Fig. 3

The chromosome 3p teQTL induces polysome half-mer formation. A Schematic overview of the polysome fractionation and RNA-seq approach. One representative polysome profile per congenic rat line is given. L, M, and H fractions indicate light, medium, and heavy polysomes, respectively. B Congenic line comparison for differences in the number of associated ribosomes per mRNA, as measured by the distribution of RNA yield across the fractions. Quantified polysome profile area under curves (AUCs) can be found in Additional file 1: Figure S4A. Bars indicate mean values. C Zoomed-in view of multiple polysomal peaks across replicates for both congenic lines, with arrows indicating possible half-mers. D Heatmap with scaled RNA-seq expression levels of all 12,471 quantified genes (mean RNA FPKM ≥ 1 across replicates, for both lines). Genes are clustered into 4 groups by k-means clustering and sorted by CDS length within each cluster. The same gene order obtained through clustering of the fold change (SHR.BN-(3S) vs SHR.BN-(3L)) comparison (3rd heatmap) was used for the individual heatmaps of SHR.BN-(3L) vs SHR.BN-(3S) (1st and 2nd heatmap). For all clusters, box plots with the CDS length distribution are shown on the right. E Scatter plots and square correlation coefficients (r²) based on standardized major axis (SMA) values between coding sequence (CDS) length and the fold change in gene expression (FC (SHR.BN-(3S) vs SHR.BN-(3L)), as measured by RNA-seq of the four isolated fractions. The correlations are significant (p value < 2.2 × 10⁻¹⁶; test of correlation coefficient against zero) and the linear model based on fitted SMA method are displayed as red lines. Ribosomal protein genes (with small CDSs) are depicted by orange dots. F Heatmaps with the scaled FC of the ribosomal configuration of the top 500 shortest and longest CDS genes. G Scatterplots showing CDS length versus fold change (FC (SHR.BN-(3S) vs SHR.BN-(3L)) for Ribo-seq and RNA-seq data, highlighting a representative selection of core- and accessory sarcomere proteins. The square correlation coefficient (r²) based on standardized major axis (SMA) is calculated using expression values of this subset of genes only. H Dot plots with Ribo-seq expression values for Ttn and a selection of cardiac thick filament proteins. Genes are sorted by CDS length from top left to bottom right. Error bars indicate mean values with standard deviation (SD). None of the displayed expression changes are genome-wide significant. I Heatmaps with polysome profiling results for selected sarcomere proteins. Expression distributions for the individual animals, as well as the scaled fold changes between SHR.BN-(3S) and SHR.BN-(3L), are given. Within each group, genes are sorted by CDS length (top to bottom). J Schematic representation of the cardiac thin filament and its composition stoichiometry as obtained from [44]. Cardiac muscle alpha actin (Actc1) and cardiac troponin T (Tnnt2) are the genes most strongly translationally regulated to achieve desired protein levels. K Bar plots showing the relative contribution of each thin filament component as measured by Ribo-seq (top) and mRNA-seq (bottom) expression levels. DESeq2-normalized expression values are corrected for reported rat heart protein turnover rates [45] and represented as a percentage of the complete thin filament. Twenty healthy rats are shown (from left to right: 5× SHR.BN-(3L) congenic animals, followed by 15× HXB/BXH RI lines as separated by local BN genotype according to the Chr. 3p teQTL). Optimal production values for 7 or 1 subunit(s) are indicated by dashed lines. See also Additional file 1: Figure S4