Marimba

A GUI for proteomics results and MRM experimental design

One of the continuing challenges for many proteomics researchers is how to move beyond discovery based high-throughput shutgun proteomics experiments to directed intensive hypothesis testing experiments like MRM.

One of the main limitations is a responsive and easily understandable GUI for viewing the high-throughput results with an eye towards designing the next step in the process.

Marimba is such a tool. It allows quick perusal and filtering of past shotgun experimental data, highlights the details that are important to an MRM experiment, and outputs an excel file of results that translate directly to best-practice MRM experiments. At least for LTQ data.

Status

After much wrangling with the Shoes GUI toolkit, I decided that it is too beta to even use as a demonstration project. For instance the official release of Shoes version 2 (Raisins) is not able to create new files to use as a local sqlite database to store results. This is a show-stopper. Bad docs, spotty performance and bogus candy functionality like custom fonts I can live without. Opening files I can’t.

As of 2/3/2009 the demonstration project is moving to a web application, in the interest of getting something done. Marimba Version 2 will be a stand-alone application, but what platform, who can say?

MRM?

What’s an MRM experiment, you ask? MRM stands for Multiple Reaction Monitoring experiment, a mass spectrometry type of experiment where one monitors a specific set of masses at specific points in time. The masses usually conforms to particular molecule of interest. For more in-depth information see {INSERT REFERENCE HERE}.

Most proteomics experiments want to move from discovery phase, where you are trying to find things of interest in your samples by brute force methods, to more accurate quantification and validation of the high-throughput results. This is typically an intensive manual labor process, as good hits to the massive original results must be sifted through and put into a format that is amendable to creating those MRM transitions.
As a simple example, let’s take the following spectra as an example:

Figure 1. A mass spectrum identifying a peptide. B ions is blue, Y ions in red. Hosted by Skitch!

The ions highlighted by red arrows represent the best candidates to monitor for an MRM experiment. Researchers would typically look up the ion series for a peptide (y1,y2,y3,…) and their masses (123,345,455,…), then go to each spectrum that has been shown to encode that peptide, jot down the time of the data acquisition, figure out from the ion series which masses correspond to which peak and then pul out the likely transitions. Rinse and repeat for several hundred spectra and peptides. Not fun.

Marimba strives to make this process a lot less labor intensive by showing the results in a format that is easily and instantly digestible to a researcher, pulling all the needed information into one window. Then all a researcher would need is to verify which tranistions are likely and export the resulting masses of interest into a CSV or Excel file.