com.chemaxon.search.mcs

Class MaxCommonSubstructure

java.lang.Object

com.chemaxon.search.mcs.MaxCommonSubstructure

All Implemented Interfaces:

chemaxon.license.Licensable

Direct Known Subclasses:

BuildupMcs, MaxCliqueMcs

@PublicAPI
public abstract class MaxCommonSubstructure
extends java.lang.Object
implements chemaxon.license.Licensable

Abstract base class of the algorithms for finding the maximum common substructure (MCS) of two molecules. More precisely, these algorithms find the maximum common edge subgraph (MCES) of the input structures. For more information, see the user's guide.

An instance of the default MCS algorithm implementation can be created using newInstance() or newInstance(McsSearchOptions). The provided MCS algorithms are powerful heuristic methods, which typically find large common substructures in a short time. However, they do not always provide the exact optimal result due to the complexity of the MCS problem (especially for large molecules). Furthermore, as the algorithms perform randomized search, different results might be obtained for equivalent molecule representations.

Warning: MCS algorithms do not perform transformations on the input molecules, so you should be aware of aromatization (and other standardization actions) before using them.

Features: Query and target structures basically play the same role in MCS search except for query features: a query molecule may contain generic query atoms (A, Q, M, X, list atom, not list atom, etc.) and query bonds (any, single or double, etc.), but query properties (e.g., valence, hydrogen count) are ignored. If exact query atom/bond matching is set to true, then generic atoms and bonds are allowed in both molecules, but they are matched in exact manner. Reactions are also supported, but Markush structures are not.

Typical usage:

 MaxCommonSubstructure mcs = MaxCommonSubstructure.newInstance();
 mcs.setMolecules(queryMol, targetMol);
 McsSearchResult result = mcs.find();
 System.out.println("Atoms in MCS: " + result.getAtomCount());
 System.out.println("Bonds in MCS: " + result.getBondCount());
 System.out.println("MCS molecule: " + MolExporter.exportToFormat(result.getAsMolecule(), "smiles"));
 

Author:

Peter Kovacs (pkovacs84), Peter Englert

Field Summary

Fields

Modifier and Type	Field and Description
static long	DEFAULT_RANDOM_SEED Default random seed.
protected static java.util.logging.Logger	LOG Logger object.
protected Molecule	queryMol Query molecule (it is neither cloned nor modified).
protected long	randomSeed Random seed (0 means using the current time).
protected SearchMode	searchMode Search mode.
protected McsSearchOptions	searchOpts Search options.
protected Molecule	targetMol Target molecule (it is neither cloned nor modified).
protected long	timeLimit Time limit in milliseconds (-1 means disabled).

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	MaxCommonSubstructure(McsSearchOptions searchOpts)

Method Summary

Modifier and Type	Method and Description
float	calculateSimilarityUpperBound() Calculates an upper bound on the similarity of the query and target molecules with respect to the specified search options.
int	calculateUpperBound() Calculates an upper bound on the number of bonds the maximum common substructure may contain with respect to the specified search options.
McsSearchResult	find() Performs MCS search according to the specified options.
protected abstract McsSearchResult	findMcs(com.chemaxon.search.sss.Matcher matcher) Finds the MCS and all related data (fragments and mappings).
Molecule	getQuery() Gets the query structure.
long	getRandomSeed() Gets the random seed value.
SearchMode	getSearchMode() Gets the current search mode.
McsSearchOptions	getSearchOptions() Returns the search options used by this instance.
Molecule	getTarget() Gets the target structure.
long	getTimeLimit() Gets the maximum allowed MCS search time in milliseconds.
boolean	hasNextResult() Returns whether there are more results available.
boolean	isLicensed() Returns information about the licensing of the product.
static MaxCommonSubstructure	newInstance() Creates a new instance of MCS search algorithm using the default search options.
static MaxCommonSubstructure	newInstance(McsSearchOptions searchOpts) Creates a new instance of MCS search algorithm using the given search options.
McsSearchResult	nextResult() Finds the next MCS search result according to the specified options.
void	setLicenseEnvironment(java.lang.String env) Sets the license environment.
void	setMolecules(Molecule query, Molecule target) Sets the two molecular structures to be matched.
void	setQuery(Molecule query) Sets the query structure.
void	setRandomSeed(long seed) Sets the random seed value.
void	setSearchMode(SearchMode mode) Sets the search mode that controls the running time and the accuracy of the algorithm.
void	setTarget(Molecule target) Sets the target structure.
void	setTimeLimit(long maxMilliseconds) Sets the maximum allowed time for MCS search.

Methods inherited from class java.lang.Object

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

Field Detail

DEFAULT_RANDOM_SEED

public static final long DEFAULT_RANDOM_SEED

Default random seed.

See Also:

Constant Field Values

randomSeed

protected long randomSeed

Random seed (0 means using the current time).

searchMode

protected SearchMode searchMode

Search mode.

timeLimit

protected long timeLimit

Time limit in milliseconds (-1 means disabled).

queryMol

protected Molecule queryMol

Query molecule (it is neither cloned nor modified).

targetMol

protected Molecule targetMol

Target molecule (it is neither cloned nor modified).

searchOpts

protected final McsSearchOptions searchOpts

Search options.

LOG

protected static final java.util.logging.Logger LOG

Logger object.

Constructor Detail

MaxCommonSubstructure

protected MaxCommonSubstructure(McsSearchOptions searchOpts)

Method Detail

newInstance

public static MaxCommonSubstructure newInstance()

Creates a new instance of MCS search algorithm using the default search options.

MaxCliqueMcs implementation is used, which turned out to be the most efficient and robust according to our benchmark tests.

