import jalview.analysis.scoremodels.PairwiseDistanceModel;
import jalview.analysis.scoremodels.ScoreMatrix;
import jalview.analysis.scoremodels.ScoreModels;
-import jalview.datamodel.BinarySequence;
-import jalview.datamodel.BinarySequence.InvalidSequenceTypeException;
-import jalview.math.Matrix;
+import jalview.math.MatrixI;
import java.io.PrintStream;
/**
* Performs Principal Component Analysis on given sequences
- *
- * @author $author$
- * @version $Revision$
*/
public class PCA implements Runnable
{
- Matrix m;
-
- Matrix symm;
+ boolean jvCalcMode = true;
- Matrix m2;
+ MatrixI symm;
double[] eigenvalue;
- Matrix eigenvector;
+ MatrixI eigenvector;
- StringBuffer details = new StringBuffer();
+ StringBuilder details = new StringBuilder(1024);
+
+ private String[] seqs;
+
+ private ScoreMatrix scoreMatrix;
/**
* Creates a new PCA object. By default, uses blosum62 matrix to generate
public PCA(String[] s, boolean nucleotides, String s_m)
{
+ this.seqs = s;
- BinarySequence[] bs = new BinarySequence[s.length];
- int ii = 0;
-
- while ((ii < s.length) && (s[ii] != null))
- {
- bs[ii] = new BinarySequence(s[ii], nucleotides);
- bs[ii].encode();
- ii++;
- }
-
- BinarySequence[] bs2 = new BinarySequence[s.length];
- ii = 0;
- ScoreMatrix smtrx = null;
+ scoreMatrix = null;
String sm = s_m;
if (sm != null)
{
- smtrx = (ScoreMatrix) ((PairwiseDistanceModel) ScoreModels
+ scoreMatrix = (ScoreMatrix) ((PairwiseDistanceModel) ScoreModels
.getInstance()
.forName(sm)).getScoreModel();
}
- if (smtrx == null)
+ if (scoreMatrix == null)
{
// either we were given a non-existent score matrix or a scoremodel that
// isn't based on a pairwise symbol score matrix
- smtrx = ScoreModels.getInstance().getDefaultModel(!nucleotides);
+ scoreMatrix = ScoreModels.getInstance().getDefaultModel(!nucleotides);
}
details.append("PCA calculation using " + sm
+ " sequence similarity matrix\n========\n\n");
- while ((ii < s.length) && (s[ii] != null))
- {
- bs2[ii] = new BinarySequence(s[ii], nucleotides);
- if (smtrx != null)
- {
- try
- {
- bs2[ii].matrixEncode(smtrx);
- } catch (InvalidSequenceTypeException x)
- {
- details.append("Unexpected mismatch of sequence type and score matrix. Calculation will not be valid!\n\n");
- }
- }
- ii++;
- }
-
- // System.out.println("Created binary encoding");
- // printMemory(rt);
- int count = 0;
-
- while ((count < bs.length) && (bs[count] != null))
- {
- count++;
- }
-
- double[][] seqmat = new double[count][bs[0].getDBinary().length];
- double[][] seqmat2 = new double[count][bs2[0].getDBinary().length];
- int i = 0;
-
- while (i < count)
- {
- seqmat[i] = bs[i].getDBinary();
- seqmat2[i] = bs2[i].getDBinary();
- i++;
- }
-
- // System.out.println("Created array");
- // printMemory(rt);
- // System.out.println(" --- Original matrix ---- ");
- m = new Matrix(seqmat);
- m2 = new Matrix(seqmat2);
-
- }
-
- /**
- * Returns the matrix used in PCA calculation
- *
- * @return java.math.Matrix object
- */
-
- public Matrix getM()
- {
- return m;
}
/**
*/
public double getEigenvalue(int i)
{
- return eigenvector.d[i];
+ return eigenvector.getD()[i];
}
/**
*/
public float[][] getComponents(int l, int n, int mm, float factor)
{
- float[][] out = new float[m.rows][3];
+ float[][] out = new float[getHeight()][3];
- for (int i = 0; i < m.rows; i++)
+ for (int i = 0; i < getHeight(); i++)
{
out[i][0] = (float) component(i, l) * factor;
out[i][1] = (float) component(i, n) * factor;
public double[] component(int n)
{
// n = index of eigenvector
- double[] out = new double[m.rows];
+ double[] out = new double[getHeight()];
- for (int i = 0; i < m.rows; i++)
+ for (int i = 0; i < out.length; i++)
{
out[i] = component(i, n);
}
{
double out = 0.0;
- for (int i = 0; i < symm.cols; i++)
+ for (int i = 0; i < symm.width(); i++)
{
- out += (symm.value[row][i] * eigenvector.value[i][n]);
+ out += (symm.getValue(row, i) * eigenvector.getValue(i, n));
}
- return out / eigenvector.d[n];
+ return out / eigenvector.getD()[n];
}
public String getDetails()
}
};
+ // long now = System.currentTimeMillis();
try
{
details.append("PCA Calculation Mode is "
+ (jvCalcMode ? "Jalview variant" : "Original SeqSpace")
+ "\n");
- Matrix mt = m.transpose();
- details.append(" --- OrigT * Orig ---- \n");
- if (!jvCalcMode)
- {
- eigenvector = mt.preMultiply(m); // standard seqspace comparison matrix
- }
- else
- {
- eigenvector = mt.preMultiply(m2); // jalview variation on seqsmace
- // method
- }
+ eigenvector = scoreMatrix.computePairwiseScores(seqs);
- eigenvector.print(ps);
+ details.append(" --- OrigT * Orig ---- \n");
+ eigenvector.print(ps, "%8.2f");
symm = eigenvector.copy();
details.append(" ---Tridiag transform matrix ---\n");
details.append(" --- D vector ---\n");
- eigenvector.printD(ps);
+ eigenvector.printD(ps, "%15.4e");
ps.println();
details.append("--- E vector ---\n");
- eigenvector.printE(ps);
+ eigenvector.printE(ps, "%15.4e");
ps.println();
// Now produce the diagonalization matrix
}
details.append(" --- New diagonalization matrix ---\n");
- eigenvector.print(ps);
+ eigenvector.print(ps, "%8.2f");
details.append(" --- Eigenvalues ---\n");
- eigenvector.printD(ps);
+ eigenvector.printD(ps, "%15.4e");
ps.println();
/*
* for (int seq=0;seq<symm.rows;seq++) { ps.print("\"Seq"+seq+"\""); for
*
* ps.print(","+component(seq, ev)); } ps.println(); }
*/
+ // System.out.println(("PCA.run() took "
+ // + (System.currentTimeMillis() - now) + "ms"));
}
- boolean jvCalcMode = true;
-
public void setJvCalcMode(boolean calcMode)
{
this.jvCalcMode = calcMode;
}
+
+ /**
+ * Answers the N dimensions of the NxN PCA matrix. This is the number of
+ * sequences involved in the pairwise score calculation.
+ *
+ * @return
+ */
+ public int getHeight()
+ {
+ // TODO can any of seqs[] be null?
+ return seqs.length;
+ }
}
git://source.jalview.org / jalview.git / blobdiff