RDKit Diversity Picker

Picks diverse rows from an input table based on tanimoto distance between fingerprints. The picking is done using the MaxMin algorithm (Ashton, M. et. al., Quant. Struct.-Act. Relat., 21 (2002), 598-604). The algorithm is quite fast, even for large datasets, but note that runtime increases rapidly with the number of rows to be picked.

Options

Molecule or fingerprint column (table 1)
The column containing the molecules or fingerprints to pick from. If molecules are selected their fingerprints will be calculated automatically with Morgan, Radius 2, 2048 bit length.
Molecule or fingerprint column to bias away from (table 2)
The column containing molecules or fingerprints to bias away from. This option has the effect of seeding the diversity pick: Molecules selected will be diverse with respect to these biasing molecules as well as each other. If molecules are provided as input their fingerprints will be calculated automatically based on input of table 1. If table 1 has fingerprints with unknown settings this calculation will fail. In this case please regenerate fingerprints in table 1 with the RDKit Fingerprint Node or select a compatible fingerprint column in table 2 instead of a molecule column.
Number to pick
Number of diverse rows to pick.
Random seed
Random number seed to use.

Input Ports

Icon
Table with either molecule or fingerprints for diversity picking
Icon
Table with either molecules or fingerprints to bias away from

Output Ports

Icon
The results of the diversity pick

Views

This node has no views

Workflows

Further 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.