It's possible, just HIGHLY unlikely. Even possible "molecularly", just HIGHLY improbable.
Sure, they can create what THEY remember, but how accurate would that be?
The differences would be phenotypically different (refering to the clone analogy), I doubt they would be genotypically different (I HIGHLY doubt it and I'm not going to say it for sure when I don't know). The phenotype can only be expressed if it is contained within the genotype.
How are you going to plan-out the expression of pleitropic traits? Codominant, incomplete-dominant traits? And again, I think Og Kush is a polyploid. MAYBE if someone decoded its genome and had some genetic markers to work with, but no breeder is playing that shit by ear and recreating it. The variation is just too great.