in progress

[jalview.git] / forester / java / src / org / forester / archaeopteryx / tools / PhylogeneticInferrer.java

diff --git a/forester/java/src/org/forester/archaeopteryx/tools/PhylogeneticInferrer.java b/forester/java/src/org/forester/archaeopteryx/tools/PhylogeneticInferrer.java
index f0ed725..8b89646 100644 (file)
--- a/forester/java/src/org/forester/archaeopteryx/tools/PhylogeneticInferrer.java
+++ b/forester/java/src/org/forester/archaeopteryx/tools/PhylogeneticInferrer.java
@@ -21,7 +21,7 @@
 // Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
 //
 // Contact: phylosoft @ gmail . com
-// WWW: www.phylosoft.org/forester
+// WWW: https://sites.google.com/site/cmzmasek/home/software/forester
 
 package org.forester.archaeopteryx.tools;
 
@@ -30,18 +30,15 @@ import java.io.FileWriter;
 import java.io.IOException;
 import java.util.ArrayList;
 import java.util.List;
-import java.util.regex.Matcher;
 
 import javax.swing.JOptionPane;
 
 import org.forester.archaeopteryx.MainFrameApplication;
-import org.forester.evoinference.distance.NeighborJoining;
+import org.forester.evoinference.distance.NeighborJoiningF;
 import org.forester.evoinference.distance.PairwiseDistanceCalculator;
 import org.forester.evoinference.matrix.distance.BasicSymmetricalDistanceMatrix;
 import org.forester.evoinference.tools.BootstrapResampler;
-import org.forester.io.parsers.FastaParser;
 import org.forester.msa.BasicMsa;
-import org.forester.msa.ClustalOmega;
 import org.forester.msa.Mafft;
 import org.forester.msa.Msa;
 import org.forester.msa.Msa.MSA_FORMAT;
@@ -49,9 +46,7 @@ import org.forester.msa.MsaInferrer;
 import org.forester.msa.MsaMethods;
 import org.forester.msa.ResampleableMsa;
 import org.forester.phylogeny.Phylogeny;
-import org.forester.phylogeny.PhylogenyNode;
-import org.forester.phylogeny.data.Accession;
-import org.forester.phylogeny.iterators.PhylogenyNodeIterator;
+import org.forester.phylogeny.PhylogenyMethods;
 import org.forester.sequence.Sequence;
 import org.forester.tools.ConfidenceAssessor;
 import org.forester.util.ForesterUtil;
@@ -98,8 +93,6 @@ public class PhylogeneticInferrer extends RunnableProcess {
         switch ( msa_prg ) {
             case MAFFT:
                 return runMAFFT( _seqs, processMafftOptions() );
-            case CLUSTAL_O:
-                return runClustalOmega( _seqs, processMafftOptions() );
             default:
                 return null;
         }
@@ -149,9 +142,10 @@ public class PhylogeneticInferrer extends RunnableProcess {
                 e.printStackTrace();
             }
         }
-        final NeighborJoining nj = NeighborJoining.createInstance();
+        final NeighborJoiningF nj = NeighborJoiningF.createInstance( false, 5 );
         final Phylogeny phy = nj.execute( m );
-        PhylogeneticInferrer.extractFastaInformation( phy );
+        PhylogenyMethods.addMolecularSeqsToTree( phy, msa );
+        PhylogenyMethods.extractFastaInformation( phy );
         return phy;
     }
 
@@ -210,7 +204,7 @@ public class PhylogeneticInferrer extends RunnableProcess {
             }
             final MsaMethods msa_tools = MsaMethods.createInstance();
             if ( _options.isExecuteMsaProcessing() ) {
-                msa = msa_tools.removeGapColumns( _options.getMsaProcessingMaxAllowedGapRatio(),
+                msa = msa_tools.deleteGapColumns( _options.getMsaProcessingMaxAllowedGapRatio(),
                                                   _options.getMsaProcessingMinAllowedLength(),
                                                   msa );
                 if ( msa == null ) {
@@ -279,20 +273,6 @@ public class PhylogeneticInferrer extends RunnableProcess {
         return msa;
     }
 
-    private Msa runClustalOmega( final List<Sequence> seqs, final List<String> opts ) throws IOException,
-            InterruptedException {
-        Msa msa = null;
-        final MsaInferrer clustalo = ClustalOmega.createInstance( _mf.getInferenceManager().getPathToLocalClustalo()
-                .getCanonicalPath() );
-        try {
-            msa = clustalo.infer( seqs, opts );
-        }
-        catch ( final IOException e ) {
-            System.out.println( clustalo.getErrorDescription() );
-        }
-        return msa;
-    }
-
     private void writeToFiles( final BasicSymmetricalDistanceMatrix m ) {
         if ( !ForesterUtil.isEmpty( _options.getIntermediateFilesBase() ) ) {
             try {
@@ -311,39 +291,7 @@ public class PhylogeneticInferrer extends RunnableProcess {
         }
     }
 
-    public static void extractFastaInformation( final Phylogeny phy ) {
-        for( final PhylogenyNodeIterator iter = phy.iteratorExternalForward(); iter.hasNext(); ) {
-            final PhylogenyNode node = iter.next();
-            if ( !ForesterUtil.isEmpty( node.getName() ) ) {
-                final Matcher name_m = FastaParser.FASTA_DESC_LINE.matcher( node.getName() );
-                if ( name_m.lookingAt() ) {
-                    System.out.println();
-                    // System.out.println( name_m.group( 1 ) );
-                    // System.out.println( name_m.group( 2 ) );
-                    // System.out.println( name_m.group( 3 ) );
-                    // System.out.println( name_m.group( 4 ) );
-                    final String acc_source = name_m.group( 1 );
-                    final String acc = name_m.group( 2 );
-                    final String seq_name = name_m.group( 3 );
-                    final String tax_sn = name_m.group( 4 );
-                    if ( !ForesterUtil.isEmpty( acc_source ) && !ForesterUtil.isEmpty( acc ) ) {
-                        ForesterUtil.ensurePresenceOfSequence( node );
-                        node.getNodeData().getSequence( 0 ).setAccession( new Accession( acc, acc_source ) );
-                    }
-                    if ( !ForesterUtil.isEmpty( seq_name ) ) {
-                        ForesterUtil.ensurePresenceOfSequence( node );
-                        node.getNodeData().getSequence( 0 ).setName( seq_name );
-                    }
-                    if ( !ForesterUtil.isEmpty( tax_sn ) ) {
-                        ForesterUtil.ensurePresenceOfTaxonomy( node );
-                        node.getNodeData().getTaxonomy( 0 ).setScientificName( tax_sn );
-                    }
-                }
-            }
-        }
-    }
-
     public enum MSA_PRG {
-        MAFFT, CLUSTAL_O;
+        MAFFT;
     }
 }