AbundanceColumn.java
/*
* Copyright 2018 Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V..
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package uk.ac.ebi.pride.jmztab2.model;
import de.isas.mztab2.model.Assay;
import de.isas.mztab2.model.IndexedElement;
import de.isas.mztab2.model.StudyVariable;
import java.util.SortedMap;
import java.util.TreeMap;
/**
* If the data exporter wishes to report only final results for 'Summary' files (i.e. following averaging over replicates),
* then these MUST be reported as quantitative values in the columns associated with the study_variable[1-n] (e.g.
* abundance_study_variable[1]). mzTab allows the reporting of abundance, standard deviation, and standard error
* for any study_variable. The unit of values in the abundance column MUST be specified in the metadata section of the mzTab file.
* The reported values SHOULD represent the final result of the performed data analysis. The exact meaning of the values will
* thus depend on the used analysis pipeline and quantitation method and is not expected to be comparable across multiple mzTab files.
*
* @author qingwei
* @author nilshoffmann
* @since 23/05/13
*
*/
public class AbundanceColumn extends MZTabColumn {
public enum Field {
ABUNDANCE_ASSAY ("abundance_assay", Double.class, 1),
ABUNDANCE_STUDY_VARIABLE ("abundance_study_variable", Double.class, 2),
ABUNDANCE_VARIATION_STUDY_VARIABLE ("abundance_variation_study_variable", Double.class, 3);
private String name;
private Class columnType;
private int position;
Field(String name, Class columnType, int position) {
this.name = name;
this.columnType = columnType;
this.position = position;
}
@Override
public String toString() {
return name;
}
}
/**
* Generate a abundance column:
* The column header is: {Section}_{Field#name()}_{IndexedElement[id]}
* The column data type: {Field#columnType()}
* The column position: always most right side, calculated by offset.
*/
private AbundanceColumn(Section section, Field field, Object element, int offset) {
super(field.name, field.columnType, true, offset + field.position + "");
setElement(element);
}
/**
* Generate a abundance optional column as measured in the given assay.The column header like
* protein_abundance_assay[1-n], the position always stay the most right of the tabled section,
* and the data type is Double.
*
* @param section SHOULD be {@link uk.ac.ebi.pride.jmztab2.model.Section#Protein}, {@link uk.ac.ebi.pride.jmztab2.model.Section#Peptide}, {@link uk.ac.ebi.pride.jmztab2.model.Section#PSM},
* or {@link uk.ac.ebi.pride.jmztab2.model.Section#Small_Molecule}.
* @param assay SHOULD not be null.
* @param offset Normally the last column's position in header, {@link uk.ac.ebi.pride.jmztab2.model.MZTabColumnFactory#getColumnMapping()},
* @return an abundance optional column as measured in the given assay.
*/
public static MZTabColumn createOptionalColumn(Section section, Assay assay, int offset) {
if (section.isComment() || section.isMetadata()) {
throw new IllegalArgumentException("Section should be Protein, Peptide, PSM or SmallMolecule.");
}
if (assay == null) {
throw new NullPointerException("Assay should not be null!");
}
return new AbundanceColumn(Section.toDataSection(section), Field.ABUNDANCE_ASSAY, assay, offset);
}
/**
* Generate an abundance optional column as measured in the given study variable.
* The header can be one of abundance_study_variable[1], abundance_coeffvar_study_variable[1].
* The position always stay the most right of the tabled section, and the data type is Double.
*
* @param section SHOULD be {@link uk.ac.ebi.pride.jmztab2.model.Section#Protein}, {@link uk.ac.ebi.pride.jmztab2.model.Section#Peptide}, {@link uk.ac.ebi.pride.jmztab2.model.Section#PSM},
* or {@link uk.ac.ebi.pride.jmztab2.model.Section#Small_Molecule}.
* @param studyVariable SHOULD not be null.
* @param order position in header for the new columns {@link uk.ac.ebi.pride.jmztab2.model.MZTabColumnFactory#getColumnMapping()}
* @return an abundance optional column as measured in the given study variable.
* @param columnHeader a processed column header with abundance_ removed.
*
*/
public static SortedMap<String, MZTabColumn> createOptionalColumns(Section section, StudyVariable studyVariable, String columnHeader, String order) {
if (section.isComment() || section.isMetadata()) {
throw new IllegalArgumentException("Section should be Protein, Peptide, PSM, SmallMolecule, SmallMoleculeFeature or SmallMoleculeEvidence.");
}
if (studyVariable == null) {
throw new NullPointerException("Study Variable should not be null!");
}
//In this case we know the real position in which the column need to star, so the offset is one less
int offset = new Integer(order)-1;
SortedMap<String, MZTabColumn> columns = new TreeMap<>();
Section dataSection = Section.toDataSection(section);
AbundanceColumn column;
if(columnHeader.startsWith("study_variable")) {
column = new AbundanceColumn(dataSection, Field.ABUNDANCE_STUDY_VARIABLE, studyVariable, offset);
columns.put(column.getLogicPosition(), column);
} else if (columnHeader.startsWith("variation_study_variable")) {
column = new AbundanceColumn(dataSection, Field.ABUNDANCE_VARIATION_STUDY_VARIABLE, studyVariable, offset);
columns.put(column.getLogicPosition(), column);
} else {
throw new IllegalArgumentException("column header "+columnHeader+" is not allowed for abundance definition!");
}
return columns;
}
/**
* <p>createOptionalColumns.</p>
*
* @param section a {@link uk.ac.ebi.pride.jmztab2.model.Section} object.
* @param studyVariable a {@link de.isas.mztab2.model.StudyVariable} object.
* @param lastOrder a {@link java.lang.Integer} object.
* @return a {@link java.util.SortedMap} object.
*/
public static SortedMap<String, MZTabColumn> createOptionalColumns(Section section, StudyVariable studyVariable, Integer lastOrder) {
if (section.isComment() || section.isMetadata()) {
throw new IllegalArgumentException("Section should be Protein, Peptide, PSM or SmallMolecule.");
}
if (studyVariable == null) {
throw new NullPointerException("Study Variable should not be null!");
}
//In this case we know the real position in which the column need to star, so the offset is one less
SortedMap<String, MZTabColumn> columns = new TreeMap<>();
Section dataSection = Section.toDataSection(section);
AbundanceColumn column;
column = new AbundanceColumn(dataSection, Field.ABUNDANCE_STUDY_VARIABLE, studyVariable, lastOrder);
columns.put(column.getLogicPosition(), column);
column = new AbundanceColumn(dataSection, Field.ABUNDANCE_VARIATION_STUDY_VARIABLE, studyVariable, lastOrder);
columns.put(column.getLogicPosition(), column);
return columns;
}
}