Fig. 7: CCDC141 nonadditively interacts with TTN and IGF1R to modify cardiomyocyte morphology.
From: Epistasis regulates genetic control of cardiac hypertrophy
a, Violin plots showing total cell diameter distributions for unaffected (blue, n = 36,327) and MYH7-R403Q variant (red, n = 46,542) cardiomyocytes from three independent cell batches (P = 2.1 × 10−36, two-sided Mann–Whitney U test). In all panels, box plots show the interquartile range (IQR) with the median line, and whiskers extend to 1.5× IQR. Overlaid points represent median diameters from individual batches. b, Gene-silencing efficiency measured by RT-qPCR in unaffected (blue) and MYH7-R403Q variant (red) cells from at least n = 3 cell batches (exact n numbers provided in Source Data). The bars, overlaid points and error bars show mean values, individual biological replicates and standard deviations, respectively. c–h, Gene-silencing effects on single-cell morphology analyzed for n = 3 independent cell batches (n = 4 for unaffected cells silencing CCDC141), with over 33,000 cells measured per condition per batch. c, Percentage change in median cell diameter for each gene-silencing condition relative to scramble controls (that is, relative size difference). The squares represent mean relative size differences across batches, with overlaid points indicating individual batches. The violins display mean relative size differences from 1,000 bootstrap samples, with error bars indicating standard deviations. *P < 0.05; **P < 6 × 10−5 (exact P values in Source Data), two-sided Mann–Whitney U test. d, Violin plots showing nonadditive interaction effects (\({\hat{\beta }}_{{\boldsymbol{12}}}\)) for unaffected (blue) and MYH7-R403Q variant (red) cells compared with marginal effects (\({\hat{\beta }}_{{\boldsymbol{1}}}\) and \({\hat{\beta }}_{{\boldsymbol{2}}}\), gray), estimated via a quantile regression model across 10,000 bootstrap samples (details in Methods). e, Scatterplots showing gene-silencing effects on cellular texture (‘norm. peak no.’, x-axis) and boundary (roundness error, y-axis) features. The markers represent mean values across cell batches, with error bars indicating standard deviations from 1,000 bootstrap samples. Bootstrapped values are visualized as density contours (scattered dots represent a density level lower than 0.2). f, Cell boundary waviness and texture irregularity were measured by roundness error (top) and normalized peak number (bottom), respectively. g,h, Representative single-cell images illustrating boundary irregularity (g) and textural variation (h). Norm. peak no., normalized peak number. Scale bars, 10 μm. Statistical details in Supplementary Data 8 and Source Data.
