How is it possible to have a npeaks value higher than the total number of samples ?
A peak group simply contains all the peaks found with in an array defined by a m/z range and a retention time range.
Depending on how dense the peaks are, it is certainly possible that a peak group will contain multiple peaks from a given sample, and even more peaks than the total number of samples.
The retention time alignment algorithm takes this into account when selecting peaks to use as standards. Check out the xcms paper for more information.
Knowing the number of peaks included in a group can help provide some diagnostic information about the quality of that group. I find it useful to know.
old forum link (https://groups.google.com/forum/#!searchin/xcms/npeaks|sort:relevance/xcms/trRMFm0vXPM/8bdOpdDnXZkJ)
old forum link2 (https://groups.google.com/forum/#!searchin/xcms/npeaks|sort:relevance/xcms/gCGrricGKH4/5ksf1_6kKpYJ)
-Colin