Possible bug in descendPeaks

[Pascallio]

Dear developers,

I might have encountered a bug for an unlikely, but possible situation in MSnbase:::descendPeaks. According to the documentation, "the peak region is defined by descending from the identified centroid/peak on both sides until the measured signal increases again.". However, in situations where the centroid is either the first or last signal in a spectrum, the other side of the peak seems to be ignored.

Some example code for reproducibility:

mz <- c(100.14, 100.145, 100.15) # masses
intensity <- c(1000, 700, 600) # intensities
peakIdx <- 1 # position of the centroid

MSnbase:::descendPeak(mz = mz, intensity = intensity, peakIdx = peakIdx) 

This results in a mass of 100.14 and does not take other values than the centroid into account (which in this example would have resulted in 100.1441). It seems that MSnbase:::descendPeak calculates the half_width, which in this case results in 0 and therefore only considers the centroid itself.

Is this the correct behaviour or should the remaining signals be taken into account until the (first) observed signal rises again?

Best, Pascal

[jorainer]

Dear Pascal! Thanks for the pointing this out! Would you be willing to provide a pull request that fixes this issue?

[Pascallio]

Hi Johannes,

Yes ofcourse, I've just written a fix and provided the pull request.

Cheers, Pascal

[lgatto]

I have merged and pushed to Bioc in devel and release. Thank you very much @Pascallio !