Re: Free LIMS for genetic analyses and beyond anywhere?
Hi Johan,
On 11/28/2011 10:21 AM, Johan Henriksson wrote:
> you're touching one of the most underdeveloped parts of free software!
A LIMS in a production environment is just fine. But in a University, when
new assays are developed all the time and compared and samples
come from various sources with various diseases in various forms,
the LIMS are very important and at the same time LIMS maintainers
intrinsically get into bits that noone has seen before. After all, we do
all think to be doing something special, some call it research. When
you pair that with time pressure to get just anything out and some
programming skills, or fairly cheap TA, pre- or post-doc labour, I
think this is why so much redundancy there is.
> for images: use OMERO http://www.openmicroscopy.org
> this is The coming standard for microscopy images
Ack.
> LIMS: difficult. to give a concrete suggestion:
> * wasn't there a talk at the last BOSC mentioning a system developed at
> EBI? (I managed to track it down once)
Sanger (same campus), yes. I just wished for someone else
"here" to be using it off-Hinxton.
> * more hardcore: would TRAC work?
> * database management: for minimal stuff, OO base and Kexi
Well, Kexi and OO-base should also merge somehow, not?
LIMS are not alone with their redundancies :)
> but this is making it easy. the problem starts as soon as you mention
> "adapted for our needs"; whenever someone says LIMS I think everyone wants
> their own thing. This goes in the far opposite direction of having a system
> everyone can use. I think a LIMS should integrate with the other local
> databases. It should also be able to operate with some electronical labbook
> (another underdeveloped area). I don't think there is a free system out
> there which can do this and is flexible so other labs can adapt it. the
> systems made are all made for in-house use.
> so by all means try to reuse something. but unless there is a system I have
> missed I think we need to develop something new. my take would be something
> like http://en.wikipedia.org/wiki/ENOVIA_MatrixOne where the first step is
> to define a sample workflow and roles. but then the design space is pretty
> open. you're not the only one needing a LIMS; would it be possible to
> collect more interested labs to design something together?
I had contacted some of the NCI folks doing the ca* databases Sebastian
had pointed to and got positive vibrations back. In terms of
"inter-site-collaborativeness" they seem difficult to beat.
Their download for caTISSUE alone is 500MB, though. So, some
"apt-get install" and say a reduction of redundancies in Java libraries
would
be an enormous help, I presume. But I did not even download it all, yet,
to make any informed guess about how difficult it would be to come up
with a working package. Anyway, given that it all would need to be a
side project for all of us, I do not see us maintaining any such beast,
at least not for the longer run, without having upstream truly with us
and basically leading the effort.
It would be very interesting to see a collaborative LIMS development that
finds its way beyond a bunch of groups that are centrally paid for
collaborating. Presumably, the way goes through some standards for
exchanging data between LIMS. One would then have multiple compatible
but independent installations with different groups. The central theme
would then be to not have it centralised.
Let's keep this thread open and collect opinions and experiences for a
while.
Steffen
> On Fri, Nov 25, 2011 at 1:39 PM, Steffen Möller <steffen_moeller@gmx.de>wrote:
>
>> Dear all,
>>
>> our lab's sample management does not scale and needs an overhaul. The
>> idea is to migrate to an Open Source solution that fits the core needs
>> (what sample sent by whom in what form in what freezer last used by whom
>> for what project) and then have this adapted to our further needs.
>> "needs" most likely mean image data (antibody staining of various tissues).
>>
>> There are various solutions within the reach of google. But if anything
>> matching our core needs is already close to our community, then anything
>> else would find it difficult to impress me. My hope is that we can get
>> something like it into our distribution in a not too far future. After
>> all, it is such bits of software that link our medical informatics folks
>> of us with the biogeneticheminformatical other half.
>>
>> Many thanks in advance
>>
>> Steffen
>>
>>
>> --
>> To UNSUBSCRIBE, email to debian-med-REQUEST@lists.debian.org
>> with a subject of "unsubscribe". Trouble? Contact
>> listmaster@lists.debian.org
>> Archive: [🔎] 4ECF8C64.5030903@gmx.de">http://lists.debian.org/[🔎] 4ECF8C64.5030903@gmx.de
>>
>>
>
Reply to: