Extended Data Fig. 1: METTL3 binds endogenous retroviral elements.

a, Accumulative plots showing the relative distances between peaks of METTL3 and histone modifications. b, Bar graph showing the overlapping ratios (calculated as Jaccard statistics, see Methods) of METTL3 peaks with repetitive elements. c, Accumulative plots showing the relative distances between METTL3 peaks and repetitive elements. d, Heat maps showing enrichment of METTL3, H3K9me3 and H4K20me3 on repetitive elements. e, Scatter plots showing the correlation between the enrichment levels of METTL3 and H3K9me3 (left) or H4K20me3 (right) on different subtypes (n = 277) of ERVK retrotransposons. Two-sided Pearson’s correlation test. f, Schematic representation of stitching of adjacent IAPEz fragments. g, Aggregation plots showing METTL3, H3K9me3, H4K20me3 and input patterns on IAPEz-int. h, Validation of METTL3 binding on IAPEz using three independent METTL3 antibodies. The mean of three biological replicates ± s.d. is shown. i, Schematic representation of uniquely mapped reads and randomly mapped non-unique reads. j, Aggregation plots and heat maps showing enrichment of METTL3, H3K9me3 and H4K20me3 on IAPEz-int elements with only uniquely mapped reads or uniquely + randomly mapped reads. Uniquely mapped ChIP–seq reads were used in a–c, e. Uniquely and randomly mapped ChIP–seq reads were used in d, g. Heat maps were ranked according to METTL3 enrichment in parental cells in descending order in j. MTA and MaSat used in ChIP–qPCR are examples of repetitive elements not bound by METTL3.