Figure 5—figure supplement 1.

Figure 5—figure supplement 1.

(A) Biplots of RNA-seq expression of X. laevis versus X. tropicalis stage 9 wild-type transcriptomes, with Z-normalized expression values; the result is similar to plotting TPM directly as shown in Figure 5A. Left plot sums L+S homeolog TPM expression per gene prior to Z-normalization, middle and right plots quantify only L or S homeolog expression, respectively. Colored points correspond to strictly zygotic genes as plotted in Figure 5A. (B) Comparison of residuals per gene (laevis minus tropicalis Z-normalized expression values), i.e. deviation from X. tropicalis, for the full L+S transcriptome versus counting expression from only one subgenome. Points below the y=x diagonal are genes whose tropicalis expression is a better predictor for laevis L+S expression compared to single homeolog expression. P values are from Wilcoxon signed-rank tests comparing residuals. (C) Similar to Figure 5A, but plotting activated genes with a maternal contribution ≥1 TPM. The maternal contribution buffers the differences for many genes. Spearman’s ρ correlations are shown. (D) Barplots showing the proportion of Pou5f3/Sox3-regulated X. laevis genes also regulated by Pou5f3/Sox3 in X. tropicalis according to Gentsch et al. X. tropicalis is more likely to have concordant regulation in genes where both homeologs are regulated by Pou5f3 and Sox3 (p=8.9 × 10^–11 for X. laevis down genes, p=0.002 for up genes, p=1.7 × 10^–6 for genes with primary activation defects in Pou5f3/Sox3 loss, χ-squared tests, 8 d.o.f.). (E) Similar to Figure 4C: barplots showing the proportion of X. laevis predicted enhancers (left) and high-confidence predicted enhancers (right) across activity categories that are acetylated in X. tropicalis, additionally showing the proportion of enhancers that lift over but are not acetylated in X. tropicalis. Among successfully lifted-over enhancers, enhancers with conserved activity between L & S are highly more likely to have acetylation in X. tropicalis (p<1 × 10^–300 for all enhancers, p=1.5 × 10^–76 for higher-confidence enhancers, χ-squared tests, 4 d.o.f). (F) Barplots showing the proportion of X. laevis genes in different homeolog activation categories whose orthologs are also activated in zebrafish as part of the first wave by maternal factors, activated by 6 h.p.f., or part of the maternal contribution. Both-activated homeologs are more likely to also be activated and directly by maternal factors (primary activation) in zebrafish (p=2.6 × 10^–11, χ-squared test, 12 d.o.f.). (G) Barplots showing the proportion of Pou5f3/Sox3-regulated X. laevis genes also regulated by Nanog/Pou5f3/Sox3 in zebrafish according to Lee et al., 2013. Both-regulated homeologs are more likely to also be regulated by maternal pluripotency factors Nanog, Pou5f3, SoxB1 in zebrafish (p=2.1 × 10^–15, χ-squared test, 4 d.o.f.). Overall proportions are higher for X. laevis genes whose primary activation depends on Pou5f3 and Sox3 (right), but there is no significant difference between homeolog expression patterns (p=0.16, χ-squared test, 4 d.o.f.). (H) Barplots showing the proportion of enhancers that lift over and are acetylated in zebrafish according to Bogdanovic et al., 2012. L+S conserved enhancers have low conservation with zebrafish, but significantly higher proportion than L- or S-only enhancers (p=2.8 × 10^–58 for all enhancers, p=5.0 × 10^–4 for higher confidence enhancers, χ-squared tests, 4 d.o.f.).