bao.tianle

Based on that papers' calculations, the sizes of a base pair (accounting for loss of a water molecule on each strand in polymerization and a proton from each phosphate hydroxyl due to the pH) should be:

GC: 347.2207+307.1966-(4x1.0079+2x15.9994)-(2x1.0079)= 616.3711
AT: 331.2213+322.2079-(4x1.0079+2x15.9994)-(2x1.0079)= 615.383

average: 615.8771

Note that this reflects the loss of water and a proton from both strands.

Without accounting for the masses lost from the 2 waters or the 2 protons lost from the backbone phosphate, the masses end up being 654.4173 for GC and 653.4292, so an average of 653.9233.  So I could see your original estimate of 660 roughly matching the value obtained if you just counted DNA as if it were a chain of AMP, TMP, GMP and CMP.

So yeah, that paper looks right.


edit: I have found references to masses of DNA with salt bound giving an average mass for a base-pair at roughly 660, so perhaps that's where your number comes from.

I think the variability might depend on your sequence (as each nucleotide is slightly different)?

Try this Invitrogen page