search

RNA-FRETools / MASH-FRET

MATLAB package to analyze single-molecule FRET data
https://rna-fretools.github.io/MASH-FRET/
GNU General Public License v3.0
8 stars 2 forks source link

Poor model selection on histograms #50

Open supazel opened 5 years ago

supazel commented 5 years ago

Suggested configuration according to BIC is wrong In the Histogram analysis tap, state configuration, the suggested configuration according to the BIC criteria is wrong.

To Reproduce Steps to reproduce the behavior:

  1. Go to 'Histogram analysis'
  2. Click on 'Start analysis'
  3. See suggested configuration

Desktop (please complete the following information):

Susann Zelger-Paulus susann.paulus@uzh.ch 20190822_error_BIC.pptx

mca-sh commented 5 years ago

Hi supazel, Thank you for the report. What is it you observed that makes you think the suggested state configuration is wrong? May be it is a misinterpretation of the BIC plot. On the pptx you attached, the axis BIC=0 is located on the right and the minimum BIC is obtained for 1 state (the longest bar).

mca-sh commented 4 years ago

Hi supazel, I understand that the suggested configuration is not satisfying, and that by eye, two Gaussians would seem to describe the histogram better. However, it seems to be a limitation from the method used in MASH-FRET. Could you send me the corresponding .mash file so I can verrify this? Thank you again for the report.

pflanzenhueter commented 3 years ago

Dear all, can we close this case? Best wishes Richard

mca-sh commented 3 years ago

Hi Richard, I was thinking of leaving it open to anybody who is having the same concern Mélodie