To what extent can de novo assembly replace current SV detection approaches (such us Pindel, Delly or CNVnator) in the detection of structural variants?