On Jun 5, 2:12*am, Mitch Haynes <mitchhay...@gmail.com> wrote:
Sent: Wednesday, July 19, 2006 9:06 AM
To: Cytometry Mailing List
Subject: Re: FloJo and DiVa file IDs
Ray,
I had this discussion with BD a while ago. This is something that
blindsided me. We actually have an SOP that we follow for tube IDs
such that they are all unique. Basically, we use accession numbers of
the format FCM00001 for experiment names and then the tube IDs would
be FCM00001_001, etc. In the older versions of Diva, this was the
"FCS" file name that showed up in the folder when experiments were
exported. However, those weren't "real" FCS files. Evidently, people
were trying to import these bogus FCS files back into Diva, or
elsewhere, and it was creating errors. In their infinite wisdom, BD's
fix was to change the names of these files to something unrecognizable
to the users so that they wouldn't be tempted to do that. They have a
system of generating their own pseudo-unique IDs in this manner. I
say pseudo-unique, because it isn't clear to me where/how the
numbering starts if, for instance, you have several instruments
running Diva in the same lab, or if upgrade to a new version by doing
a clean install. Or perhaps the machine crashes. Can the numbering
be set to pick up where if left off? I think what they really mean is
unique to that particular incarnation of the database on that local
machine which isn't unique enough in my mind. This really screwed up
my system. So asked the same question you just did, "Is there a way
to retain my Tube IDs and/or can I configure Diva to use the same
unique filenaming convention that I do". Of course, the answer was
no. Even worse, even if you export as FCS files, although
recognizable, the files aren't exactly the same as the entered ID.
The specimen name and an underscore are added before the tube ID and
".fcs" is added to the end. This is more aggravation for me to deal
with in regard to data management in a GLP setting. It's too bad. We
like exporting entire experiments for archiving purposes, but it is
easier to reconcile file name/tube ID discrepancies when exported as
FCS, so we will probably have to make a change. At present we've just
been adding a file note to the archive documents and explaining that
the original naming scheme will reappear when imported back into
Diva. This isn't entirely true when FCS files are imported into Diva
since you have to import them into an open experiment, which could be
named differently than the exported FCS files. Again this is an issue
for me in a GLP validated instrument setting and would warrant more
workarounds or further explanation as to why reimported raw data
doesn't look "exactly" as it did when collected.
Let me know if you here anything different from what I've told you as
the situation seems to change frequently and this may no be the most
current info.
Regards,
Dave
David McFarland
Principal Scientist
GlaxoSmithKline
"Ray Hester" <rhes...@jaguar1.usouthal.edu>
17-Jul-2006 16:16
To "Cytometry Mailing List"
<cytome...@flowcyt.cyto.purdue.edu>
cc
Subject FloJo and DiVa file IDs
Can someone tell us how to export DiVa files as 'Experiments', not as
'FCS files', so that the individual tubes within the 'Experiment'
retain
their specific identities, e.g., 'T4/8', rather than some numerical
value, e.g., '2437'. When we export DiVa data as 'FCS files', the
specific tube ID is retained.
We have a FloJo 7.1 USB dongle and it's being used primarily on PCs.
Thanks.
Ray Hester
Univ. of South Alabama
