It has recently become possible to directly estimate the germ-line de novo 11 mutation (dnm) rate by sequencing the whole genome of father-mother-12 offspring trios, and this has been conducted in human 1-5 , chimpanzee 6 , 13 mice 7 , birds 8 and fish 9 . In these studies dnm's are typically defined as 14 variants that are heterozygous in the offspring while being absent in both 15 parents. They are assumed to have occurred in the germ-line of one of the 16 parents and to have been transmitted to the offspring via the sperm cell or 17 oocyte. This definition assumes that detectable mosaïcism in the parent in 18 which the mutation occurred is negligible.However, instances of 19 detectable mosaïcism or premeiotic clusters are well documented in 20 humans and other organisms, including ruminants 10-12 . We herein take 21advantage of cattle pedigrees to show that as much as ~30% to ~50% of 22 dnm's present in a gamete may occur during the early cleavage cell 23 .
CC-BY-NC-ND 4.0 International license not peer-reviewed) is the author/funder. It is made available under aThe copyright holder for this preprint (which was . http://dx.doi.org/10.1101/079863 doi: bioRxiv preprint first posted online Oct. 9, 2016;
Mosaic DNMsPage 2/27 divisions in males and females, respectively, resulting in frequent 24 detectable mosaïcism and a high rate of sharing of multiple dnm's between 25 siblings. This should be taken into account to accurately estimate the 26 mutation rate in cattle and other species. 27To study the process of dnm's in the cattle germ-line, we sequenced the whole 28 genome of 54 animals from four pedigrees. Grand-parents, parents and offspring 29 (referred to as probands) were sequenced at average 26-fold depth (min = 21), 30 and grand-offspring at average 21-fold depth (min = 10). The source of DNA was 31 venous blood for females and sperm for males. The genome of one male proband 32 (Pr 1) was sequenced both from semen (26-fold depth) and blood DNA (37-fold 33 depth) ( Figure 1A). 34Using the standard definition, we identified 190 candidate dnm's as variants that 35 were (i) detected in a proband, (ii) absent in both parents (and grand-parents 36 when available), (iii) transmitted to at least one grand-offspring, and (iv) not 37 previously reported in unrelated individuals from the 1,000 Bulls project 13 38 (Suppl. Figure 1&2 and Suppl. Table 1). For confirmation, we developed 39 amplicons spanning 113 candidate dnm's and sequenced them at average depth 40 of ~2,187 in the 54 animals plus 55 relatives ( Figure 1A). This confirmed the 41 genuine nature of 110/113 variants, demonstrating the excellent specificity of 42 our bioinformatics pipeline. The three remaining ones were also detected in one 43 of the parents (although not in the grand-parents) in the confirmation, and 44 momentarily ignored. 45We first examined what proportion of dnm's detected in a proband might 46 actually have occurred during its development rather than being inherited via 47 . CC-BY-NC-ND 4.0 International license not peer-reviewed) i...