Update coordinates code

PGB-LIV · Feb 5, 2016 · 9cea8b2 · 9cea8b2
1 parent c5d8b90
commit 9cea8b2
Show file tree

Hide file tree

Showing 6 changed files with 56 additions and 14 deletions.
diff --git a/README.txt b/README.txt
@@ -1,8 +1,10 @@
-
+
 
 MzidLib - A library of routines for manipulating data in the PSI's mzIdentML standard. 
 ProteoAnnotator is a completely automated software pipeline for integration of mass spectrometry (MS) based proteomic evidence into genome databases. 
-
+
+
+To download MzidLib and ProteoAnnotator: http://www.proteoannotator.org/?q=installation
 
 
 Citation:

diff --git a/documentation/~$oteoAnnotator_user_guide.docx b/documentation/~$oteoAnnotator_user_guide.docx
diff --git a/nbactions.xml b/nbactions.xml
@@ -10,7 +10,7 @@
                 <goal>org.codehaus.mojo:exec-maven-plugin:1.2.1:exec</goal>
             </goals>
             <properties>
-                <exec.args>-Xmx10G -classpath %classpath uk.ac.liv.mzidlib.MzIdentMLLib ProteoAnnotator -inputGFF Homo_sapiens.GRCh38.77.gff3 -inputFasta Homo_sapiens.GRCh38.pep.all.fa -spectrum_files /home/fawaz/pilot/green/PXD000561/Adult_Adrenalgland_Gel_Velos_2_f13.mgf -outputFolder /home/fawaz/pilot/green/PXD000561/withoutmsgf -prefix PXD000561_withoutmsgf_ -compress false -searchParameters green_nomods.txt -peptideThreshValue 0.01 -proteinThreshValue 0.01 -inputPredicted "Homo_sapiens.GRCh38.77.abinitio.gff3;Homo_sapiens.GRCh38.pep.abinitio.fa" -enableMsgf true</exec.args>
+                <exec.args>-Xmx10G -classpath %classpath uk.ac.liv.mzidlib.MzIdentMLLib AddGenomeCoordinatesForPeptides E:\tobias\PXD000764\34937\PXD000764_34937_combined_fdr_peptide_threshold.mzid E:\tobias\fawaz.mzid -inputGff E:\tobias\Homo_sapiens.GRCh38.77.gff3 -outputGff E:\tobias\fawaz.gff3</exec.args>
                 <exec.executable>java</exec.executable>
                 <skipTests>true</skipTests>
             </properties>
@@ -25,7 +25,7 @@
                 <goal>org.codehaus.mojo:exec-maven-plugin:1.2.1:exec</goal>
             </goals>
             <properties>
-                <exec.args>-Xdebug -Xrunjdwp:transport=dt_socket,server=n,address=${jpda.address} -Xmx10G -classpath %classpath uk.ac.liv.mzidlib.MzIdentMLLib ProteoAnnotator -inputGFF Homo_sapiens.GRCh38.77.gff3 -inputFasta Homo_sapiens.GRCh38.pep.all.fa -spectrum_files /home/fawaz/pilot/green/PXD000561/Adult_Adrenalgland_Gel_Velos_2_f13.mgf -outputFolder /home/fawaz/pilot/green/PXD000561/withoutmsgf -prefix PXD000561_withoutmsgf_ -compress false -searchParameters green_nomods.txt -peptideThreshValue 0.01 -proteinThreshValue 0.01 -inputPredicted "Homo_sapiens.GRCh38.77.abinitio.gff3;Homo_sapiens.GRCh38.pep.abinitio.fa" -enableMsgf true</exec.args>
+                <exec.args>-Xdebug -Xrunjdwp:transport=dt_socket,server=n,address=${jpda.address} -Xmx10G -classpath %classpath uk.ac.liv.mzidlib.MzIdentMLLib AddGenomeCoordinatesForPeptides E:\tobias\PXD000764\34937\PXD000764_34937_combined_fdr_peptide_threshold.mzid E:\tobias\fawaz.mzid -inputGff E:\tobias\Homo_sapiens.GRCh38.77.gff3 -outputGff E:\tobias\fawaz.gff3</exec.args>
                 <exec.executable>java</exec.executable>
                 <jpda.listen>true</jpda.listen>
             </properties>
@@ -40,7 +40,7 @@
                 <goal>org.codehaus.mojo:exec-maven-plugin:1.2.1:exec</goal>
             </goals>
             <properties>
-                <exec.args>-Xmx10G -classpath %classpath uk.ac.liv.mzidlib.MzIdentMLLib ProteoAnnotator -inputGFF Homo_sapiens.GRCh38.77.gff3 -inputFasta Homo_sapiens.GRCh38.pep.all.fa -spectrum_files /home/fawaz/pilot/green/PXD000561/Adult_Adrenalgland_Gel_Velos_2_f13.mgf -outputFolder /home/fawaz/pilot/green/PXD000561/withoutmsgf -prefix PXD000561_withoutmsgf_ -compress false -searchParameters green_nomods.txt -peptideThreshValue 0.01 -proteinThreshValue 0.01 -inputPredicted "Homo_sapiens.GRCh38.77.abinitio.gff3;Homo_sapiens.GRCh38.pep.abinitio.fa" -enableMsgf true</exec.args>
+                <exec.args>-Xmx10G -classpath %classpath uk.ac.liv.mzidlib.MzIdentMLLib AddGenomeCoordinatesForPeptides E:\tobias\PXD000764\34937\PXD000764_34937_combined_fdr_peptide_threshold.mzid E:\tobias\fawaz.mzid -inputGff E:\tobias\Homo_sapiens.GRCh38.77.gff3 -outputGff E:\tobias\fawaz.gff3</exec.args>
                 <exec.executable>java</exec.executable>
             </properties>
         </action>

