DNA printing

[Dan Bolser]

Here is my brain storm after reading some of the above....

1000 bp per second / 3,000,000,000 bp = 3,000,000 seconds = 1 month
for one human genome.... i.e. any solution has to be highly parallel
with a suitable 'linking step' to produce the final DNA.

What length of DNA fragment gives the optimal (time, cost) trade off
with a particular linking strategy?

Regulating the enzyme:
* On / Off
* One bp per 'cycle'
* Which nucleotide to add?

Regulating the availablility of nucleotides:
* Washing,
* Cycling,
* 'switching'.

So it seems we can dream up many different strategies depending on
weather we regulate the enzyme 'mechanically' (including
electronically, physicaly, or optically) or 'chemically' via a cycling
or controlling nucleotide availability.

I think trying to obtain a certain sequence stocastically is bound to
fail, so the addition of each nucleotide needs to be precicesly
controlled.

I like the idea of mechanically perturbing the polymerase using some
kind of nano teathering into four differnt states, each state likely
to 'spontainiously' introduce one of the four nucleotides.

The text about the spontainious production of DNA repeats, template
free, gives me a lot of hope that different pols with different
tendencies can be 'evolved', and then that info used in desgning such
a machine.

Can we re-engineer the polymerase entirely into a purely 'nano-
fabrication' type of design that would be trivial to control (compared
to the task of making it work, controlling it would be trivial).

Can the new oxford nanopore technology be adapted to synthesis? Since
they got their sequencer working, perhaps we can adapt their platform
for synthesis?

</brainstorm>