Ternary and contour plots showing the relative enrichment of the isoform usage ratios across ESC, PE and PGC for four genes with isoform switches (Fig. 5f and Extended Data Fig. 10d) and corresponding density. Each gene has two isoforms, where one is used predominantly in ESC (Down, deep color), and the other is used mainly in PE and/or PGC (Up, light color); 100 bootstrap sampling of cDNA-ONT reads for four genes was conducted, with sizes corresponding to real abundances of four genes (Real data) and quartiles (Q 3, Q 2, Q 1) of all gene abundances in ESC. Isoforms used solely in one of the ESC, PE and PGC will be located at the triangle vertices without contour. The contour plot is not generated for extremely sparse points such as LIQA at the 25th percentile. The distribution of all gene abundances in ESC are shown on the right and the bootstrap sampling sizes are indicated correspondingly. Due to technological issues, the FLAMES and TALON are not shown here (Supplementary Note 13).
