96 lane upgrade

Margaret Robertson (mar@howard.genetics.utah.edu)
Mon, 17 Aug 1998 16:24:55 -0600 (MDT)

I'm interested in getting some discussion/input from those labs who have
the 96 lane upgrade working. I've spoken to only a few people so far and
there is some indication that the data is far from satisfactory. I've had
the upgrade since the end of July and still am trying to get robust
results. It would be great to hear how
everyone else is getting on. Please reply to everyone at the abrf
discussion group, so that we can all share the info.
Here is my 2 cents worth:
1. Lack of resolution is the main problem. On a 4x gel, a 5% long ranger
starts to break down at about 400 bases compared to 500 bases using the
same gel conditions on the Xl upgrade-64lanes/36cm. We have used the new
rapid ranger gel mix from FMC with one good run going out to over 650 bases
with 98.6% accuracy, not reproducible, but looking promising!
The only 48cm gel we've run so far has given promising results: 48cm
4.75% long ranger, 2x, 10 hours, gave 600 bases with good resolution and
98.8% accuracy. We had to run semi adaptive base calling and force the
base spacing to 13 to get the data to look good. One channel averaging
seems to give sharper data on the 96 lane upgrade.
2. Weak signal: We are now keeping the CCD gain set at 4
We sequence 600ng DNA with 3.2 pmoles primer, etoh ppt, resuspend in
4ul dye, load 1ul ~150ng DNA per well. This gives signal strengths for
each dye in the range 200-300.
3. Gel tracking: 64 lanes tracks robustly with the new neural net
tracker, about 10% of gels need to be manually altered.
All of the 96 lane gels we have run, have required manual tracking.
4. Our computers are 7300/200 and 7200/120 with 32Mb Ram each
with virtual memory set to 50Mb. If analysis is set for automatic
analysis, even if only 64 lanes are being collected, the run pauses at
5hours into the run with a power fail message. The computer has to be
restarted and the run continues. These computers are supposed to be the
bare minimum required to run the 96 lane upgrade. We CANNOT run automatic
analysis, but we can collect data without any problems and dump the data
to our remote G3s for analysis.Today, we have upgraded memory in both
collection computers to 96 Mb each and will see if analysis can be
automatically fired off without any conflicts. The new g3 computers are
wonderful, running system 8.0, so haven't seen any conflicts there.

Thanks in advance for any useful info with the 96 lane upgrade.
Margaret Robertson
Director DNA Sequencing Facility
University of Utah.