Package de.julielab.genemapper.filtering

Class UnspecificNameFilter

java.lang.Object

de.julielab.genemapper.filtering.UnspecificNameFilter

public class UnspecificNameFilter
extends Object

Also known as the CrazyRegExFilter. Removes gene names that are too unspecific ("DNA binding protein") and those that most likely refer to a gene family, process, cell, disease, experimental technique (GST, LPS), etc.

Author:

Joerg

Field Summary

Fields

Modifier and Type	Field	Description
static String	blacklist_pattern
static String	leftWordBoundary	characters that appear on the left side of a proper word: white space, bracket starts, ..
static String	rightWordBoundary	characters that appear on the right side of a proper word: white space, comma, bracket ends, ..

Constructor Summary

Constructors

Constructor	Description
UnspecificNameFilter()

Method Summary

Modifier and Type	Method	Description
void	filter(de.julielab.geneexpbase.genemodel.GeneDocument document)	Removes all recognized gene names that are made up entirely of unspecific keywords or where names point to something else than a gene name, e.g.
static boolean	isAminoAcid(String name)	Checks if the given term matches an amino acid, using three-letter code (Ala) and full names (Alanine, alanines).
static boolean	isCellLine(String name, String sentence)	Returns true if the name is followed by a keyword indicating that this name is a cell line.
static boolean	isChromosome(String name, String sentence)	Checks if the given name found in a sentence refers to a (human!) chromosome: it should have at least a chromosome number and arm (13p), maybe also bands (13p.23), and the name should also appear next to the word 'chromosome'
static boolean	isDiseaseName(String name)	Returns true if the name represents a disease name.
static boolean	isNegativePair(String name, String sentence)	Some gene names refer to the actual gene only in very few cases.
static boolean	isSpecies(String name)	Naive implementation: checks if the name refers to a species.
static boolean	isTissueCellCompartment(String name)
static boolean	isUnspecific(String term)
static boolean	isUnspecific(String term, HashSet<Integer> speciesIDs)	Decides whether a proposed gene name is too unspecific: mainly checks for compound names that do not have any reference concering the exact identity of a protein.
static boolean	isUnspecificAbbreviation(String name, String sentence)
static boolean	isUnspecificAbbreviation(String name, String sentence, HashSet<Integer> speciesIDs)
static boolean	isUnspecificSingleWord(String name)	Checks a gene name against a list of single word, case sensitive words that are pretty much always false positives: "aim", "fat", "up".
static boolean	isUnspecificSingleWordCaseInsensitive(String name)
static boolean	keepDiseaseName(String name, String sentence)	Returns true if this disease name is mentioned together with a locus or genes/proteins in this sentence.
static boolean	textHasPlural(String name, String text)

Methods inherited from class java.lang.Object

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

Field Detail

leftWordBoundary

public static String leftWordBoundary

characters that appear on the left side of a proper word: white space, bracket starts, ..

rightWordBoundary

public static String rightWordBoundary

characters that appear on the right side of a proper word: white space, comma, bracket ends, ..

blacklist_pattern

public static String blacklist_pattern

Constructor Detail

UnspecificNameFilter

public UnspecificNameFilter()

Method Detail

isUnspecific

public static boolean isUnspecific(String term)

Parameters:

term -

Returns:

isUnspecific

public static boolean isUnspecific(String term,
                                   HashSet<Integer> speciesIDs)

Decides whether a proposed gene name is too unspecific: mainly checks for compound names that do not have any reference concering the exact identity of a protein. Also removes units and diseases/syndromes, single letter names
Examples: protease, DNA binding protein, nerve growth factor, human polymerase, cell-surface glycoprotein.

Parameters:

term -

Returns:

isUnspecificSingleWord

public static boolean isUnspecificSingleWord(String name)

Checks a gene name against a list of single word, case sensitive words that are pretty much always false positives: "aim", "fat", "up".

Parameters:

name -

Returns:

isUnspecificSingleWordCaseInsensitive

public static boolean isUnspecificSingleWordCaseInsensitive(String name)

Parameters:

name -

Returns:

isUnspecificAbbreviation

public static boolean isUnspecificAbbreviation(String name,
                                               String sentence)

Parameters:

name -

sentence -

Returns:

isUnspecificAbbreviation

public static boolean isUnspecificAbbreviation(String name,
                                               String sentence,
                                               HashSet<Integer> speciesIDs)

Returns:

isAminoAcid

public static boolean isAminoAcid(String name)

Checks if the given term matches an amino acid, using three-letter code (Ala) and full names (Alanine, alanines). Does not check against one-letter codes (A).

Parameters:

name -

Returns:

isDiseaseName

public static boolean isDiseaseName(String name)

Returns true if the name represents a disease name.

Parameters:

name -

Returns:

keepDiseaseName

public static boolean keepDiseaseName(String name,
                                      String sentence)

Returns true if this disease name is mentioned together with a locus or genes/proteins in this sentence. In these cases, we prefer to keep the name of the gene.

Parameters:

name -

sentence -

Returns:

isTissueCellCompartment

public static boolean isTissueCellCompartment(String name)

Parameters:

name -

Returns:

isCellLine

public static boolean isCellLine(String name,
                                 String sentence)

Returns true if the name is followed by a keyword indicating that this name is a cell line.

Parameters:

name -

sentence -

Returns:

isSpecies

public static boolean isSpecies(String name)

Naive implementation: checks if the name refers to a species.

Parameters:

name -

Returns:

true if the name refers to a species

isChromosome

public static boolean isChromosome(String name,
                                   String sentence)

Checks if the given name found in a sentence refers to a (human!) chromosome:
it should have at least a chromosome number and arm (13p), maybe also bands (13p.23), and the name should also appear next to the word 'chromosome'

Parameters:

name -

sentence -

Returns:

true if the given name refers to a chromosome

isNegativePair

public static boolean isNegativePair(String name,
                                     String sentence)

Some gene names refer to the actual gene only in very few cases. Examples are
- GST (glutathione-S-transferase, an experimental technique for pulldowns),
- LPS (lipopolysaccharide in most cases, not IRF6),
- polymerase (... chain reaction), ...
which need to be filtered out.

Parameters:

name -

sentence -

Returns:

textHasPlural

public static boolean textHasPlural(String name,
                                    String text)

Parameters:

name -

Returns:

filter

public void filter(de.julielab.geneexpbase.genemodel.GeneDocument document)

Removes all recognized gene names that are made up entirely of unspecific keywords or where names point to something else than a gene name, e.g. a tissue, species, amino acid.