A bovine GWAS reveals determinants of mobilization rate and dynamics of endogenous retroviruses

We develop a novel method to measure the rate of ERVK[2-1-LTR] mobilization in the male germline, and show that this rate averages 1 per 150 sperm cells, yet varies more than ten-fold between animals. We perform a genome wide association study with 430 bulls and identify eight loci that affect ERVK[2-1-LTR] mobilization rate. We provide strong evidence that ERVK[2-1-LTR] elements in four of these, cause the association. We generate an exhaustive catalogue of ERVK[2-1-LTR] elements for the bovine, and show that is composed for 15% of 10 Kb <em>C</em>-type autonomous elements encoding GAG, PRO, POL and ENV genes, and 85% of 6.8 Kb <em>D</em>-type non-autonomous elements devoid of functional genes. We show that &gt;25% of the variance of ERVK[2-1-LTR] mobilization rate is determined by the number of <em>C</em>-type elements in the genome, yet that <em>de novo</em> ERVK[2-1-LTR] insertions are dominated by <em>D</em>-type elements.