RE: Do your have any DIVA problems?

No, this is not true.  We routinely collect multiple millions of 
events per tube (at 20 parameters).  In an early version of the DiVa 
software (now more than 2 years ago), there was a limitation of 7 
million events (roughly) per experiment, but that has long been 
relieved.

We also don't experience the high frequency crashing that seems to 
pervade many users.  (We used to!).  Several steps to take:  One--we 
never keep any experiments in the "database" for more than a week. 
The smaller the database, the better.  Since we do all of our data 
analysis with FlowJo, there's no need to keep it in the "database". 
(Note, I use quotations around "database" because, despite what the 
intent was of the original programmers, the software doesn't function 
as a "database" in any reasonable way).  Two--we defragment the disk 
frequently.  Three--we delete old data files from the disk (anything 
over a month).	All of our data is stored on a separate TByte store 
anyway.

We only have occasional crashes, although if you watch our users, 
you'll occasionally see them flinch as the mouse moves over certain 
areas of the screen, as if they have been taught by negative 
reinforcement: "Never click here because you will lose your data"...

mr


At 9:56 AM -0400 10/19/04, Fischer, Randy (NIH/NIAMS) wrote:
>Obviously I missed this when it first came out, but is it true the LSRII
>with the DiVa is limited to collecting 100K events/tube?  Mario and Steve,
>you guys are doing >14 colors per tube, and statistically, small populations
>could not be acquired in enough numbers with this limitation. As you two
>provide the impetus for the rest of us to push the frontiers with huge
>polychomatic experiments, how does this work? Or do you just limit your
>experiments to a few tubes?
>
>I do find it hard to believe BD would not have addressed this problem before
>now, it was a problem with the original DiVa unit we had (gone now).
>
>Randy T. Fischer
>NIH/NIAMS
>Building 10, Room 6D50
>9000 Rockville Pike
>Bethesda, MD 20892
>(301) 594-3537
>fischer1@mail.nih.gov
>
>>  ----------
>>  From:	Christopher A. Worth
>>  Sent:	Friday, October 15, 2004 4:12 PM
>>  To: cytometry@flowcyt.cyto.purdue.edu
>>  Subject:	RE: Do your have any DIVA problems?
>>
>>
>>  That sounds like a plan.
>>
>>  Customer 'our cytometer that you built isn't working right with our
>>  experiments"
>>
>>  Vendor 'your experiments aren't working right with OUR cytometer.'
>>
>>  Customer 'hmmmm.... what to do? '
>>
>>  Vendor 'change your experiments!'
>>
>>  Customer 'or I could change vendors'
>>
>>  sounds simple enough to me.
>>
>>  Chris
>>
>>
>>
>>
>>  --Original Message Text---
>>  From: Akos Szilvasi
>>  Date: Fri, 15 Oct 2004 10:12:39 -0400
>>
>>
>>  I pasted below a response to our Diva problems.
>>
>>  In short there are some serious limitations "built in" to Diva that
>>  renders the LSR II to a less than perfect cytometer. 100,000 event limit
>>  is unacceptable for any cytometer. Especially not for one that can collect
>>  over a dozen parameters. If you apply complex gating to your analyses, run
>>  an Indo-1 Ca++ flux experiment or go after rare events you may need more
>>  than one or two hundred thousand cells collected. The recommended 50
>>  tube/experiment is a new feature I have not seen in the color catalog I
>>  received. I guess the delay of the HTS's release for LSR II is due to a
>>  shortage of 50 well Falcon plates.
>>
>>  What is the solution? We are very interested in BD's official response. We
>>  are about to buy three more of the new BD cytometers.
>>
>>  Akos
>>
>>  ..............................................................
>>
>>  As a general rule, when collecting 10 parameters per Tube, approximately
>>  50
>>  Tubes per Experiment can be recorded with 100,000 events per Tube.
>>  Collecting more parameters (such as Area and Height) or having multiple
>>  gates decreases the number of Tubes to 25 or fewer per Experiment. To save
>>  memory (and disk space for permanent storage), save only parameters that
>  > are being used.
>>
>>  . once they notice that Diva is slowing down,
>>  that's a sign that their open Experiment is getting too large and they
>>  should split it up into smaller experiments. They should also check the
>>  size of their database and clean it out on a regular basis.
>>
>>
>>
>>
>>