Machine learning based identification potential feature genes for prediction of drug efficacy in nonalcoholic steatohepatitis animal model.

BACKGROUND: Nonalcoholic Steatohepatitis (NASH) results from complex liver conditions involving metabolic, inflammatory, and fibrogenic processes. Despite its burden, there has been a lack of any approved food-and-drug administration therapy up till now. PURPOSE: Utilizing machine learning (ML) algorithms, the study aims to identify reliable potential genes to accurately predict the treatment response in the NASH animal model using biochemical and molecular markers retrieved using bioinformatics techniques. METHODS: The NASH-induced rat models were administered various microbiome-targeted therapies and herbal drugs for 12 weeks, these drugs resulted in reducing hepatic lipid accumulation, liver inflammation, and histopathological changes. The ML model was trained and tested based on the Histopathological NASH score (HPS); while (0-4) HPS considered Improved NASH and (5-8) considered non-improved, confirmed through rats' liver histopathological examination, incorporates 34 features comprising 20 molecular markers (mRNAs-microRNAs-Long non-coding-RNAs) and 14 biochemical markers that are highly enriched in NASH pathogenesis. Six different ML models were used in the proposed model for the prediction of NASH improvement, with Gradient Boosting demonstrating the highest accuracy of 98% in predicting NASH drug response. FINDINGS: Following a gradual reduction in features, the outcomes demonstrated superior performance when employing the Random Forest classifier, yielding an accuracy of 98.4%. The principal selected molecular features included YAP1, LATS1, NF2, SRD5A3-AS1, FOXA2, TEAD2, miR-650, MMP14, ITGB1, and miR-6881-5P, while the biochemical markers comprised triglycerides (TG), ALT, ALP, total bilirubin (T. Bilirubin), alpha-fetoprotein (AFP), and low-density lipoprotein cholesterol (LDL-C). CONCLUSION: This study introduced an ML model incorporating 16 noninvasive features, including molecular and biochemical signatures, which achieved high performance and accuracy in detecting NASH improvement. This model could potentially be used as diagnostic tools and to identify target therapies.

Machine-Learning-Based Identification of Key Feature RNA-Signature Linked to Diagnosis of Hepatocellular Carcinoma.

Background: Hepatocellular carcinoma (HCC) is the third prime cause of malignancy-related mortality worldwide. Early and accurate identification of HCC is crucial for good prognosis, efficacy of therapy, and survival rates of the patients. We aimed to develop a machine-learning model incorporating differentially expressed RNA signatures with laboratory parameters to construct an RNA signature-based diagnostic model for HCC. Methods: We have used five classifiers (KNN, RF, SVM, LGBM, and DNNs) to predict the liver disease (HCC). The classifiers were trained on 187 samples and then tested on 80 samples. The model included 22 features (age, sex, smoking, cirrhosis, non-cirrhosis, albumin, ALT, AST bilirubin (total and direct), INR, AFP, HBV Ag, HCV Abs, RQmiR-1298, RQmiR-1262, RQmiR-106b-3p, RQmRNARAB11A, and RQSTAT1, RQmRNAATG12, RQLnc-WRAP53, RQLncRNA- RP11-513I15.6). Results: LGBM achieved the highest accuracy of 98.75% in predicting HCC among all models surpassing Random Forest (96.25%), DNN (91.25%), SVC (88.75%), and KNN (87.50%). Conclusion: Our machine-learning model incorporating the expression data of RAB11A/STAT1/ATG12/miR-1262/miR-1298/miR-106b-3p/lncRNA-RP11-513I15.6/lncRNA-WRAP53 signature and clinical data represents a potential novel diagnostic model for HCC.

Comprehensive machine learning models for predicting therapeutic targets in type 2 diabetes utilizing molecular and biochemical features in rats.

Introduction: With the increasing prevalence of type 2 diabetes mellitus (T2DM), there is an urgent need to discover effective therapeutic targets for this complex condition. Coding and non-coding RNAs, with traditional biochemical parameters, have shown promise as viable targets for therapy. Machine learning (ML) techniques have emerged as powerful tools for predicting drug responses. Method: In this study, we developed an ML-based model to identify the most influential features for drug response in the treatment of type 2 diabetes using three medicinal plant-based drugs (Rosavin, Caffeic acid, and Isorhamnetin), and a probiotics drug (Z-biotic), at different doses. A hundred rats were randomly assigned to ten groups, including a normal group, a streptozotocin-induced diabetic group, and eight treated groups. Serum samples were collected for biochemical analysis, while liver tissues (L) and adipose tissues (A) underwent histopathological examination and molecular biomarker extraction using quantitative PCR. Utilizing five machine learning algorithms, we integrated 32 molecular features and 12 biochemical features to select the most predictive targets for each model and the combined model. Results and discussion: Our results indicated that high doses of the selected drugs effectively mitigated liver inflammation, reduced insulin resistance, and improved lipid profiles and renal function biomarkers. The machine learning model identified 13 molecular features, 10 biochemical features, and 20 combined features with an accuracy of 80% and AUC (0.894, 0.93, and 0.896), respectively. This study presents an ML model that accurately identifies effective therapeutic targets implicated in the molecular pathways associated with T2DM pathogenesis.