Package org.openscience.cdk.similarity

Class Tanimoto

java.lang.Object

org.openscience.cdk.similarity.Tanimoto

public class Tanimoto
extends Object

Calculates the Tanimoto coefficient for a given pair of two fingerprint bitsets or real valued feature vectors. The Tanimoto coefficient is one way to quantitatively measure the "distance" or similarity of two chemical structures.

You can use the FingerPrinter class to retrieve two fingerprint bitsets. We assume that you have two structures stored in cdk.Molecule objects. A tanimoto coefficient can then be calculated like:

   BitSet fingerprint1 = Fingerprinter.getBitFingerprint(molecule1);
   BitSet fingerprint2 = Fingerprinter.getBitFingerprint(molecule2);
   float tanimoto_coefficient = Tanimoto.calculate(fingerprint1, fingerprint2);
  

The FingerPrinter assumes that hydrogens are explicitely given, if this is desired!

Note that the continuous Tanimoto coefficient does not lead to a metric space

Author:

steinbeck

Source code:

main

Belongs to CDK module:

fingerprint

Keywords:

jaccard, similarity, tanimoto

Created on:

2005-10-19

Field Summary

Fields

Modifier and Type	Field	Description
static String	EMPTY_FINGERPRINTS_PROVIDED

Method Summary

Modifier and Type	Method	Description
static float	calculate(double[] features1, double[] features2)	Evaluates the continuous Tanimoto coefficient for two real valued vectors.
static float	calculate(BitSet bitset1, BitSet bitset2)	Evaluates Tanimoto coefficient for two bit sets.
static float	calculate(Map<String,Integer> features1, Map<String,Integer> features2)	Evaluate continuous Tanimoto coefficient for two feature, count fingerprint representations.
static double	calculate(IBitFingerprint fingerprint1, IBitFingerprint fingerprint2)	Evaluates Tanimoto coefficient for two IBitFingerprint.
static double	calculate(ICountFingerprint fp1, ICountFingerprint fp2)	Evaluate continuous Tanimoto coefficient for two feature, count fingerprint representations.
static double	method1(ICountFingerprint fp1, ICountFingerprint fp2)	Calculates Tanimoto distance for two count fingerprints using method 1.
static double	method2(ICountFingerprint fp1, ICountFingerprint fp2)	Calculates Tanimoto distance for two count fingerprints using method 2 [J.A. Grant, J.A. Haigh, B.T. Pickup, A. Nicholls and R.A. Sayle. J. Chem. Inf. Model.. 2006. 46].

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

EMPTY_FINGERPRINTS_PROVIDED

public static final String EMPTY_FINGERPRINTS_PROVIDED

See Also:

Constant Field Values

Method Detail

calculate

public static float calculate(BitSet bitset1,
                              BitSet bitset2)
                       throws CDKException

Evaluates Tanimoto coefficient for two bit sets.

Parameters:

bitset1 - A bitset (such as a fingerprint) for the first molecule

bitset2 - A bitset (such as a fingerprint) for the second molecule

Returns:

The Tanimoto coefficient

Throws:

CDKException - if bitsets are not of the same length

calculate

public static double calculate(IBitFingerprint fingerprint1,
                               IBitFingerprint fingerprint2)

Evaluates Tanimoto coefficient for two IBitFingerprint.

Parameters:

fingerprint1 - fingerprint for the first molecule

fingerprint2 - fingerprint for the second molecule

Returns:

The Tanimoto coefficient

Throws:

IllegalArgumentException - if bitsets are not of the same length

calculate

public static float calculate(double[] features1,
                              double[] features2)
                       throws CDKException

Evaluates the continuous Tanimoto coefficient for two real valued vectors.

Parameters:

features1 - The first feature vector

features2 - The second feature vector

Returns:

The continuous Tanimoto coefficient

Throws:

CDKException - if the features are not of the same length

calculate

public static float calculate(Map<String,Integer> features1,
                              Map<String,Integer> features2)

Evaluate continuous Tanimoto coefficient for two feature, count fingerprint representations.

Note that feature/count type fingerprints may be of different length. Uses Tanimoto method from 10.1021/ci800326z

Parameters:

features1 - The first feature map

features2 - The second feature map

Returns:

The Tanimoto coefficient

calculate

public static double calculate(ICountFingerprint fp1,
                               ICountFingerprint fp2)

Evaluate continuous Tanimoto coefficient for two feature, count fingerprint representations.

Note that feature/count type fingerprints may be of different length. Uses Tanimoto method from 10.1021/ci800326z

Parameters:

fp1 - The first fingerprint

fp2 - The second fingerprint

Returns:

The Tanimoto coefficient

See Also:

method1(org.openscience.cdk.fingerprint.ICountFingerprint, org.openscience.cdk.fingerprint.ICountFingerprint), method2(org.openscience.cdk.fingerprint.ICountFingerprint, org.openscience.cdk.fingerprint.ICountFingerprint)

method1

public static double method1(ICountFingerprint fp1,
                             ICountFingerprint fp2)

Calculates Tanimoto distance for two count fingerprints using method 1. The feature/count type fingerprints may be of different length. Uses Tanimoto method from [Andreas Steffen, Thierry Kogej, Christian Tyrchan and Ola Engkvist. J. Chem. Inf. Model.. 2009. 49].

Parameters:

fp1 - count fingerprint 1

fp2 - count fingerprint 2

Returns:

a Tanimoto distance

method2

public static double method2(ICountFingerprint fp1,
                             ICountFingerprint fp2)

Calculates Tanimoto distance for two count fingerprints using method 2 [J.A. Grant, J.A. Haigh, B.T. Pickup, A. Nicholls and R.A. Sayle. J. Chem. Inf. Model.. 2006. 46].

Parameters:

fp1 - count fingerprint 1

fp2 - count fingerprint 2

Returns:

a Tanimoto distance