package jalview.ws.jws2; import jalview.api.AlignCalcWorkerI; import jalview.datamodel.AlignmentAnnotation; import jalview.datamodel.Annotation; import jalview.gui.AlignFrame; import jalview.ws.jws2.jabaws2.Jws2Instance; import jalview.ws.params.WsParamSetI; import java.text.MessageFormat; import java.util.ArrayList; import java.util.Arrays; import java.util.HashMap; import java.util.List; import java.util.Map; import java.util.LinkedHashMap; import java.util.TreeSet; import java.util.regex.Pattern; import compbio.data.sequence.RNAStructReader.AlifoldResult; import compbio.data.sequence.RNAStructScoreManager; import compbio.data.sequence.Range; import compbio.data.sequence.Score; import compbio.metadata.Argument; public class RNAalifoldClient extends JabawsAlignCalcWorker implements AlignCalcWorkerI { String methodName; AlignFrame af; // keeps track of whether the RNAalifold result includes base contact probabilities boolean bpScores; public RNAalifoldClient(Jws2Instance sh, AlignFrame alignFrame, WsParamSetI preset, List paramset) { super(sh, alignFrame, preset, paramset); if (arguments == null) arguments = new ArrayList(); af = alignFrame; methodName = sh.serviceType; nucleotidesAllowed = true; proteinAllowed = false; } @Override public String getServiceActionText() { return "Submitting RNA alignment for Secondary Structure prediction using " + "RNAalifold Service"; } @Override public void updateResultAnnotation(boolean immediate) { if (immediate || !calcMan.isWorking(this) && scoremanager != null) { List ourAnnot = new ArrayList(); // Unpack the ScoreManager List structs = ((RNAStructScoreManager) scoremanager).getStructs(); List> data = ((RNAStructScoreManager) scoremanager).getData(); System.out.println("Length of RNAStructScoreManager is: " + structs.size()); // test to see if this data object contains base pair contacts Score fscore = data.get(0).first(); this.bpScores = (fscore.getMethod().equals( AlifoldResult.contactProbabilities.toString())); // add annotation for the consensus sequence alignment createAnnotationRowforScoreHolder(ourAnnot, getCalcId(), structs.get(0), null, null); // Add annotations for the mfe Structure createAnnotationRowforScoreHolder(ourAnnot, getCalcId(), structs.get(1), data.get(1), null); // decide whether to add base pair contact probability histogram int count = 2; if (bpScores) { createAnnotationRowforScoreHolder(ourAnnot, getCalcId(), structs.get(2), data.get(0), data.get(2)); count++; } // Now loop for the rest of the Annotations (if there it isn't stochastic output // only the centroid and MEA structures remain anyway) for (int i = count; i < structs.size(); i++) { // The ensemble values should be displayed in the description of the // first (or all?) Stochastic Backtrack Structures. if (!data.get(i).first().getMethod().equals( AlifoldResult.ensembleValues.toString())) { createAnnotationRowforScoreHolder(ourAnnot, getCalcId(), structs.get(i), data.get(i), null); } } if (ourAnnot.size() > 0) { updateOurAnnots(ourAnnot); ap.adjustAnnotationHeight(); } } } protected void createAnnotationRowforScoreHolder( List ourAnnot, String calcId, String struct, TreeSet data, TreeSet descriptionData) { /* If contactProbability information is returned from RNAalifold it is stored * in the first TreeSet object corresponding to the String Id which * holds the consensus alignment. The method enumeration is then updated to * AlifoldResult.contactProbabilties. This line recreates the same * data object as was overwritten with the contact probabilites data. */ if (data == null) data = compbio.data.sequence.RNAStructReader .newEmptyScore(AlifoldResult.consensusAlignment); if (descriptionData == null) descriptionData = data; String[] typenameAndDescription = constructTypenameAndDescription( descriptionData.first()); String typename = typenameAndDescription[0]; String description = typenameAndDescription[1]; AlignmentAnnotation annotation = alignViewport.getAlignment() .findOrCreateAnnotation(typename, calcId, false, null, null); constructAnnotationFromScoreHolder(annotation, struct, data); /* update annotation description with the free Energy, frequency in ensemble * or other data where appropriate. * * Doesnt deal with AlifoldResult.ensembleValues, the free energy of ensemble * and frequency of mfe structure in