Isn't the problem the absence of random DNA?

I wouldn't call random data 'complex', but it is easy to sequence when assembling short reads.