Fig. 3: Several early changes, including heterochronies, correlate with supplementary cusps in the mouse upper molar.

a Percentage of mesenchymal tissue in tooth germs estimated by deconvolution of the RNAseq time series with mesenchyme and epithelium-specific marker genes (left) and with direct volume quantification in 3D reconstructed tooth germs (right). Time scale in relative 0–10 scale (x axis) for each species. Color code indicated for each molar. b Relative duration of the first morphological stages highlights a longer period of lingual growth in mouse upper molar (1-SEK, purple). Stage durations from Fig. 1b. PEK and SEK: primary and secondary enamel knots respectively. c Levels of activation of BMP, SHH and WNT pathways in buccal and lingual sides of the mouse molars at 15.0 dpc (1-SEK stage). Measurements made with an in silico method, ROMA, using two lists of target genes to estimate both an epithelial (epi) and a mesenchymal (mes) pathway activity in the 15.0 buccal and lingual RNAseq samples. Drawing on the left represents the dataset design. d Proportion of naive lingual tissue in mouse and hamster molars estimated by deconvolution of the RNAseq time series with lingual and buccal marker genes.