to paly a role in tumorigenesis, However, the role of cuproptosis in melanoma metastasis are not clear. Studying the correlation beteen the molecular subtypes of cuproptosis-related genes (CRGs) and metastasis of melanoma may provide some guidance for the prognosis of melanoma.Methods: We collected 1085 melanoma samples in The Cancer Genome Atlas(TCGA) and Gene Expression Omnibus(GEO) databases, constructed CRGs molecular subtypes and gene subtypes according to clinical characteristics, and investigated the role of CRGs in melanoma metastasis. We randomly divide the samples into train set and validation set according to the ratio of 1:1. A prognostic model was constructed using data from the train set and then validated on the validation set. We performed tumor microenvironment analysis and drug sensitivity analyses for high and low risk groups based on the outcome of the prognostic model risk score. Finally, we established a metastatic model of melanoma.Results: According to the expression levels of 12 cuproptosis-related genes, we obtained three subtypes of A1, B1, and C1. Among them, C1 subtype had the best survival outcome. Based on the differentially expressed genes shared by A1, B1, and C1 genotypes, we obtained the results of three gene subtypes of A2, B2, and C2. Among them, the B2 group had the best survival outcome. Then, we constructed a prognostic model consisting of 6 key variable genes, which could more accurately predict the 1-, 3-, and 5-year overall survival rates of melanoma patients. Besides, 98 drugs were screened out. Finally, we explored the role of cuproptosis-related genes in melanoma metastasis and established a metastasis model using seven key genes.Conclusions: In conclusion, CRGs play a role in the metastasis and prognosis of melanoma, and also provide new insights into the underlying pathogenesis of melanoma.
- 19 October 2022
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 13, 986214
Open AccessThe role of cuproptosis-related gene in the classification and prognosis of melanoma
Background: Melanoma, as one of the most aggressive and malignant cancers, ranks first in the lethality rate of skin cancers. Cuproptosis has been shown to paly a role in tumorigenesis, However, the role of cuproptosis in melanoma metastasis are not clear. Studying the correlation beteen the molecular subtypes of cuproptosis-related genes (CRGs) and metastasis of melanoma may provide some guidance for the prognosis of melanoma.Methods: We collected 1085 melanoma samples in The Cancer Genome Atlas(TCGA) and Gene Expression Omnibus(GEO) databases, constructed CRGs molecular subtypes and gene subtypes according to clinical characteristics, and investigated the role of CRGs in melanoma metastasis. We randomly divide the samples into train set and validation set according to the ratio of 1:1. A prognostic model was constructed using data from the train set and then validated on the validation set. We performed tumor microenvironment analysis and drug sensitivity analyses for high and low risk groups based on the outcome of the prognostic model risk score. Finally, we established a metastatic model of melanoma.Results: According to the expression levels of 12 cuproptosis-related genes, we obtained three subtypes of A1, B1, and C1. Among them, C1 subtype had the best survival outcome. Based on the differentially expressed genes shared by A1, B1, and C1 genotypes, we obtained the results of three gene subtypes of A2, B2, and C2. Among them, the B2 group had the best survival outcome. Then, we constructed a prognostic model consisting of 6 key variable genes, which could more accurately predict the 1-, 3-, and 5-year overall survival rates of melanoma patients. Besides, 98 drugs were screened out. Finally, we explored the role of cuproptosis-related genes in melanoma metastasis and established a metastasis model using seven key genes.Conclusions: In conclusion, CRGs play a role in the metastasis and prognosis of melanoma, and also provide new insights into the underlying pathogenesis of melanoma.
- Open Access
- 5 September 2022
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 13, 967299
RAB20 deficiency promotes the development of silicosis via NLRP3 inflammasome
Silicosis is a worldwide serious occupational disease that is caused by inhalation of silica crystals. However, little is known about the pathogenesis mechanism of silicosis. We performed single-cell sequencing in bronchoalveolar lavage fluid (BALF) from mine workers with silicosis and their co-workers who did not develop silicosis, and found that the RAB20 deficiency in monocytes/macrophages was strongly linked to the development of silicosis. In the silicosis murine model, RAB20 knockout markedly enhanced the silica crystal-induced pulmonary interstitial fibrosis and respiratory dysfunction. Moreover, this process is strongly accompanied by IL-1β release and NLRP3 activation. In vitro, RAB20 knockout macrophages aggravated the crystalline silica-induced IL-1β release and NLRP3 inflammasome activation partly by increased ratio of crystalline silica/phagosomal areas/volumes to induce lysosomal injury. Thus, these findings provide novel molecular insights into the intricate mechanisms underlying lysosomal protein RAB20 that are necessary for environmental irritant-mediated innate immunity, and shed light on the future development of novel therapy target for the prevention of silicosis.
- 5 September 2022
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 13, 967299
Open AccessRAB20 deficiency promotes the development of silicosis via NLRP3 inflammasome
Silicosis is a worldwide serious occupational disease that is caused by inhalation of silica crystals. However, little is known about the pathogenesis mechanism of silicosis. We performed single-cell sequencing in bronchoalveolar lavage fluid (BALF) from mine workers with silicosis and their co-workers who did not develop silicosis, and found that the RAB20 deficiency in monocytes/macrophages was strongly linked to the development of silicosis. In the silicosis murine model, RAB20 knockout markedly enhanced the silica crystal-induced pulmonary interstitial fibrosis and respiratory dysfunction. Moreover, this process is strongly accompanied by IL-1β release and NLRP3 activation. In vitro, RAB20 knockout macrophages aggravated the crystalline silica-induced IL-1β release and NLRP3 inflammasome activation partly by increased ratio of crystalline silica/phagosomal areas/volumes to induce lysosomal injury. Thus, these findings provide novel molecular insights into the intricate mechanisms underlying lysosomal protein RAB20 that are necessary for environmental irritant-mediated innate immunity, and shed light on the future development of novel therapy target for the prevention of silicosis.
