We are working on a paper describing the algorithms used by msInspect. One thing you should know is that they only work on high-resolution data acquired in profile-mode. They rely on persistence of isotopic profiles over time. If the data is not profile or is non-continuous over time, the algorithms may have trouble.

If you don't see peptides found that you know are there, you might just open the mzXML file in msInspect and click on the mz/elution time location of those known peptides. Do you see clear isotopic peaks in your scan? Are they persisted across scans? The right-hand pane will show you the detail and you can use the bottom pane to look at some of the intermediate results (e.g. wavelet peaks). Those might give you a hint.

We have not worked much with Q-STAR data so it is possible our algorithms have not been tuned properly for that data.

One quick note: you might want to try a different feature extraction algorithm. msInspect includes several to choose from, and the one to use depends somewhat on the properties of your instrument.

You might try something like:

java -Xmx384M -jar viewerApp.jar --findPeptides --strategy=FeatureStrategyPeakClusters --out=mixture.features.tsv mixture.mzXML

Provided that your MS1 data was acquired as Mark describes (high mass accuracy, profile mode) this will likely do better than the strategy you first used.

Even so, you will want to filter the resulting feature set (either in Excel or using the --filter option to msInspect). Note that your output contains about ~75% charge "zero" features. These are isolated peaks for which no isotopic distribution was observed; we are unable to assign correct charge to them. So you might consider filtering out features with charge less than 1. You can also use the "kl" column, which indicates how much this feature deviates from an idealized isotopic distribution; higher values indicate greater deviance and (in general) poor features. You might also remove features whose intensity falls below some threshold. The maximum "kl" score to use and the minimum intensity to use depend on the instrument.

Please let me know if this helps.

Thanks for the quick response. I will try what you suggested. The mzXML files from QSTAR wiff files don't include points with zero intensities. Those zero points can be recovered if the calibration equation is known. Can this cause trouble for msInspect?

-Peicheng

We do not rely on zeros being encoded in the input data.

Please follow up and let us know if you have any success. In the past we've had to sometimes tweak the code as users try different platforms.

I rerun mxInspect with a similar mzXML file of the same peptide mixture, with the command line of:

java -Xmx384M -jar viewerApp.jar --findPeptides --strategy=FeatureStrategyPeakClusters --out=mixture1.tsv --minMz=300 --maxMz=1500 mixture1.mzXML

The output file is attached. Why all masses are shown as "0"? Does each row mean a chromatographic peak?

In addition, if each row means a chromatographic peak, then some peaks are missing, such as m/z=483.3, which is not a weak peak, and is clearly there if viewed with ABI Analyst.

If feasible, I can attach the mzXML file in the next post for anyone to try, which is 305M in size.

Please don't attach 350M files in this forum. We perhaps need an FTP site to deal with files like that (do you have one?).

The values in the mass column should not be zero. I'd encourage you to open the file interactively see which features the tools is identifying and which it is not (but you think it should).

In the meantime, I'll look into putting up a newer build.