Dear Nina, All,

We finished MakeMNA/MakeQNA service, and presented it:

http://195.178.207.160/OpenTox/MakeMNA
http://195.178.207.160/OpenTox/MakeQNA

Could you check and give comments? And how could we integrate our service
into ToxCreate? 

Example POST:
 
curl -X POST -d
dataset_uri=https://ambit.uni-plovdiv.bg:8443/ambit2/dataset/2765 -d
dataset_service=https://ambit.uni-plovdiv.bg:8443/ambit2/dataset
http://195.178.207.160/OpenTox/MakeMNA

But as you asked me,  if the MNA are still described as one string feature,
nobody else will be able to make sense of them ... We are suggesting such
structure: representation line contains only 1 or 0, which means presents or
absence MNA descriptor. For each structure we'll generate fixed number of
MNA descriptors for example 500. So each partners could be used this string
us independent variables (which the number 500) for regression analysis. For
QNA representation Chebyshev polynomials will be used. So each string will
be include 100 independent variables. Variable means  Chebyshev polynomial
value obtained QNA descriptor. 

And more, how kind date format do you use for uploading data for lasar
regression? 

Best regards
Dmitry