Extended Data Fig. 4: YTHDC2, m6A-modified HERV-H RNAs, and LTR7 DNA physically interact with each other.

a, Heatmap showing enriched YTHDC2 binding (normalized against input control) along all the HERV-H transcripts, determined by RIP-seq. b, Aggregated reads counts of YTDHC2 eCLIP-seq data showing enriched YTHDC2 binding across all HERV-H RNAs. Red line: YTDHC2 pulldown, blue line: Input signal. c, A representative genome browser snapshot illustrating the binding of YTHDC2 on one represent LTR7/HERV-H locus revealed by YTHDC2 eCLIP-seq and RIP-seq. d, The 1,119 mRNAs associated with enriched YTHDC2 binding (YTHDC2 RIP/input fold enrichment > 2, p-value<0.05) are expressed at a comparable level with those mRNAs with less enriched YTHDC2 binding. P values determined by two-sided Wilcoxon signed-rank test. n = 2 biologically independent experiments. In the plots, center lines represent the median value, box limits the 25th and 75th percentiles, and whiskers denote minima and maxima (1.5× the interquartile range). e, A ‘CCGGACGC’ motif harboring the conserved m⁶A motif GGACU is enriched in the 1,119 mRNA sequences associated with enriched YTHDC2 binding. f, Coomassie blue staining showing the recombinant protein of the YTH-___domain of human YTHDC2 (1288-1418aa) purified from E.coli. g-i, MAGRI data showing the patterns of HERV-H RNA interacting with the indicated DNA sequences. HERV-H RNAs interact with the entire LTR7/HERV-H elements while exhibiting strongest interactions with the 5′- promoter LTR7s (TSS LTR7) (g). HERV-H RNAs were separated based on whether they are bound by YTHDC2 (h, blue line) or modified by m6A (i, blue line). The YTHDC2-bound and m6A-modified HERV-H RNAs exhibited stronger interactions with their TSS LTRs compared to those without YTHDC2 binding or m6A modification.