Fig. 5: INSCRIBE identifies authentic TDP-43 binding sites through confident C-to-U edit clusters.

a HOMER de novo motif discovery of TDP-43 INSCRIBE replicates identified the canonical TDP-43 motif (UGUGUG), using a cumulative hypergeometric distribution for p values. b The actual fraction of UGUGUG-containing TDP-43-INSCRIBE edit clusters/TDP-43 eCLIP IDR peaks (dots) compared with that of the 20 permuted clusters (box plot, n = 20). The actual fraction was plotted using the confident edit clusters derived from 3 INSCRIBE technical replicates. The box shows the quartiles while the whiskers extend to show the rest of the distribution; the median is represented by the center line. Z-scores: TDP-43-INSCRIBE, 22.22; Enzyme-only control, 0.33; K562 TDP-43 eCLIP, 146.11. c The distribution of the nearest UGUGUG distance to clusters of confident edits reveals a peak at 0nt for TDP-43-INSCRIBE (orange), indicating an abundance of UGUGUG sequences in close proximity to the center of the edit cluster, which demonstrates TDP43-binding driven C-to-U editing. d The actual fraction of eCLIP-overlapping TDP-43-INSCRIBE edit clusters (diamonds) compared to that of the 20 permuted clusters (box plot, n = 20). The actual fraction was plotted using the confident edit clusters derived from 3 INSCRIBE technical replicates. The box shows the quartiles while the whiskers extend to show the rest of the distribution; the median is represented by the center line. Z-scores: TDP-43-INSCRIBE, 15.61; Enzyme-only control, −0.46. e Integrative genome viewer (IGV) tracks of example RBFOX2 target genes (TARDBP, NUCKS1, XPO1, FUS) with 1500 bp windows showing RBFOX2 eCLIP peaks (orange), RBFOX2-INSCRIBE (blue) and the enzyme-only control (gray). The 3-replicate overlay of SAILOR-quantified edit fraction of INSCRIBE is shown along with the FLARE-determined confident edit clusters.