DMC Learner

DMC is an independent implementation of:
Godden, Jeffrey; Furr, John R.; Xue, Ling; Stahura, Florence L.; Bajorath, Juergen
Molecular Similarity Analysis and Virtual Screening by Mapping of Consensus Positions in Binary-Transformed Chemical Descriptor Spaces with Variable Dimensionality
J. Chem. Inf. Comput. Sci. (2004), 44, 21–29.
A bitstring is constructed from the descriptor values of active molecules. Bit positions are set to 0 if the descriptor value of the active molecule is below and to 1 if the value is above the median of the value distribution calculated for a huge reference set of molecules. For each bit the conservation percentage is calculated. Only bits with less than the allowed bit variability are included in the model. The prediction is done so that the variability needed to predict the molecule as active if written to the database.
The Model Evaluate node can be used to calculate predictions.

Options

Molecule Field
Field containing the active molecules
Partial Charges
Some descriptors require the precalculation of partial charges. For consistency reasons charges have to be calculated the same way for the training and the test set.
Bitvariability
Allow the maximum allowed variability for a bit to be included in the model. Please note that higher values will result in more selective models. For example a value of 10 means that positions with 90% conservations will be included in the model. If one position exposes this variability only 90% of the actives would be retrieved by the model. If two positions have only 90% conservation, only 81% of the actives would be retrieved and so on.

Input Ports

Icon
Active Molecules

Output Ports

Icon
MOE DMC Model

Popular Predecessors

  • No recommendations found

Popular Successors

  • No recommendations found

Views

DMC Learner
Summary

Workflows

  • No workflows found

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.