Author Topic: 2D analysis software  (Read 13722 times)

brian

  • Guest
2D analysis software
« on: April 04, 2006, 05:00:00 PM »
We are in the process of decding which 2D analysis software to buy. At the current state we are choosing between imagemaster and pdquest. It is very difficult to decide which is best.

Does anyone have any reasons for choosing one over the other?


Michael

  • Guest
Re: 2D analysis software (brian)
« Reply #1 on: April 04, 2006, 05:00:00 PM »
How extensively have you looked at software packages?  Some are outrageously priced and some aren't.  No packages are hands-off, despite the many claims otherwise.  The criteria that are key, to us anyway, are the quality of matching (or degree of mismatching), number of unmatched spots, and ease of use (to correct the warping/matching errors).   Another issue is uniformity of spot shape.  If you have the same spot on two gels, fold change you get will be influenced by the spot shapes which affects the volume of the spot.  Personally, I haven't used either of the packages above in a long, long time so not in their current iterations.  Some packages/companies are coming to the understanding that missing data is a huge problem (which arises from mismatching and unmatched spots), but not many have addressed it.  DIGE helps in this regard, but I'm not sure about DeCyder.  There are two companies that I know of that have directly addressed the issue.  One has a solution in place and one is either there or very, very close.

Char

  • Guest
(No subject)
« Reply #2 on: April 04, 2006, 05:00:00 PM »
ATM there is still no software package that really adresses all problems.
DeCyder is obscenely expensive and also does not adress some statistical problems (at least in the version that we have) for instance.
Another one that I am testing is 2D-Delta. From what I have seen it has at least all functionalities from other packages, and is quite cheaper (around a third of the price tag from DeCyder I think). The best thing however is that they are apparently quite eager to add suggested features into their packages (I suppose that is also because they collaborate and are in fact partly derived from a nearby proteomics group).
After having used (an admittedly two years old version) PDquest and 2D-Delta I'd go for the latter.

Dambort

  • Guest
Re: (Char)
« Reply #3 on: April 04, 2006, 05:00:00 PM »
I should not say this. But, as you know from previous posts my answers are quite direct. I may run into problems with some suppliers, but anyway... I have experience in phoretix (nonlinear dynamics, also called image master) and PDQuest, but both software packages have difficulties in auto-modi, like the spot warping algorithms. This is the main problem, because run-to-run variations cause spot clusters to move with gels. For PDQuest I had to do manual spot matching in order to correct for this problem. Many people recommend to use proteomweaver. Unfortunately I do not have experience with this software, but I think it is possible to download it for free!

best regards

Ambort


Brian

  • Guest
Re: (Dambort)
« Reply #4 on: April 04, 2006, 05:00:00 PM »
Thanks for the comments. Ambort, I believe we are many who really appreciate your direct replies!
Concerning proteomwaever I looked for it but had difficulties in finding it until I got redirected to Bio-Rad's homepage where it appears as a "coming soon" product. So, I guess it is to late to get it for free.

At the moment we are staining our gels with colloidal CBB. But we have ambitions to use DIGE in the future. As more of you mentioned Decyder is very expensive and perhaps has some problems. We do have access to a single license decyder at our institution and several groups are using it, so we need higher capacity so to speak. what is your opinion on the DIGE functionality that have been implemented to PDQuest and Imagemaster for example (suggestions to other software packages are very welcome!)


Dambort

  • Guest
Re: (Brian)
« Reply #5 on: April 05, 2006, 05:00:00 PM »
Aha, I suppose BioRad is going to replace PDQuest by Proteomweaver! For DIGE Decyder was recommended. I do not know if it makes sense to use non-DIGE specific software like PDQuest or Image master for image analysis of DIGE gels!

best regards

Ambort


Char

  • Guest
(No subject)
« Reply #6 on: April 05, 2006, 05:00:00 PM »
Embarassingly I have to add that the software from Decodon is not 2D-delta but Delta2D.
Proteomeweaver has a good reputation but is also quite expensive.
I suppose the best you can do is to download demo versions and try to match your gels with it.
I am not sure if the new version of PDQuest or Imagemaster (never used that before) has implemented DIGE analysis, but Delta2D has.

Edit: just browsed over to the ImageMaster site and found that the current ImageMaster version also supports Dige analysis.


Modified by Char at 12:58 AM 4/6/2006


Brian

  • Guest
Re: (Char)
« Reply #7 on: April 05, 2006, 05:00:00 PM »
Yes both imagemaster and pdquest supports DIGE in their new versions. IE you can buy a reasonably priced version that does not support DIGE analysis and a very expensive one which does. So, I was interested in your opinion about their performance in DIGE analysis., but I guess it is so new that no one have really tried it our yet.

