001/* BioJava development code 002 * 003 * This code may be freely distributed and modified under the 004 * terms of the GNU Lesser General Public Licence. This should 005 * be distributed with the code. If you do not have a copy, 006 * see: 007 * 008 * http://www.gnu.org/copyleft/lesser.html 009 * 010 * Copyright for this code is held jointly by the individual 011 * authors. These should be listed in @author doc comments. 012 * 013 * For more information on the BioJava project and its aims, 014 * or to join the biojava-l mailing list, visit the home page 015 * at: 016 * 017 * http://www.biojava.org/ 018 * 019 */ 020package org.biojava.nbio.ronn; 021 022import org.biojava.nbio.core.sequence.ProteinSequence; 023import org.biojava.nbio.core.sequence.compound.AminoAcidCompound; 024import org.biojava.nbio.data.sequence.FastaSequence; 025import org.biojava.nbio.data.sequence.SequenceUtil; 026 027import java.io.FileInputStream; 028import java.io.FileNotFoundException; 029import java.io.IOException; 030import java.util.ArrayList; 031import java.util.List; 032import java.util.Map; 033import java.util.TreeMap; 034 035 036/** 037 * This class gives public API to RONN functions. 038 * It is build on top of the command line client. Due to this fact a few things 039 * could be improved and extended pending the refactoring of the command line client. 040 * 041 * The input sequence limitations - the input sequence must not contain any ambiguous characters, 042 * and have a minimum length of 19 amino acids. 043 * 044 * @author Peter Troshin 045 * @version 1.0 046 * @since 3.0.2 047 * 048 */ 049public class Jronn { 050 051 // Load models 052 private static final ModelLoader loader = new ModelLoader(); 053 static { 054 try { 055 loader.loadModels(); 056 } catch (NumberFormatException e) { 057 throw new RuntimeException("Fails to load models!" + e.getMessage(), e); 058 } catch (IOException e) { 059 throw new RuntimeException("Fails to load models!" + e.getMessage(), e); 060 } 061 } 062 063 064 /** 065 * Holder for the ranges, contain pointers to starting and ending position 066 * on the sequence which comprises a disordered region. Immutable. 067 * @author pvtroshin 068 */ 069 public static class Range { 070 /** 071 * Range starting position counts from 1 (the first position on the sequence is 1) 072 */ 073 public final int from; 074 /** 075 * The range ending position includes the last residue. 076 */ 077 public final int to; 078 079 public final float score; 080 public Range(int from, int to, float score) { 081 assert from>=0; 082 assert from<to; 083 this.from = from; 084 this.to = to; 085 this.score = score; 086 } 087 088 @Override 089 public String toString() { 090 return "Range" + " From:" + from + "\t" + "to: " + to + "\n"; 091 } 092 093 @Override 094 public int hashCode() { 095 final int prime = 31; 096 int result = 1; 097 result = prime * result + from; 098 result = prime * result + to; 099 return result; 100 } 101 102 @Override 103 public boolean equals(Object obj) { 104 if (this == obj) 105 return true; 106 if (obj == null) 107 return false; 108 if (getClass() != obj.getClass()) 109 return false; 110 Range other = (Range) obj; 111 if (from != other.from) 112 return false; 113 if (to != other.to) 114 return false; 115 return true; 116 } 117 118 119 } 120 121 /** 122 * Calculates the probability value for each residue in the protein sequence, 123 * telling the probability that the residue belongs to disordered region. 124 * In general, values greater than 0.5 considered to be in the disordered regions. 125 * 126 * @param sequence an instance of FastaSequence object, holding the name and the sequence. 127 * @return the probability scores for each residue in the sequence 128 */ 129 public static float[] getDisorderScores(FastaSequence sequence) { 130 return predictSerial(sequence); 131 } 132 133 /** 134 * Calculates the probability value for each residue in the protein sequence, 135 * telling the probability that the residue belongs to disordered region. 136 * In general, values greater than 0.5 considered to be in the disordered regions. 137 * 138 * @param sequence an instance of FastaSequence object, holding the name and the sequence. 139 * @return the probability scores for each residue in the sequence 140 */ 141 public static float[] getDisorderScores(ProteinSequence sequence) { 142 143 FastaSequence seq = convertProteinSequencetoFasta(sequence); 144 145 return predictSerial(seq); 146 } 147 148 /** Utility method to convert a BioJava ProteinSequence object to the FastaSequence 149 * object used internally in JRonn. 150 * 151 * @param sequence 152 * @return 153 */ 154 public static FastaSequence convertProteinSequencetoFasta(ProteinSequence sequence){ 155 StringBuffer buf = new StringBuffer(); 156 for (AminoAcidCompound compound : sequence) { 157 158 String c = compound.getShortName(); 159 160 if (! SequenceUtil.NON_AA.matcher(c).find()) { 161 buf.append(c); 162 } else { 163 buf.append("X"); 164 } 165 } 166 167 return new FastaSequence(sequence.getAccession().getID(),buf.toString()); 168 } 169 170 private static float[] predictSerial(FastaSequence fsequence) { 171 ORonn.validateSequenceForRonn(fsequence); 172 ORonn ronn; 173 float[] disorder = null; 174 try { 175 ronn = new ORonn(fsequence, loader); 176 disorder = ronn.call().getMeanScores(); 177 } catch (NumberFormatException e) { 178 throw new RuntimeException("Jronn fails to load models " + e.getLocalizedMessage(), e); 179 } catch (IOException e) { 180 throw new RuntimeException("Jronn fails to load models " + e.getLocalizedMessage(), e); 181 } 182 return disorder; 183 } 184 185 /** 186 * Calculates the disordered regions of the sequence. More formally, the regions for which the 187 * probability of disorder is greater then 0.50. 