on magnetic separation for nucleic acids