X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FStructureFrequency.java;h=7b0858d502d70a7a5aac3582478afcc6cea7f1e0;hb=37de9310bec3501cbc6381e0c3dcb282fcaad812;hp=29c0a0b809d149640e9e5f0ba0e062a074ad26fb;hpb=a7c0ad7574fa26219f1908582e696dbbcd608548;p=jalview.git
diff --git a/src/jalview/analysis/StructureFrequency.java b/src/jalview/analysis/StructureFrequency.java
index 29c0a0b..7b0858d 100644
--- a/src/jalview/analysis/StructureFrequency.java
+++ b/src/jalview/analysis/StructureFrequency.java
@@ -1,25 +1,34 @@
/*
- * Jalview - A Sequence Alignment Editor and Viewer (Version 2.6)
- * Copyright (C) 2010 J Procter, AM Waterhouse, G Barton, M Clamp, S Searle
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
*
* This file is part of Jalview.
*
* Jalview is free software: you can redistribute it and/or
* modify it under the terms of the GNU General Public License
- * as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
- *
+ * as published by the Free Software Foundation, either version 3
+ * of the License, or (at your option) any later version.
+ *
* Jalview is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty
* of MERCHANTABILITY or FITNESS FOR A PARTICULAR
* PURPOSE. See the GNU General Public License for more details.
*
- * You should have received a copy of the GNU General Public License along with Jalview. If not, see .
+ * You should have received a copy of the GNU General Public License
+ * along with Jalview. If not, see .
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
*/
package jalview.analysis;
-import java.util.*;
+import jalview.datamodel.AlignmentAnnotation;
+import jalview.datamodel.Annotation;
+import jalview.datamodel.SequenceFeature;
+import jalview.datamodel.SequenceI;
+import jalview.util.Comparison;
+import jalview.util.Format;
-import jalview.datamodel.*;
+import java.util.ArrayList;
+import java.util.Hashtable;
/**
* Takes in a vector or array of sequences and column start and column end and
@@ -32,6 +41,8 @@ import jalview.datamodel.*;
*/
public class StructureFrequency
{
+ public static final int STRUCTURE_PROFILE_LENGTH = 74;
+
// No need to store 1000s of strings which are not
// visible to the user.
public static final String MAXCOUNT = "C";
@@ -44,300 +55,223 @@ public class StructureFrequency
public static final String PROFILE = "P";
- public static final Hashtable[] calculate(Vector sequences, int start,
- int end)
- {
- return calculate(sequences, start, end, false);
- }
+ public static final String PAIRPROFILE = "B";
- public static final Hashtable[] calculate(Vector sequences, int start,
- int end, boolean profile)
+ /**
+ * Returns the 3' position of a base pair
+ *
+ * @param pairs
+ * Secondary structure annotation
+ * @param indice
+ * 5' position of a base pair
+ * @return 3' position of a base pair
+ */
+ public static int findPair(SequenceFeature[] pairs, int indice)
{
- SequenceI[] seqs = new SequenceI[sequences.size()];
- int width = 0;
- for (int i = 0; i < sequences.size(); i++)
+
+ for (int i = 0; i < pairs.length; i++)
{
- seqs[i] = (SequenceI) sequences.elementAt(i);
- if (seqs[i].getLength() > width)
+ if (pairs[i].getBegin() == indice)
+
{
- width = seqs[i].getLength();
- }
- }
- Hashtable[] reply = new Hashtable[width];
+ return pairs[i].getEnd();
- if (end >= width)
- {
- end = width;
+ }
}
-
- calculate(seqs, start, end, reply, profile);
-
- return reply;
+ return -1;
}
+ /**
+ * Method to calculate a 'base pair consensus row', very similar to nucleotide