Best regards,

Brian


domthe

  • Guest
(No subject)
« Reply #8 on: April 06, 2006, 05:00:00 PM »
Hi
I have used PDQuest in my earlier laboratory for normal gels (not for DIGE). Now in new lab, there is Delta 2D software, I have not used it. I plan to do DIGE soon, and perhaps I will use Delta 2D for analysis.
Could anyone explain what you mean by 'DIGE functionality' of software. Can we use a normal software (one without DIGE function) for  warping and comaprison for images taken from a DIGE expt? I mean take images (3 images for all the three Cy dyes)  and compare using normal software?

Than you for the advice.


Char

  • Guest
Re: (domthe)
« Reply #9 on: April 06, 2006, 05:00:00 PM »
In theory yes. The problem is actually the normalization using the internal standard, which is not supported in some systems. In theory you could export the data from each channel and use an external programme for normalization and then re-import the normalized data, though.

Reiner Westermeier

  • Guest
Re: (Char)
« Reply #10 on: April 09, 2006, 05:00:00 PM »
There are in reality only two software packages which support spot co-detection accross the different channels:
1) Ettan DeCyder
2) Imagemaster Platinum Version 6.0 DIGE enabled.
If you do not have the capability of the co-detection algorithm, you cannot make the real use of DIGE.

Sjouke Hoving

  • Full Member
  • ***
  • Posts: 176
Re: (Reiner Westermeier)
« Reply #11 on: April 10, 2006, 05:00:00 PM »
Progenesis from Nonlinear Dynamics can handle DIGE gels too, but this is not the cheapest option.

Sjouke


brian

  • Guest
Re: (Reiner Westermeier)
« Reply #12 on: April 12, 2006, 05:00:00 PM »
Reiner, are you saying that the new version of PDQuest (8.0 advanced) do not really support spot co-detection accross different channels (or DIGE as they claim)????

hansv

  • Guest
Re: (Reiner Westermeier)
« Reply #13 on: April 16, 2006, 05:00:00 PM »
Dear Reiner

It seems your -highly valued- opinion here is a bit troubled by your affiliation. What you call co-detection across channels is in reality simply a projection of spot boundaries across channels. The spot is detected in one (the Cy2) channel and the boundary is copied to the other channels. That is something that other (non-GE) software supports as well, e.g. Progenesis. The recurring discussion on what the best image analysis software is underlines the current situation: there is no software around that is clearly better (i.e. with different datasets) than another. I think it is generally accepted that Decyder is not a good choice, because of its poor performance in gel-to-gel matching. And as others have noted, it is a DIGE-only software. On the other hand, it is probably the package that is best integrated with Typhoon and most easy-to-use, so it might be a good choice if you use only DIGE with small datasets (and if you can get a good deal on the scanner/software package). The main other players, Progenesis, Melanie (equals the latest Imagemaster), PDQuest, Proteomeweaver and possibly Delta 2D all have their strengths and weaknesses and it depends completely on the kind of application, the degree of spot editing, the user group (few experts or many novices) the integration with spotpicker, the amount of money etc. etc. what the best choice is. Currently none of these products (I can't give an opinion on Delta 2D) would be a 'wrong' choice and certainly none of these providers can claim to have a clear advantage over another. And sooner rather than later all of these will be completely DIGE-compatible, if they aren't already, which means: automatic recognition of subgels based on DIGE nomenclature, 'co-detection', normalization based on the internal standard etc.
Sorry not to be more helpful
Hans


PeterCampbell

  • Guest
Re: (hansv)
« Reply #14 on: April 16, 2006, 05:00:00 PM »
I have just done my frist DIGE experiment and used DeCyder 5 for the analysis.  There is also a more expensive version 6.  I have been very impressed by the normalisation that is done using the reference/Cy2 image of each gel.  This has allowed statistics I am confident of and also make good biological sense.  Also I am much more confident of the spot matching between gels using the Cy2 channel that should be the same between all gels.  Where one gel is warp relative to another it is possible to get the matching very precise.  BUT! The software is still relatively dumb when it does the matching; it just uses an array of spot positions ignoring their relative intensities.  Knowing that the reference patterns (intensities and positions) should be identical with ideal gels is one of the great advantages of having a standard and has allowed confidence for manual matching.  I have had to do A LOT of manual matching in iterative rounds of ever fine tuning the matches and re-running the stats analysis.  It is just a pity the software didn't use a more sophisticated matching of of the reference channel images at the start.
Still I am overall very happy with my first attempt at DIGE.
Peter.