Re: [Open-ms-general] FeatureFinderMetabo parameters

[Oliver K. <oli...@un...>]

Hi Ivan,

even though I haven’t checked the code itself, my educated guess
would be that the thresholds you mention are on a peak/chromatogram
basis. Feature intensity, however, is the sum over multiple peaks.
You should expect the max height of the peaks contained in the 
feature to be no lower than noise_threshold_int. Other than that
there are no real predictions to be made on min feature intensity,
as it will depend on chromatographic peak width and number of isotope
traces identified.

Hth,
  O.

> On 31.08.2015, at 17:06, Ivan Protsyuk <iva...@em...> wrote:
> 
> Hi everyone,
> 
> I have a general question about FeatureFinderMetabo algorithm settings. There are "noise_threshold_int" and "chrome_peak_snr". According to their descriptions, it seems logical to me that the minimal intensity of features detected by the algorithm is "noise_threshold_int * chrome_peak_snr". However, minimal intensity is usually several times more than this value.
> 
> For example, if I try to find only features with high intensity, I set "noise_threshold_int = 10^6; chrome_peak_snr = 3" and expect to get features with intensity of 3*10^6 at least. However, the lowest value I see is 11 * 10^6.
> 
> So, the question is where's a mistake in my reasoning? Other words, how can I estimate the lowest intensity value produced by the algorithm based on its parameters?
> 
> Kind regards,
> Ivan Protsyuk
> 
> 
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Open-ms-general mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/open-ms-general