+ * consensus but takes into account a given structure
+ *
+ * @param sequences
+ * @param start
+ * @param end
+ * @param result
+ * @param profile
+ * @param rnaStruc
+ */
public static final void calculate(SequenceI[] sequences, int start,
- int end, Hashtable[] result)
+ int end, Hashtable[] result, boolean profile,
+ AlignmentAnnotation rnaStruc)
{
- calculate(sequences, start, end, result, false);
- }
- public static final void calculate(SequenceI[] sequences, int start,
- int end, Hashtable[] result, boolean profile)
- {
Hashtable residueHash;
- int maxCount, nongap, i, j, v, jSize = sequences.length;
String maxResidue;
- char c;
+ char[] struc = rnaStruc.getRNAStruc().toCharArray();
+
+ SequenceFeature[] rna = rnaStruc._rnasecstr;
+ char c, s, cEnd;
+ int bpEnd = -1;
+ int jSize = sequences.length;
+ int[] values;
+ int[][] pairs;
float percentage;
- int[] values = new int[255];
-
- char[] seq;
-
- for (i = start; i < end; i++)
+ for (int i = start; i < end; i++) // foreach column
{
- residueHash = new Hashtable();
- maxCount = 0;
- maxResidue = "";
- nongap = 0;
+ int canonicalOrWobblePairCount = 0, canonical = 0;
+ int otherPairCount = 0;
+ int nongap = 0;
+ maxResidue = "-";
values = new int[255];
-
+ pairs = new int[255][255];
+ bpEnd = -1;
+ if (i < struc.length)
+ {
+ s = struc[i];
+ }
+ else
+ {
+ s = '-';
+ }
+ if (s == '.' || s == ' ')
+ {
+ s = '-';
+ }
- for (j = 0; j < jSize; j++)
+ if (!Rna.isOpeningParenthesis(s))
{
- if (sequences[j]==null)
+ if (s == '-')
{
- System.err.println("WARNING: Consensus skipping null sequence - possible race condition.");
- continue;
+ values['-']++;
}
- seq = sequences[j].getSequence();
- if (seq.length > i)
- {
- c = seq[i];
-
- if (c == '.' || c == ' ')
- {
- c = '-';
- }
+ }
+ else
+ {
+ bpEnd = findPair(rna, i);
- if (c == '-')
- {
- values['-']++;
- continue;
- }
- else if ('a' <= c && c <= 'z')
+ if (bpEnd > -1)
+ {
+ for (int j = 0; j < jSize; j++) // foreach row
{
- c -= 32; // ('a' - 'A');
- }
+ if (sequences[j] == null)
+ {
+ System.err
+ .println("WARNING: Consensus skipping null sequence - possible race condition.");
+ continue;
+ }
- nongap++;
- values[c]++;
+ c = sequences[j].getCharAt(i);
+ cEnd = sequences[j].getCharAt(bpEnd);
- }
- else
- {
- values['-']++;
+ if (Comparison.isGap(c) || Comparison.isGap(cEnd))
+ {
+ values['-']++;
+ continue;
+ }
+ nongap++;
+ /*
+ * ensure upper-case for counting purposes
+ */
+ if ('a' <= c && 'z' >= c)
+ {
+ c += 'A' - 'a';
+ }
+ if ('a' <= cEnd && 'z' >= cEnd)
+ {
+ cEnd += 'A' - 'a';
+ }
+ if (Rna.isCanonicalOrWobblePair(c, cEnd))
+ {
+ canonicalOrWobblePairCount++;
+ if (Rna.isCanonicalPair(c, cEnd))
+ {
+ canonical++;
+ }
+ }
+ else
+ {
+ otherPairCount++;
+ }
+ pairs[c][cEnd]++;
+ }
}
}
- for (v = 'A'; v < 'Z'; v++)
+ residueHash = new Hashtable();
+ if (profile)
{
- if (values[v] < 2 || values[v] < maxCount)
- {
- continue;
- }
+ // TODO 1-dim array with jsize in [0], nongapped in [1]; or Pojo
+ residueHash.put(PROFILE, new int[][] { values,
+ new int[] { jSize, (jSize - values['-']) } });
- if (values[v] > maxCount)
+ residueHash.put(PAIRPROFILE, pairs);
+ }
+ values['('] = canonicalOrWobblePairCount;
+ values['['] = canonical;
+ values['{'] = otherPairCount;
+ /*
+ * the count is the number of valid pairs (as a percentage, determines
+ * the relative size of the profile logo)
+ */
+ int count = canonicalOrWobblePairCount;
+
+ /*
+ * display '(' if most pairs are canonical, or as
+ * '[' if there are more wobble pairs.