ensemble. How to deal with these? */ annotation.description = description; annotation.belowAlignment = false; // annotation.showAllColLabels = true; annotation.validateRangeAndDisplay(); ourAnnot.add(annotation); } private AlignmentAnnotation constructAnnotationFromScoreHolder( AlignmentAnnotation annotation, String struct, TreeSet data) { Annotation[] anns = new Annotation[struct.length()]; if (data != null && data.size() > 1 && data.first().getMethod().equals( AlifoldResult.contactProbabilities.toString())) { // The base pair probabilities are stored in a set in scoreholder. we want a map LinkedHashMap basePairs = new LinkedHashMap(); for (Score score : data) { // The Score objects contain a set of size one containing the range and // an ArrayList of size one containing the probabilty basePairs.put(score.getRanges().first(), new Float(score.getScores().get(0))); } for (int i = 0; i < struct.length(); i++) { // Return all the contacts associated with position i LinkedHashMap contacts = isContact(basePairs, i+1); String description = ""; float prob = 0f; if (contacts.size() == 0) { description = "No Data"; } else { for (Range contact : contacts.keySet()) { float t = contacts.get(contact); if (t > prob) prob = t; description += Integer.toString(contact.from) + "->" + Integer.toString(contact.to) + ": " + Float.toString(t) + "% | "; } } anns[i] = new Annotation(struct.substring(i, i+1), description, isSS(struct.charAt(i)), prob); } } else if (data == null || data.size() == 1) { for (int i = 0; i < struct.length(); i++) { anns[i] = new Annotation(struct.substring(i, i+1), "", isSS(struct.charAt(i)), Float.NaN); } annotation.graph = 0; // No graph } annotation.annotations = anns; return annotation; } private String[] constructTypenameAndDescription(Score score) { String description = ""; String typename = ""; String datatype = score.getMethod(); if (datatype.equals(AlifoldResult.mfeStructure.toString())) { description = MessageFormat.format("Minimum Free Energy Structure. Energy: {0} = {1} + {2}", score.getScores().get(0), score.getScores().get(1), score.getScores().get(2)); typename = "MFE Structure"; } else if (datatype.equals(AlifoldResult.contactProbabilityStructure.toString())) { description = MessageFormat.format("Base Pair Contact Probabilities. " + "Energy of Ensemble: {0} Frequency of Ensemble: {1}", score.getScores().get(0), score.getScores().get(1)); typename = "Contact Probabilities"; } else if (datatype.equals(AlifoldResult.centroidStructure.toString())) { description = MessageFormat.format("Centroid Structure. Energy: {0} = {1} + {2}", score.getScores().get(0), score.getScores().get(1), score.getScores().get(2)); typename = "Centroid Structure"; } else if (datatype.equals(AlifoldResult.stochBTStructure.toString())) { if (score.getScores().size() > 0) { description = MessageFormat.format("Probability: {0} Energy: {1}", score.getScores().get(0), score.getScores().get(1)); } else description = "Stochastic Backtrack Structure"; } else if (datatype.equals(AlifoldResult.MEAStucture.toString())) { description = MessageFormat.format("Maximum Expected Accuracy Values: '{' {0} MEA={1} '}", score.getScores().get(0), score.getScores().get(1)); typename = "MEA Structure"; } else if (datatype.equals(AlifoldResult.consensusAlignment.toString())) { typename = "RNAalifold Consensus"; description = "Consensus Alignment Produced by RNAalifold"; } else { typename = datatype; description = typename; } return new String[] {typename, description}; } // Check whether, at position i there is a base contact and return all the // contacts at this position. Should be in order of descending probability. private LinkedHashMap isContact(LinkedHashMap basePairs, int i) { LinkedHashMap contacts = new LinkedHashMap(); for (Range contact : basePairs.keySet()) { // finds the contacts associtated with position i ordered by the natural // ordering of the Scores TreeSet in ScoreManager which is, descending probability if (contact.from == i || contact.to == i) contacts.put(contact, basePairs.get(contact)); } return contacts; } private char isSS (char chr) { String regex = "\\(|\\)|\\{|\\}|\\[|\\]"; char ss = (Pattern.matches(regex, Character.toString(chr))) ? 'S': ' '; return ss; } public String getCalcId() { return SequenceAnnotationWSClient.AAConCalcId; } }