We have got an unusual occurrence last week on a 373 Stretch sequencer. All
36 analyses were missing about 20 bases at about position 300 in their
sequences. There were almost no indication that anything went wrong with
that run, and it took us a while to detect the problem, essentially by
finding a gap when comparing data with known sequences.
We now think that this problem occurred when an automatic backup somehow
interfered with the collection of data. Has anybody ever seen that? We use
Retrospect 3.0 to backup all our machine over an ethernet network, but will
be extra cautious not to allow it to connect to the one collecting data
from the sequencer when its operating.
What surprised me the most was that the traces and base calls contained
absolutely no clue that anything had gone wrong, and the problem might very
well go unnoticed if the sequences cannot be independantly doublechecked.
BEWARE OF ANY PROGRAM OR NETWORK ACCESS TO THE COLLECTING COMPUTER!!!!
Coordonnateur des services scientifiques
1125 Charles-Eugene Marchand tel. (418) 656-5496
Universite Laval fax. (418)
Sainte-Foy, Que G1K 7P4