+ */
+ if (canonicalOrWobblePairCount > 0 || otherPairCount > 0)
+ {
+ if (canonicalOrWobblePairCount >= otherPairCount)
{
- maxResidue = String.valueOf((char) v);
+ maxResidue = (canonicalOrWobblePairCount - canonical) < canonical ? "("
+ : "[";
}
- else if (values[v] == maxCount)
+ else
{
- maxResidue += String.valueOf((char) v);
+ maxResidue = "{";
}
- maxCount = values[v];
}
+ residueHash.put(MAXCOUNT, new Integer(count));
+ residueHash.put(MAXRESIDUE, maxResidue);
- if (maxResidue.length() == 0)
+ percentage = ((float) count * 100) / jSize;
+ residueHash.put(PID_GAPS, new Float(percentage));
+
+ percentage = ((float) count * 100) / nongap;
+ residueHash.put(PID_NOGAPS, new Float(percentage));
+
+ if (result[i] == null)
{
- maxResidue = "-";
+ result[i] = residueHash;
}
- if (profile)
+ if (bpEnd > 0)
{
- residueHash.put(PROFILE, new int[][]
- { values, new int[]
- { jSize, nongap } });
- }
- residueHash.put(MAXCOUNT, new Integer(maxCount));
- residueHash.put(MAXRESIDUE, maxResidue);
+ values[')'] = values['('];
+ values[']'] = values['['];
+ values['}'] = values['{'];
+ values['('] = 0;
+ values['['] = 0;
+ values['{'] = 0;
+ maxResidue = maxResidue.equals("(") ? ")"
+ : maxResidue.equals("[") ? "]" : "}";
+
+ residueHash = new Hashtable();
+ if (profile)
+ {
+ residueHash.put(PROFILE, new int[][] { values,
+ new int[] { jSize, (jSize - values['-']) } });
- percentage = ((float) maxCount * 100) / (float) jSize;
- residueHash.put(PID_GAPS, new Float(percentage));
+ residueHash.put(PAIRPROFILE, pairs);
+ }
- percentage = ((float) maxCount * 100) / (float) nongap;
- residueHash.put(PID_NOGAPS, new Float(percentage));
- result[i] = residueHash;
+ residueHash.put(MAXCOUNT, new Integer(count));
+ residueHash.put(MAXRESIDUE, maxResidue);
+
+ percentage = ((float) count * 100) / jSize;
+ residueHash.put(PID_GAPS, new Float(percentage));
+
+ percentage = ((float) count * 100) / nongap;
+ residueHash.put(PID_NOGAPS, new Float(percentage));
+
+ result[bpEnd] = residueHash;
+ }
}
}
-
- /**
- * Method to calculate a 'base pair consensus row', very similar
- * to nucleotide consensus but takes into account a given structure
- * @param sequences
- * @param start
- * @param end
- * @param result
- * @param profile
- * @param rnaStruc
- */
- public static final void calculate(SequenceI[] sequences, int start,
- int end, Hashtable[] result, boolean profile, AlignmentAnnotation rnaStruc){
-
- Hashtable residueHash;
- String maxResidue;
- char[] seq, struc=rnaStruc.getRNAStruc().toCharArray();
- SequenceFeature[] rna =rnaStruc._rnasecstr;
- char c,s,cEnd;
- int count,nonGap=0,i,bpEnd=-1,j,jSize = sequences.length;
- int[] values = new int[255];
- float percentage;
-
- for (i = start; i < end; i++) //foreach column
- {
- residueHash = new Hashtable();
- maxResidue="";
- values = new int[255];
- bpEnd=-1;
-
- s = struc[i];
- if (s == '.' || s == ' ')
- {
- s = '-';
- }
-
- if(s == '-'){
- values['-']++;
- }
- else
- {
- for (j = 0; j < jSize; j++) //foreach row
- {
- if (sequences[j]==null)
- {
- System.err.println("WARNING: Consensus skipping null sequence - possible race condition.");
- continue;
- }
- seq = sequences[j].getSequence();
-
