Fig 1.
Basic characteristics of GenoSkyline-Plus annotation.
(A) Odds ratio of predicting functionality. Each box represents the odds ratio for the same data type across 127 GenoSkyline-Plus tracks. (B) Histogram of predicted functional proportion across 127 annotation tracks. Dashed line marks the mean functional proportion. (C) Distribution of tracks with predicted functionality. For example, 26% of exon regions are predicted to be functional in more than 10 GenoSkyline-Plus tracks.
Fig 2.
Identify tissue and cell type-specific functionality.
Predicted functional proportions of different classes of previously identified tissue-specific non-coding elements. Sample sizes count the number of non-coding elements with specificity for the titled tissue or cell type (see Methods). Darker bars represent annotation tracks that physiologically match the tissue to which the corresponding set of non-coding elements are specific. (A) miRNAs with TSI > 0.75 identified in Ludwig et al. (B) lncRNAs with TSI > 0.75 identified in Derrien et al. (C) Enhancers with differential expression within a cell type facet identified by Andersson et al. (D) Predicted functional elements based on GenoSkyline-Plus annotations in the IL-17A LCR. Orange boxes mark identified CNS sites. (E) Predicted functional proportion in CNS sites and their 200-bp flanking regions across different T-cell subsets.
Fig 3.
Enrichment analysis for 45 human complex traits.
(A) Relationship between GWAS sample size, total count of significant associations, and signal enrichment in the functional genome. Traits significantly enriched in at least one annotation are highlighted in red. (B) Enrichment in the general functional genome predicted by GenoCanyon annotation. The dashed line marks the Bonferroni-corrected significance cutoff. (C) Enrichment across 7 broadly defined tissue tacks. Asterisks highlight significance after correcting for 45 traits and 7 tissues. (D) Enrichment in 66 tissue and cell tracks. Asterisks highlight significant enrichment after correcting for 45 traits and 66 annotations. Details for annotation tracks and different traits are summarized in S2 and S3 Tables.
Fig 4.
Tissue and cell type-specific enrichment for AD and PD.
(A) Enrichment in 7 broadly defined tissue tracks. (B) Enrichment analysis using 66 GenoSkyline-Plus tissue and cell tracks. Dashed lines indicate Bonferroni-corrected significance cutoff. (C) Percentage of variants covered by each annotated category and percentage of heritability explained by variants in that category.
Fig 5.
Identify genetic correlation between LOAD and PD.
(A) estimated chromosome-by-chromosome heritability percentage for LOAD and PD. (B) chromosome-by-chromosome heritability in the monocyte functional genome. (C-D) Association peaks in pleiotropic loci SLC9A9 and AIM1. The upper and the lower panels represent associations for LOAD and PD, respectively. Monocyte-specific functional regions are highlighted by red dots at the bottom of the figure above the gene annotations.
Fig 6.
Reprioritize AD GWAS loci using functional annotations.
(A) Effect size estimates for 10 SNPs of interest in the discovery and the replication cohort. Intervals in the discovery stage indicate 95% confidence. Asterisk indicates significant effects in the replication cohort. Red and green squares highlight loci identified using monocyte or liver annotation track, respectively. (B) The successfully replicated SCIMP locus. The vertical axis shows the GenoWAP posterior probability based on monocyte annotation track. Functional regions in monocyte are highlighted by red dots.