The follow are just some comments on the recent postings related to the
3700. First I wanted to suggest that people might supply some or all of
the following information when relevant.
1)How many runs on the capillaries in total and since the last acid wash
and maybe an estimate on how frequently the capillaries were washed
2)Your conditions for loading your samples(buffer used, injection
parameters and foil piercing or not)
3)running conditions (polymer type and lot number, voltage, capillary
temp and cuvette temp).
4)an estimate on range of phred >20 scores if you use it.
Our 3700 is currently being checked out by ABI(PEB, perkin elmer etc.)
because of our concerns on how consistently the machine performs. I
have not run it for a while now and the above parameters were changed
during our trouble shooting period so I won't list these things at the
moment.
On the observation by Stuart that the data quality tails off at 200-300
bases. I think the two possible conditions you describe is that there
is signal but the resolution becomes poor or that the signal is top
heavy(signal at start that fades too early).
On the top heavy signal you might try running the sample again. If it
works better the second time it would suggest that there were too many
small ions either because of the sequencing reaction (which you should
see on the gel also) or because of salts in the sample. Many people
still resuspend their templates in buffer which I believe does not help
and EtOH percipitation on the cleanup can leave salts in the sample.
You should also consider your loading solution and your injection
parameters.
If it is a poor resolution it can be either the capillary or the sample
or both. We have found (after running ALOT of pGEM plates) that after an
acid wash we get a good consistent phred20 (between 650-700 on all
capillaries) but after about a week that the phred can range from 600 to
350 on pGEM and this appears to have a general left to right trend. The
lower phreds being on the right capillaries. We also developed a problem
of a totally random poor resolution on about 40% of the capillaries per
run. This appeared to be resolved by changing the board that controls
the loading robot and syringes???
My understanding is that the inline filter is only for the polymer
going to the cuvette for the sheath flow. The capillary fill is not
filtered so bubbles can be a problem. I think just being careful when
changing polymer and also be sure that the polymer fill syringe does not
leak when filling is the best you can do. If you get a big bubble in
the line I would run the change polymer wizard. I did see big bubbles
develop in the POP5(I don't know why) but I don't think they got into
the line.
On reloading the sample, in the past this worked pretty good. I have
seen data where we ran the same plate three runs in a row and the phred
scores practically overlayed each other when we plotted them. However
more recently I found that the Gs were going off much quicker(I was
loading in water with no foil piercing on all runs).
We found that running a whole plate of pGEMs was very informative on
capillary to capillary performance and I would encourage everyone to
have some similar reference point. It is a pity the the Bigdye
standards are so unreasonably priced.
Hope my ramblings were of some interest. Looking forward to seeing
other comments. Thanks
Martin
