I have been reading your comments abouts 'bad' results of all kinds from the
3700 DNA sequencer. We have also seen many of the same problems that have been
mentioned and maybe some additional ones. We have seen:
- broad "lopey" peaks from the beginning with no accurate reads.
-samples which start out OK but then go "lopey"
-samples which appear to emerge late from the capillary, show no good
sequence
-blank samples
-samples that start out OK and then just stop (as if from short template)
-multi-color spikes that are off-scale which obscure base calling for about
three or so bases. These are supposed to be air, but do not show as air in
the array view upon close scrutiny post-run.
The bad news is these have all been exhibited from a full plate (i.e., 96
capillaries) of Big Dye Terminator Standard, properly prepared, for the 3700
instrument. We had been seeing these patterns in our customers' samples for
some time (probably from the start of use of the instrument) and had been
advised they were samples of "poor" quality, or dirty. We do not perform the
reactions for our customers - they do them. So, we could believe that was true
until we ran the plate with only Big Dye Standards.
What we have been told by service is the specifications for this instrument is
that if 85% of capillaries from a run of standards pass - then the instrument is
OK, i.e. it passes specifications. So, what this tells me is that on the
average we can expect any of these patterns of failure at any time - they have
not repeated themselves on the same capillaries...yet.
Therefore, we keep a running tally of each capillary for each run as to whether
that particular capillary shows;, blank, lopey, poor quality, retained, too many
"air" spikes, short sequence. Those samples are subject to be re-run depending
upon the customers wishes. And...if a particular capillary shows a problem
three times in a row then we will run Big Dye in order to decide if it should be
blocked out. It hasn't happened yet that we have had to block a capillary.
These problems are just random.
PE Biosystems has been very cooperative in sending service engineers to help
find the source of these problems but no definite source has yet been found in
our lab.
I would like to suggest that as many of you that can perform a run with nothing
but Big Dye standard in all wells and then report your AUTOSCORE results.
Thank you,
Millie McAdams
919-684-2652
919-684-5458 FAX
mcada002@mc.duke.edu
This archive was generated by hypermail 2b29 : Thu May 18 2000 - 09:47:35 EDT