I'm part of a Brazilian group inside Fiocruz that creates new methodologies for shotgun proteomics data analysis. We already have several software solutions that are available at our lab's homepage (http://proteomics.fiocruz.br/), like PatternLab for Proteomics, for example.

In my opinion, this comes down to resources. The really big proteomics labs will have bioinformatics people to write, modify, or fill existing software gaps. Many times, they make these available to the broader community.

Other smaller labs also release open source software. These fill a niche, but often making a pipeline can be complicated with this approach. Also, this software may well work on the developers PC, but is near impossible to get running on your own PC. Unfortunately, this happens far too often in my experience.

Obviously vendor software comes at a considerable cost, but when you consider the time involved in developing your own, and the cost of employing a competent programmer, it all adds up. Vendor software is also more likely to be well documented and comes with support.

We use a variety of vendor software and open source software. We do not have the bioinformaticians to allow us to generate our own software. I think we'd use them more to fill the gaps between different software applications, rather than write something from the ground up.

Expertise, cost and time are some of the factors that keep most of the proteomics labs away from developing their own software. Vendor software are usually user friendly and are supported well. However, for specific requirement some labs use their own software such as proteogenomics etc. One can actually count the number of labs that use their in house algorithms and other software tools.

Thanks everyone for your response. Data analysis is a tough job and with so many software and possibilities, selecting best one is indeed a challenging task. We also use vendor supplied software at the moment. I am glad to note that most of you agree with my point of view. Thanks again!