- if (seq.length > i)
- {
- c = seq[i];
-
- //standard representation for gaps in sequence and structure
- if (c == '.' || c == ' ')
- {
- c = '-';
- }
-
- if (c == '-')
- {
- values['-']++;
- continue;
- }
- if(s == '('){
- bpEnd=rna[(rna.length-1-nonGap)].getEnd();
- cEnd=seq[bpEnd];
- if(checkBpType(c,cEnd)){
- values['H']++; //H means it's a helix (structured)
- }
- }
- }
- }
- nonGap++;
- }
- /*UPDATE this for new values
- if (profile)
- {
- residueHash.put(PROFILE, new int[][]
- { values, new int[]
- { jSize, nongap } });
- }
- */
-
- count=values['H'];
-
- if (count == 0)
- {
- maxResidue = "-";
- }else{
- maxResidue="H";
- }
-
- residueHash.put(MAXCOUNT, new Integer(count));
- residueHash.put(MAXRESIDUE, maxResidue);
-
- percentage = ((float) count * 100) / (float) jSize;
- residueHash.put(PID_GAPS, new Float(percentage));
-
- //percentage = ((float) count * 100) / (float) nongap;
- //residueHash.put(PID_NOGAPS, new Float(percentage));
- if(result[i]==null){
- result[i] = residueHash;
- }
- if(bpEnd>0){
- result[bpEnd]=residueHash;
- }
-
- }
- }
-
- /**
- * Method to check if a base-pair is a canonical or a wobble bp
- * @param up 5' base
- * @param down 3' base
- * @return True if it is a canonical/wobble bp
- */
- public static boolean checkBpType(char up, char down){
- if(up>'Z'){up-=32;}
- if(down>'Z'){down-=32;}
-
- switch (up){
- case 'A':
- switch (down){
- case 'T': return true;
- case 'U': return true;
- }
- break;
- case 'C':
- switch (down){
- case 'G': return true;
- }
- break;
- case 'T':
- switch (down){
- case 'A': return true;
- case 'G': return true;
- }
- break;
- case 'G':
- switch (down){
- case 'C': return true;
- case 'T': return true;
- case 'U': return true;
- }
- break;
- case 'U':
- switch (down){
- case 'A': return true;
- case 'G': return true;
- }
- break;
- }
- return false;
- }
-
/**
* Compute all or part of the annotation row from the given consensus
* hashtable
@@ -353,18 +287,7 @@ public class StructureFrequency
public static void completeConsensus(AlignmentAnnotation consensus,
Hashtable[] hconsensus, int iStart, int width,
boolean ignoreGapsInConsensusCalculation,
- boolean includeAllConsSymbols)
- {
- completeConsensus(consensus, hconsensus, iStart, width,
- ignoreGapsInConsensusCalculation, includeAllConsSymbols, null); // new
- // char[]
- // { 'A', 'C', 'G', 'T', 'U' });
- }
-
- public static void completeConsensus(AlignmentAnnotation consensus,
- Hashtable[] hconsensus, int iStart, int width,
- boolean ignoreGapsInConsensusCalculation,
- boolean includeAllConsSymbols, char[] alphabet)
+ boolean includeAllConsSymbols, long nseq)
{
float tval, value;
if (consensus == null || consensus.annotations == null
@@ -374,9 +297,23 @@ public class StructureFrequency
// initialised properly
return;
}
+ String fmtstr = "%3.1f";
+ int precision = 2;
+ while (nseq > 100)
+ {
+ precision++;
+ nseq /= 10;
+ }
+ if (precision > 2)
+ {
+ fmtstr = "%" + (2 + precision) + "." + precision + "f";
+ }
+ Format fmt = new Format(fmtstr);
+
for (int i = iStart; i < width; i++)
{
- if (i >= hconsensus.length)
+ Hashtable hci;
+ if (i >= hconsensus.length || ((hci = hconsensus[i]) == null))
{
