Running a range of thresholds vs finding quantiles of all peaks

[benayang]

Hello, I was wondering whether finding quantiles of peaks from a numerical threshold of 1 without an IgG control is the same as calling SEACR with a range of numerical thresholds. For example, instead of calling peaks with a numerical threshold of 0.01 to find the top 1% of peaks, could I get the same result by calling peaks with a threshold of 1 and then find the top 1% of those peaks based on the total signal in column 4? That way I could call SEACR just once and then find the subset of peaks that looks "best."

[mpmeers]

Hi,

You could indeed do that--using a "threshold" of 1 should return 100% of signal blocks, and you could then implement your own cutoff on the total signal reported in column 4. The numeric threshold is just a very simple way to do this automatically if you have an existing idea of what proportion you want to return.

Mike