search

glasgowcompbio / ms2ldaviz

Substructural discovery in untargeted metabolomics data using LDA topic modelling.
http://ms2lda.org
MIT License
11 stars 8 forks source link

Consistency in M2M_Doc threshold #64

Closed justinjjvanderhooft closed 7 years ago

justinjjvanderhooft commented 7 years ago

The default doc_m2m_threshold is 0.05 according to the global options, but it seems to be 0.01 in the newer experiment. In the network, 0.05 is prefilled and thus the degrees do not match within the motifs and the network. See snipit from one of the urineNGS016 motifs (loss of NH3)

image

justinjjvanderhooft commented 7 years ago

adding an experiment option and setting the m2m_document threshold to 0.05 indeed changes the numbers: image So it seems like 0.01 is somehow the default value, whereas 0.05 is mentioned. I think we can go with 0.05.

sdrogers commented 7 years ago

Sounds like it's not picking up the global option. I can check that.

sdrogers commented 7 years ago

Ok, fixed. The problem was that if it couldn't find the option for the particular experiment it just randomly picked the first one it could find (maybe from another experiment). It now explicitly finds the global one. Note that this could cause crashes if the global options aren't set...