Fig. 1: Selecting lncRNAs to predict HRD and PARP inhibitor sensitivity. | Communications Biology

Fig. 1: Selecting lncRNAs to predict HRD and PARP inhibitor sensitivity.

From: Long non-coding RNAs as a biomarker for homologous recombination deficiency and parp inhibitor sensitivity in high-grade serous ovarian cancers

Fig. 1

a Flowchart of the lncRNA selection process based on their predictive value for HRD-score, PARPi7-score, eCARD-score, CCNE1 amplification status or BRCA1/2 mutation status using multiple feature extraction methods. b Venn diagram showing overlap of different feature extraction methods. c Heatmap of pearson correlation matrix of different scores against the expression of 29 lncRNAs. d Left: Prediction of HRD-scores using different algorithms with 10-time cross validation on the TCGA ovarian cancer dataset. Right: Correlation of predicted HRD-score and measured HRD-score on the test dataset using the random forest algorithm. e Left: Prediction of PARPi7-scores using different algorithms with 10-time cross validation on the TCGA ovarian cancer dataset. Right: Correlation of predicted-PARPi7 and measured PARPi7 on the test dataset using the random forest algorithm. f Heat map of the average expression of 29 lncRNA in different cancer types of the pan-cancer dataset of the TCGA data.

Back to article page