Skip to main content
. 2019 Sep 24;8:e46922. doi: 10.7554/eLife.46922

Figure 4. Identification of post-PGCS germline mosaicism in the second generation.

Figure 4.

(a) Mosaic variants occurring during or after primordial germ cell specification (PGCS) were defined as DNMs present in multiple third-generation siblings, and absent from progenitors in the family. (b) Comparison of mutation spectra in autosomal single-nucleotide germline mosaic variants (red, n = 288) and germline de novo variants observed in the third generation (non-shared) (blue, n = 22,644). Asterisks indicate significant differences at a false-discovery rate of 0.05 (Benjamini-Hochberg procedure), using a Chi-squared test of independence. P-values for each comparison are: C > G: 6.84e-2, T > G: 0.169, T > A: 0.236, T > C: 1.51e-2, C > A: 4.31e-3, C > T: 0.385, CpG >TpG: 2.26e-6. (c) For each third-generation individual, we calculated the number of their DNMs that was shared with at least one sibling, and plotted this number against the individual’s paternal age at birth. The red line shows a Poisson regression (identity link) predicting the mosaic number as a function of paternal age at birth. (d) We fit a Poisson regression predicting the total number of germline single-nucleotide DNMs observed in the third-generation individuals as a function of paternal age at birth, and plotted the regression line (with 95% CI) in blue. In red, we plotted the line of best fit (with 95% CI) produced by the regression detailed in (c). (e) For each third-generation individual, we divided the number of their DNMs that occurred during or post-PGCS in a parent (i.e., that were shared with a sibling) by their total number of DNMs (germline +germline mosaic), and plotted this fraction of shared germline mosaic DNMs against their paternal age at birth.