On Oct 7, 4:38 pm, jlund256 <
jlund...@gmail.com> wrote:
> > I'm a little disappointed that their pictorial of a garage lab doesn't
> > include any automation devices. Liquid handling robots have
>
> Having tried them, I found almost no use for them on small or medium
> scale.
> pipetting by hand
> (with a multi-channel pipettor) a better choice for thousands of
> PCRs.
For low cost bio, I have been assuming that protocols could be created
which may produce lower yield with more work effort at much a lower
cost per experiment -- that is where an automated process might win
over a manual process. For example it seems most protocols now
involve use of a costly limited-use kit (with proprietary/patented
engineered reagents) as these kits boost reliability and yield in
single experiments. However if the sample is an unlimited volume
(which could be made via PCR in some cases), then yield may not be a
factor if specificity can be held constant across a larger number of
runs -- basically, repeat the same experiment many times, something a
robot is good at (well, kind of, they're supposed to be anyway) until
the noise can be measured and subtracted across all the results.
This is all theoretical and I'm not a biologist; it's just what I've
been assuming might be a great way to go. The industrial revolution
model, i.e. doing it by hand might result in better quality yet
automation can pump out more cars per day at far lower cost (with
results which may have great precision but horrible accuracy). Is it
possible? All of this would only hold true if such protocols (and
cheap reagents) exist. Kits were created to save humans time (among
other things) so protocols have been optimized to use them. Robots
have plenty of time to kill so protocols might be optimized to
leverage lots of robot time.