diff --git a/pom.xml b/pom.xml
@@ -263,7 +263,20 @@
                     </execution>
                 </executions>
             </plugin>
-
+          <!--  TO BE ADDED: Fawaz<plugin>
+                <groupId>org.apache.maven.plugins</groupId>
+                <artifactId>maven-javadoc-plugin</artifactId>
+                <version>2.9.1</version>
+                <executions>
+                    <execution>
+                        <id>attach-javadocs</id>
+                        <goals>
+                            <goal>jar</goal>
+                        </goals>
+                    </execution>
+                </executions>
+            </plugin>
+-->
             <!-- Assembly plugin -->
             <plugin>
                 <groupId>org.apache.maven.plugins</groupId>

diff --git a/src/main/java/bgi/ipeak/IPeak.java b/src/main/java/bgi/ipeak/IPeak.java
@@ -127,10 +127,7 @@ public class IPeak {
     private File[] xtandem_list;
     private File[] omssa_list;
 
-    /**
-     * @param args
-     * @throws Exception
-     */
+
     public static void main(String[] args) throws Exception {
         IPeak ip = new IPeak();
         CmdLineParser parser = new CmdLineParser(ip);

diff --git a/src/main/java/uk/ac/liv/mzidlib/fasta/AddGenomeCoordinatesForPeptides.java b/src/main/java/uk/ac/liv/mzidlib/fasta/AddGenomeCoordinatesForPeptides.java
@@ -360,6 +360,9 @@ public void writeMappingResults() {
                         List<CDS_Information> outputCDS = new ArrayList();
                         int countCDS = 0;
                         boolean sorted = false;
+                        if (gffData.get(0).getStrand().equals("-")){
+                            sortedCDS = sortCDSAccordingToStartPositionTemp(sortedCDS);
+                        }
                         for (int j = 0; j < sortedCDS.size(); j++) {
                             CDS_Information object = sortedCDS.get(j);
                             long s = object.getStart();
@@ -406,8 +409,14 @@ public void writeMappingResults() {
                         //System.out.println("END");
                         for (int j = 0; j < outputCDS.size(); j++) {
                             CDS_Information cDS_Information = outputCDS.get(j);
-                            peptideEvidence.getUserParam().add(makeUserParam("start_map", String.valueOf(cDS_Information.getStart())));
-                            peptideEvidence.getUserParam().add(makeUserParam("end_map", String.valueOf(cDS_Information.getEnd())));
+                            // Added by Fawaz based on Tobias email 29/01/2016
+                            if (j==0){
+                                peptideEvidence.getUserParam().add(makeUserParam("start_map", String.valueOf(cDS_Information.getStart())));
+                                peptideEvidence.getUserParam().add(makeUserParam("end_map", String.valueOf(cDS_Information.getEnd()+3)));
+                            }else{
+                                peptideEvidence.getUserParam().add(makeUserParam("start_map", String.valueOf(cDS_Information.getStart())));
+                                peptideEvidence.getUserParam().add(makeUserParam("end_map", String.valueOf(cDS_Information.getEnd()+1)));
+                            }
 
                         }
 
@@ -522,10 +531,10 @@ long[] determineTheLocationOfSeqOnCds(ProteinResults pr, List<CDS_Information> c
         // Get locations from the protein object
         long start = pr.getstart();
         long end = pr.getEnd();
-
+        
         // sort the CDS collection according to the strand
         List<CDS_Information> sortedCDS = sortCDSAccordingToStartPosition(cdsColl);
-
+       
         long mapped_start = getMappedCordinates(start, sortedCDS, pr);
         long mapped_end = getMappedCordinates(end, sortedCDS, pr);
         if (mapped_start == -1) {
@@ -623,6 +632,27 @@ List<CDS_Information> sortCDSAccordingToStartPosition(List<CDS_Information> cdsC
         return cdsColl;
 
     }
+
+    List<CDS_Information> sortCDSAccordingToStartPositionTemp(List<CDS_Information> cdsCollection) {
+
+        List<CDS_Information> cdsColl = new ArrayList<CDS_Information>(cdsCollection);
+
+
+
+
+            for (int i = cdsColl.size() - 1; i >= 0; i--) {
+                int j = cdsColl.size() - 1 - i;
+                if (j < i) {
+                    CDS_Information temp = cdsColl.get(i);
+                    cdsColl.set(i, cdsColl.get(j));
+                    cdsColl.set(j, temp);
+                }
+            }
+
+
+        return cdsColl;
+
+    }
 
     /**
      *