Returns:

a new instance of the default algorithm implementation

See Also:

newInstance(McsSearchOptions)

newInstance

public static MaxCommonSubstructure newInstance(McsSearchOptions searchOpts)

Creates a new instance of MCS search algorithm using the given search options.

MaxCliqueMcs implementation is used, which turned out to be the most efficient and robust according to our benchmark tests.

Parameters:

searchOpts - the search options (not null)

Returns:

a new instance of the default algorithm implementation

isLicensed

public final boolean isLicensed()

Returns information about the licensing of the product.

Specified by:

isLicensed in interface chemaxon.license.Licensable

Returns:

true if the product is correctly licensed

setLicenseEnvironment

public final void setLicenseEnvironment(java.lang.String env)

Sets the license environment.

Specified by:

setLicenseEnvironment in interface chemaxon.license.Licensable

setMolecules

public final void setMolecules(Molecule query,
                               Molecule target)

Sets the two molecular structures to be matched.

Parameters:

query - query molecule (not null)

target - target molecule (not null)

setQuery

public final void setQuery(Molecule query)

Sets the query structure.

Parameters:

query - query molecule (not null)

setTarget

public final void setTarget(Molecule target)

Sets the target structure.

Parameters:

target - target molecule (not null)

getQuery

public final Molecule getQuery()

Gets the query structure.

Returns:

the query molecule

getTarget

public final Molecule getTarget()

Gets the target structure.

Returns:

the target molecule

getSearchOptions

public final McsSearchOptions getSearchOptions()

Returns the search options used by this instance. For more information, see McsSearchOptions.

Returns:

search options (not null)

setSearchMode

public final void setSearchMode(SearchMode mode)

Sets the search mode that controls the running time and the accuracy of the algorithm. The default option is SearchMode.NORMAL. For more information, see SearchMode.

Parameters:

mode - search mode (not null)

getSearchMode

public final SearchMode getSearchMode()

Gets the current search mode. For more information, see SearchMode.

Returns:

current search mode (not null)

setTimeLimit

public final void setTimeLimit(long maxMilliseconds)

Sets the maximum allowed time for MCS search. If the given limit is exceeded, the search process terminates with the best result obtained so far. When searching for multiple results (without changing the input molecules), the time limit is considered only on the first call to nextResult() (or hasNextResult()).

This is an optional limit, which is set to 1 minute by default. You can use a negative parameter value to disable it.

If the search process seems to be too slow, consider using FAST search mode instead of decreasing the time limit (see setSearchMode(SearchMode)).

Parameters:

maxMilliseconds - maximum running time in milliseconds (negative value means disabled)

See Also:

#getTerminationCause()

getTimeLimit

public final long getTimeLimit()

Gets the maximum allowed MCS search time in milliseconds. If no such limit is specified, this method returns -1. For more information, see setTimeLimit().

Returns:

maximum running time in milliseconds or -1 if no such limit is specified

setRandomSeed

public final void setRandomSeed(long seed)

Sets the random seed value. 0 means using the current time.

Parameters:

seed - random seed

getRandomSeed

public final long getRandomSeed()

Gets the random seed value. 0 means using the current time.

Returns:

seed random seed

find

public final McsSearchResult find()

Performs MCS search according to the specified options. This method returns a valid result object even if the maximum common substructure is empty or the search time limit is reached.

If multiple MCS results are desired, use hasNextResult() and nextResult().

Returns:

the McsSearchResult object containing the found common substructure

Throws:

java.lang.IllegalStateException - if the input molecules are not set before the call to this method or the method is called more than once for the same input molecules

hasNextResult

public final boolean hasNextResult()

Returns whether there are more results available. See nextResult().

Returns:

true if nextResult() can be called to obtain a new McsSearchResult object

nextResult

public final McsSearchResult nextResult()

Finds the next MCS search result according to the specified options. The first call to this method always returns a result, even if the maximum common substructure is empty or the search time limit is reached.

If multiple search results are desired, this method can be called repeatedly. The results typically correspond to equivalent common substructures but with different mappings. (See McsSearchResult for more information.) When this method is called multiple times, hasNextResult() should be called first to avoid exception.

The search state is reset after a call to setQuery(Molecule), setTarget(Molecule) or setMolecules(Molecule, Molecule).

Returns:

the McsSearchResult object containing the found common substructure

Throws:

java.lang.IllegalStateException - if the input molecules are not set before the first call to this method or there are no more results to return

calculateUpperBound

public final int calculateUpperBound()

Calculates an upper bound on the number of bonds the maximum common substructure may contain with respect to the specified search options. The returned value is an upper bound on McsSearchResult.getBondCount().

As this method is much faster than executing the MCS algorithm (find()), it can be used for pre-filtering.

Returns:

an upper bound on the size (bond count) of the MCS

Throws:

java.lang.IllegalStateException - if the input molecules are not set

calculateSimilarityUpperBound

public final float calculateSimilarityUpperBound()

Calculates an upper bound on the similarity of the query and target molecules with respect to the specified search options. The returned value is an upper bound on McsSearchResult.getSimilarity().

As this method is much faster than executing the MCS algorithm (find()), it can be used for pre-filtering.

Returns:

an upper bound on the (MCS based) similarity of the query and the target

Throws:

java.lang.IllegalStateException - if the input molecules are not set

findMcs

protected abstract McsSearchResult findMcs(com.chemaxon.search.sss.Matcher matcher)

Finds the MCS and all related data (fragments and mappings). Deriving classes must implement this method to calculate an (approximate) maximum common substructure.

This method is called from find(), nextResult(), and hasNextResult() with a properly initialized matcher object.

Parameters:

matcher - the matcher object, which also contains the preprocessed query and target molecules