Extended Data Fig. 5: METTL3 regulates SETDB1–TRIM28 recruitment.

a, Venn diagram showing overlaps of the IAPEz elements bound by METTL3, SETDB1 and TRIM28. b, Scatter plots showing correlation of METTL3 and SETDB1 (left) or TRIM28 (right) on IAPEz-int elements (n = 2,542). Two-sided Pearson’s correlation test. c, Box plots showing SETDB1 (left) and TRIM28 (right) enrichment levels on IAPEz-int elements in parental and Mettl3 KO cell lines. P = 0 (left), P = 0 (right), two-sided paired t-test. d, ChIP–qPCR showing binding patterns of SETDB1 (left) and TRIM28 (right) on IAPEz-int elements in parental and Mettl3 KO cells. The mean of three biological replicates ± s.d. is shown. *P < 0.05, **P < 0.01, two-sided t-test. Exact P values are provided in the Source Data. e, f, Box plots showing density fold changes (log₂[Mettl3 KO/parental]) of SETDB1 (e) and TRIM28 (f) on different types of repetitive elements upon Mettl3 KO. g, Co-immunoprecipitation-coupled western blot showing interactions of SETDB1 (left) and TRIM28 (right) with reintroduced METTL3 (wild-type or catalytically mutated) in Mettl3 KO cells with or without triptolide treatment. Uniquely mapped ChIP–seq reads were used in b, c, e, f. For the box plots in c, e and f, the middle line and lower and upper hinges of the box plot correspond to the median and the first and third quartiles, respectively. The whiskers extend from the hinges to no further than 1.5 × IQR from the hinge. Outlying points are plotted individually. Blots are representative of two independent experiments in g. For blots source data, see Supplementary Fig. 1.

Source data