Fig. 3 | Nature Communications

Fig. 3

From: A semi-supervised approach for predicting cell-type specific functional consequences of non-coding variation using MPRAs

Fig. 3

Organismal-level prediction of regulatory variants. Each figure presents the AUPR (area under the precision recal curve), AUROC (area under the receiver operating characteristics curve), and COR (Pearson correlation between predicted and true labels) for different organism-level prediction methods. X axis presents AUPR; Y axis presents AUROC; the bubble size presents COR. Fine mapping eQTLs: Uniformly processed fine-mapping eQTLs from 11 studies (31,118 positive variants, 36,540 control variants); Experimentally validated SNPs: manually curated experimentally validated regulatory SNPs (76 positive variants, 156 control variants); Refined causal SNPs: refined causal SNPs in the non-coding regions from different resources including HGMD, ClinVar, OregAnno, and variants from fine-mapping candidate causal SNPs for 39 immune and non-immune diseases (5247 positive variants, 20,988 control variants); Allelic imbalanced SNPs: allelic imbalanced SNPs in chromatin accessibility using a large number of DNase-seq assays (8592 positive variants, 9678 control variants)

Back to article page