There are, currently, no tools for building molecules in VNL, but there's a lot of other good software packages out there for that, or you can probably just find ready structures of DNA bases as XYZ files via Google. Once you have that, you can import XYZ files into VNL for use in transport structures.
DNA as a whole is a bit large for real calculations, but base pairs are ok.
I'm not sure how you set up the band gap calculation, which parameters and what structure... If it's a basic small crystal, one can usually do a band structure in ATK within a few minutes, but of course it can take longer if it's a bit larger structure. You may want to try the SemiEmpirical method, it's a lot faster, and quite accurate in many cases.