188 * 189 * 190 * @param sequence an instance of FastaSequence object, holding the name and the sequence. 191 * @return the array of ranges if there are any residues predicted to have the 192 * probability of disorder greater then 0.5, null otherwise. 193 * 194 */ 195 public static Range[] getDisorder(FastaSequence sequence) { 196 float[] scores = getDisorderScores(sequence); 197 return scoresToRanges(scores, RonnConstraint.DEFAULT_RANGE_PROBABILITY_THRESHOLD); 198 } 199 200 /** 201 * Convert raw scores to ranges. Gives ranges for given probability of disorder value 202 * @param scores the raw probability of disorder scores for each residue in the sequence. 203 * @param probability the cut off threshold. Include all residues with the probability of disorder greater then this value 204 * @return the array of ranges if there are any residues predicted to have the 205 * probability of disorder greater then {@code probability}, null otherwise. 206 */ 207 public static Range[] scoresToRanges(float[] scores, float probability) { 208 assert scores!=null && scores.length>0; 209 assert probability>0 && probability<1; 210 211 int count=0; 212 int regionLen=0; 213 List<Range> ranges = new ArrayList<Range>(); 214 for(float score: scores) { 215 count++; 216 // Round to 2 decimal points before comparison 217 score = (float) (Math.round(score*100.0)/100.0); 218 if(score>probability) { 219 regionLen++; 220 } else { 221 if(regionLen>0) { 222 ranges.add(new Range(count-regionLen, count-1,score)); 223 } 224 regionLen=0; 225 } 226 } 227 // In case of the range to boundary runs to the very end of the sequence 228 if(regionLen>1) { 229 ranges.add(new Range(count-regionLen+1, count,scores[scores.length-1])); 230 } 231 return ranges.toArray(new Range[ranges.size()]); 232 233 } 234 235 /** 236 * Calculates the probability of disorder scores for each residue in the sequence for 237 * many sequences in the input. 238 * 239 * @param sequences the list of the FastaSequence objects 240 * @return the Map with key->FastaSequence, value->probability of disorder for each residue 241 * @see #getDisorder(FastaSequence) 242 */ 243 public static Map<FastaSequence,float[]> getDisorderScores(List<FastaSequence> sequences) { 244 Map<FastaSequence,float[]> results = new TreeMap<FastaSequence, float[]>(); 245 for(FastaSequence fsequence : sequences) { 246 results.put(fsequence, predictSerial(fsequence)); 247 } 248 return results; 249 } 250 251 /** 252 * Calculates the disordered regions of the sequence for many sequences in the input. 253 * 254 * @param sequences sequences the list of the FastaSequence objects 255 * @return 256 * @see #getDisorder(FastaSequence) 257 */ 258 public static Map<FastaSequence,Range[]> getDisorder(List<FastaSequence> sequences) { 259 Map<FastaSequence,Range[]> disorderRanges = new TreeMap<FastaSequence,Range[]>(); 260 for(FastaSequence fs: sequences) { 261 disorderRanges.put(fs, getDisorder(fs)); 262 } 263 return disorderRanges; 264 } 265 266 /** 267 * Calculates the disordered regions of the protein sequence. 268 * @param fastaFile input file name containing the sequence in FASTA 269 * @return the Map with key->FastaSequence, value->the list of disordered regions for each sequence 270 * @throws FileNotFoundException if the input file cannot be found 271 * @throws IOException of the system cannot access or read from the input file 272 * @see #getDisorder(FastaSequence) 273 * @see #Jronn.Range 274 */ 275 public static Map<FastaSequence,Range[]> getDisorder(String fastaFile) throws FileNotFoundException, IOException { 276 final List<FastaSequence> sequences = SequenceUtil.readFasta(new FileInputStream(fastaFile)); 277 return getDisorder(sequences); 278 } 279 280 /** 281 * TODO 282 * 283 * High performance method for calculating disorder. Use multiple threads to achieve the speedup. 284 * 285 * @param fastaFile fully qualified path to the input FASTA file 286 * @param outputFile file name of the file for the results 287 * @param threadNumber the number of threads to use, default 288 * @param controls the format of the result file 289 * @throws FileNotFoundException if input file in not found 290 * @throws IOException if the input or the output files cannot be accessed 291 * @see ORonn.ResultLayout 292 293 public static void calculateDisorder(String fastaFile, String outputFile, int threadNumber, ResultLayout layout) throws FileNotFoundException, IOException { 294 final List<FastaSequence> sequences = SequenceUtil.readFasta(new FileInputStream(fastaFile)); 295 InputParameters in = new InputParameters(); 296 in.setFilePrm(fastaFile, InputParameters.inputKey); 297 in.setFilePrm(outputFile, InputParameters.outputKey); 298 //in.setThreadNum(Integer.toString(threadNumber)); 299 ORonn.predictParallel(sequences, in, loader); 300 } 301 */ 302}