Fig 1: The performance of the random forest model in the local cohort. (A, B) The qPCR experiments indicated that RNMT and RBM24 were both down-regulated in the knee cartilage tissue extracted from OA patients. (C, D) The ROC analysis (C) and the confusion matrix (D) of the random forest model in the local cohort.
Fig 2: Feature selection via machine learning algorithms. (A) 15 genes were determined by Lasso regression. (B) The parameters of the variables in the Lasso regression model. (C) 37 genes were identified by SVM-RFE algorithm. (D) Boruta algorithm helped to select 37 genes. (E) RNMT and RBM24 were con-determined by the PPI network analysis and machine learning algorithms.
Fig 3: RNMT and RBM24 were associated with genesis of OA. (A–C) The diagnosis value of RNMT and RBM24 in the training cohort (A), the GSE117999 cohort (B), and the local cohort (C). (D) The mean decrease accuracy (up) and the mean decrease Gini (bottom) of RNMT and RBM24 in the random forest model. (E, F) The qPCR experiments (E) and Western Blotting (F) displayed that RNMT and RBM24 were both down-regulated in the CHON-001 cells treated with 10 ng/mL IL-1ß.
Fig 4: The functionally-associated genes and GO enrichment. (A, B) The Top 20 genes associated with RBM24 (A) and RNMT (B). GO functional annotation of the associated genes of RBM24 (C) and RNMT (D). Abbreviation: GO: gene ontology.
Supplier Page from Thermo Fisher Scientific for RNMT Antibody