- Open Access
- 1 June 2022
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 13, 858466
Identification and Validation of the Diagnostic Characteristic Genes of Ovarian Cancer by Bioinformatics and Machine Learning
Background: Ovarian cancer (OC) has a high mortality rate and poses a severe threat to women’s health. However, abnormal gene expression underlying the tumorigenesis of OC has not been fully understood. This study aims to identify diagnostic characteristic genes involved in OC by bioinformatics and machine learning.Methods: We utilized five datasets retrieved from the Gene Expression Omnibus (GEO) database, The Cancer Genome Atlas (TCGA) database, and the Genotype-Tissue Expression (GTEx) Project database. GSE12470 and GSE18520 were combined as the training set, and GSE27651 was used as the validation set A. Also, we combined the TCGA database and GTEx database as validation set B. First, in the training set, differentially expressed genes (DEGs) between OC and non-ovarian cancer tissues (nOC) were identified. Next, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), Disease Ontology (DO) enrichment analysis, and Gene Set Enrichment Analysis (GSEA) were performed for functional enrichment analysis of these DEGs. Then, two machine learning algorithms, Least Absolute Shrinkage and Selector Operation (LASSO) and Support Vector Machine-Recursive Feature Elimination (SVM-RFE), were used to get the diagnostic genes. Subsequently, the obtained diagnostic-related DEGs were validated in the validation sets. Then, we used the computational approach (CIBERSORT) to analyze the association between immune cell infiltration and DEGs. Finally, we analyzed the prognostic role of several genes on the KM-plotter website and used the human protein atlas (HPA) online database to analyze the expression of these genes at the protein level.Results: 590 DEGs were identified, including 276 upregulated and 314 downregulated DEGs.The Enrichment analysis results indicated the DEGs were mainly involved in the nuclear division, cell cycle, and IL−17 signaling pathway. Besides, DEGs were also closely related to immune cell infiltration. Finally, we found that BUB1, FOLR1, and PSAT1 have prognostic roles and the protein-level expression of these six genes SFPR1, PSAT1, PDE8B, INAVA and TMEM139 in OC tissue and nOC tissue was consistent with our analysis.Conclusions: We screened nine diagnostic characteristic genes of OC, including SFRP1, PSAT1, BUB1B, FOLR1, ABCB1, PDE8B, INAVA, BUB1, TMEM139. Combining these genes may be useful for OC diagnosis and evaluating immune cell infiltration.
- 1 June 2022
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 13, 858466
Open AccessIdentification and Validation of the Diagnostic Characteristic Genes of Ovarian Cancer by Bioinformatics and Machine Learning
Background: Ovarian cancer (OC) has a high mortality rate and poses a severe threat to women’s health. However, abnormal gene expression underlying the tumorigenesis of OC has not been fully understood. This study aims to identify diagnostic characteristic genes involved in OC by bioinformatics and machine learning.Methods: We utilized five datasets retrieved from the Gene Expression Omnibus (GEO) database, The Cancer Genome Atlas (TCGA) database, and the Genotype-Tissue Expression (GTEx) Project database. GSE12470 and GSE18520 were combined as the training set, and GSE27651 was used as the validation set A. Also, we combined the TCGA database and GTEx database as validation set B. First, in the training set, differentially expressed genes (DEGs) between OC and non-ovarian cancer tissues (nOC) were identified. Next, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), Disease Ontology (DO) enrichment analysis, and Gene Set Enrichment Analysis (GSEA) were performed for functional enrichment analysis of these DEGs. Then, two machine learning algorithms, Least Absolute Shrinkage and Selector Operation (LASSO) and Support Vector Machine-Recursive Feature Elimination (SVM-RFE), were used to get the diagnostic genes. Subsequently, the obtained diagnostic-related DEGs were validated in the validation sets. Then, we used the computational approach (CIBERSORT) to analyze the association between immune cell infiltration and DEGs. Finally, we analyzed the prognostic role of several genes on the KM-plotter website and used the human protein atlas (HPA) online database to analyze the expression of these genes at the protein level.Results: 590 DEGs were identified, including 276 upregulated and 314 downregulated DEGs.The Enrichment analysis results indicated the DEGs were mainly involved in the nuclear division, cell cycle, and IL−17 signaling pathway. Besides, DEGs were also closely related to immune cell infiltration. Finally, we found that BUB1, FOLR1, and PSAT1 have prognostic roles and the protein-level expression of these six genes SFPR1, PSAT1, PDE8B, INAVA and TMEM139 in OC tissue and nOC tissue was consistent with our analysis.Conclusions: We screened nine diagnostic characteristic genes of OC, including SFRP1, PSAT1, BUB1B, FOLR1, ABCB1, PDE8B, INAVA, BUB1, TMEM139. Combining these genes may be useful for OC diagnosis and evaluating immune cell infiltration.