Forcing 3700 to do wrapped runs in 384 well plates

Phillip San Miguel pmiguel at
Tue Dec 14 08:48:39 EST 1999

    I'm not sure if this is common knowledge among 3700
owners--but I didn't know until last week, so I'll presume
most of us don't know. If you use a 384 well plate on the
3700 it divides the plate into 4 quadrants and creates a run
for each quadrant. This is what one would expect. However,
its default quadrants are columns 1-6, 7-12, 13-18 and 19-24
instead of "wrapped" 96-well plate sectors[1]. This may not
matter if you use formamide, but it creates a problem for
samples resuspended in water. With water-resuspened samples,
it is necessary to add extra water to the wells of samples
in the later runs to compensate for the evaporation before
they are loaded. Using a Hydra (or any standard 96-well
pipettor) to add extra water in columns 7-12 won't work. So
one must either use an 8 channel pipetor by hand or program
a Biomek (for example) to do it.

    But there is a way around this. You can force the 3700
to load any way you would like[2]. The 3700 collection
software automatically orders runs in the alphabetical order
of their run modules[3]. To force the 3700 to load they way
you want: just make four copies your normal run module using
the "save as" option from the module creation wizard and put
alphabetical prefixes on the names. I used "A1, A2, B1, and
B2" as the prefixes[4]. Then I made the plate record in
excel spreadsheet. I put the wells in wrapped quadrant order
[5]. This makes it simple to assign the correct module to
each quadrant as a block in excel. Then I export the file as
a tab delimited text file with a ".plt" extension and
imported it into the data collection program as a plate
record. After I link the plate record I check the pane with
the list of runs that are scheduled by clicking on each run
and seeing that the 384 well icon produces the desired
"checker-board" blue and white pattern. It works fine.

    Is it worth the trouble? Well it isn't much trouble
after you get it set up. And a 384 well plate costs less
than a 96 well PE plate. (Yes, that's right, per sample, the
96 well PE plate is more than four-fold the price of a 384
well plate.) So I think it is worth the trouble.

Phillip San Miguel
Purdue Genomics Center

[1]By "wrapped" I mean the interleaved sectors A1, A2, B1
and B2. That is, the sectors as defined by a standard 8 x 12
(96-well) pin tool with 9 mm spacing.
[2]Within the confines of "paired wells", of course--the
loading robot has two sample needles, with 18 mm spacing,
and it draws samples into both simultaneously and loads them
simultaneously. So, for example, wells A1 and A5 (of a 384
well plate) will always be in the same run. I don't know
anyway around that.
[3]I didn't know this either, at first. It can be irritating
under certain circumstances. But as long as you know about
it, then it is a feature, not a bug.
[4]By "prefix" I mean whatever you choose to type in front
of the module name.
[5]I.e., A1-O23, A2-O24, B1-B23, B2-B24 -- I hope this is
making sense, if not I can go into more detail.


More information about the Autoseq mailing list