// happens if sequences calculated over were shorter than alignment
// width
@@ -384,72 +321,85 @@ public class StructureFrequency
continue;
}
value = 0;
+ Float fv;
if (ignoreGapsInConsensusCalculation)
{
- value = ((Float) hconsensus[i].get(StructureFrequency.PID_NOGAPS))
- .floatValue();
+ fv = (Float) hci.get(StructureFrequency.PID_NOGAPS);
}
else
{
- value = ((Float) hconsensus[i].get(StructureFrequency.PID_GAPS))
- .floatValue();
+ fv = (Float) hci.get(StructureFrequency.PID_GAPS);
}
-
- String maxRes = hconsensus[i].get(StructureFrequency.MAXRESIDUE).toString();
- String mouseOver = hconsensus[i].get(StructureFrequency.MAXRESIDUE) + " ";
+ if (fv == null)
+ {
+ consensus.annotations[i] = null;
+ // data has changed below us .. give up and
+ continue;
+ }
+ value = fv.floatValue();
+ String maxRes = hci.get(StructureFrequency.MAXRESIDUE).toString();
+ String mouseOver = hci.get(StructureFrequency.MAXRESIDUE) + " ";
if (maxRes.length() > 1)
{
mouseOver = "[" + maxRes + "] ";
maxRes = "+";
}
- int[][] profile = (int[][]) hconsensus[i].get(StructureFrequency.PROFILE);
- if (profile != null && includeAllConsSymbols)
+ int[][] profile = (int[][]) hci.get(StructureFrequency.PROFILE);
+ int[][] pairs = (int[][]) hci.get(StructureFrequency.PAIRPROFILE);
+
+ if (pairs != null && includeAllConsSymbols) // Just responsible for the
+ // tooltip
+ // TODO Update tooltips for Structure row
{
mouseOver = "";
- if (alphabet != null)
+
+ /*
+ * TODO It's not sure what is the purpose of the alphabet and wheter it
+ * is useful for structure?
+ *
+ * if (alphabet != null) { for (int c = 0; c < alphabet.length; c++) {
+ * tval = ((float) profile[0][alphabet[c]]) 100f / (float)
+ * profile[1][ignoreGapsInConsensusCalculation ? 1 : 0]; mouseOver +=
+ * ((c == 0) ? "" : "; ") + alphabet[c] + " " + ((int) tval) + "%"; } }
+ * else {
+ */
+ int[][] ca = new int[625][];
+ float[] vl = new float[625];
+ int x = 0;
+ for (int c = 65; c < 90; c++)
{
- for (int c = 0; c < alphabet.length; c++)
+ for (int d = 65; d < 90; d++)
{
- tval = ((float) profile[0][alphabet[c]])
- * 100f
- / (float) profile[1][ignoreGapsInConsensusCalculation ? 1
- : 0];
- mouseOver += ((c == 0) ? "" : "; ") + alphabet[c] + " "
- + ((int) tval) + "%";
+ ca[x] = new int[] { c, d };
+ vl[x] = pairs[c][d];
+ x++;
}
}
- else
+ jalview.util.QuickSort.sort(vl, ca);
+ int p = 0;
+
+ /*
+ * profile[1] is {total, ungappedTotal}
+ */
+ final int divisor = profile[1][ignoreGapsInConsensusCalculation ? 1
+ : 0];
+ for (int c = 624; c > 0; c--)
{
- Object[] ca = new Object[profile[0].length];
- float[] vl = new float[profile[0].length];
- for (int c = 0; c < ca.length; c++)
- {
- ca[c] = new char[]
- { (char) c };
- vl[c] = (float) profile[0][c];
- }
- ;
- jalview.util.QuickSort.sort(vl, ca);
- for (int p = 0, c = ca.length - 1; profile[0][((char[]) ca[c])[0]] > 0; c--)
+ if (vl[c] > 0)
{
- if (((char[]) ca[c])[0] != '-')
- {
- tval = ((float) profile[0][((char[]) ca[c])[0]])
- * 100f
- / (float) profile[1][ignoreGapsInConsensusCalculation ? 1
- : 0];
- mouseOver += ((p == 0) ? "" : "; ") + ((char[]) ca[c])[0]
- + " " + ((int) tval) + "%";
- p++;
+ tval = (vl[c] * 100f / divisor);
+ mouseOver += ((p == 0) ? "" : "; ") + (char) ca[c][0]
