troubles about "npeaks" and detected features in multigroup

February 16, 2016, 10:26:32 AM

Hi,
Since few days I’m trying to do some experiments with XCMS online in its multigroup mode.
Actually, until now I just used 1 “96wells plate” against 27 times another “96wells plate” (numerical duplicates).
In the final annotated diffreport I observed something that I can’t understand.
-   I was expecting that “npeaks” values would be the sum of all detected peaks in the different groups. That’s not the case. Sometimes that’s ok, but sometimes it isn’t, and in that case “npeaks” is always greater than the sum.
-   To try to understand that, I looked to one of the concerned detected feature. According to the table, this feature in only detected 86 times in only 1 96wells plate, but “npeaks”=89. That’s surprising because if it is detected in that plate, it should have been detected in the others 26 copies.
-   Then, I looked the areas measured for the same well in the 27 numerical copies, and all values are as expected exactly the same.

Then, my questions are : why can I get more “npeaks” than the sum of all detected peaks, and why sometimes areas for a same feature in numerical duplicates are considered and sometimes they aren’t?

Thank you for your help.

Christophe