+ + (char) ca[c][1] + " " + fmt.form(tval) + "%";
+ p++;
- }
}
-
}
+
+ // }
}
else
{
- mouseOver += ((int) value + "%");
+ mouseOver += (fmt.form(value) + "%");
}
consensus.annotations[i] = new Annotation(maxRes, mouseOver, ' ',
value);
@@ -457,56 +407,80 @@ public class StructureFrequency
}
/**
- * get the sorted profile for the given position of the consensus
+ * get the sorted base-pair profile for the given position of the consensus
*
* @param hconsensus
- * @return
+ * @return profile of the given column
*/
public static int[] extractProfile(Hashtable hconsensus,
boolean ignoreGapsInConsensusCalculation)
{
- int[] rtnval = new int[64];
+ int[] rtnval = new int[STRUCTURE_PROFILE_LENGTH]; // 2*(5*5)+2
int[][] profile = (int[][]) hconsensus.get(StructureFrequency.PROFILE);
+ int[][] pairs = (int[][]) hconsensus
+ .get(StructureFrequency.PAIRPROFILE);
+
if (profile == null)
+ {
return null;
- Object[] ca = new Object[profile[0].length];
- float[] vl = new float[profile[0].length];
- for (int c = 0; c < ca.length; c++)
+ }
+
+ // TODO fix the object length, also do it in completeConsensus
+ // Object[] ca = new Object[625];
+ int[][] ca = new int[625][];
+ float[] vl = new float[625];
+ int x = 0;
+ for (int c = 65; c < 90; c++)
{
- ca[c] = new char[]
- { (char) c };
- vl[c] = (float) profile[0][c];
+ for (int d = 65; d < 90; d++)
+ {
+ ca[x] = new int[] { c, d };
+ vl[x] = pairs[c][d];
+ x++;
+ }
}
- ;
jalview.util.QuickSort.sort(vl, ca);
- rtnval[0] = 1;
- for (int c = ca.length - 1; profile[0][((char[]) ca[c])[0]] > 0; c--)
+
+ int valuesCount = 0;
+ rtnval[1] = 0;
+ int offset = 2;
+ final int divisor = profile[1][ignoreGapsInConsensusCalculation ? 1 : 0];
+ for (int c = 624; c > 0; c--)
{
- if (((char[]) ca[c])[0] != '-')
+ if (vl[c] > 0)
{
- rtnval[rtnval[0]++] = ((char[]) ca[c])[0];
- rtnval[rtnval[0]++] = (int) (((float) profile[0][((char[]) ca[c])[0]]) * 100f / (float) profile[1][ignoreGapsInConsensusCalculation ? 1
- : 0]);
+ rtnval[offset++] = ca[c][0];
+ rtnval[offset++] = ca[c][1];
+ rtnval[offset] = (int) (vl[c] * 100f / divisor);
+ rtnval[1] += rtnval[offset++];
+ valuesCount++;
}
}
- return rtnval;
+ rtnval[0] = valuesCount;
+
+ // insert profile type code in position 0
+ int[] result = new int[rtnval.length + 1];
+ result[0] = AlignmentAnnotation.STRUCTURE_PROFILE;
+ System.arraycopy(rtnval, 0, result, 1, rtnval.length);
+ return result;
}
- enum base {A,T,g,C};
-
-
- public static void main(String args[]){
- //Short test to see if checkBpType works
- ArrayList test = new ArrayList();
- test.add("A");
- test.add("c");
- test.add("g");
- test.add("T");
- test.add("U");
- for (String i : test) {
- for (String j : test) {
- System.out.println(i+"-"+j+": "+StructureFrequency.checkBpType(i.charAt(0),j.charAt(0)));
- }
- }
+ public static void main(String args[])
+ {
+ // Short test to see if checkBpType works
+ ArrayList test = new ArrayList();
+ test.add("A");
+ test.add("c");
+ test.add("g");
+ test.add("T");
+ test.add("U");
+ for (String i : test)
+ {
+ for (String j : test)
+ {
+ System.out.println(i + "-" + j + ": "
+ + Rna.isCanonicalOrWobblePair(i.charAt(0), j.charAt(0)));
+ }
+ }
}
}