package jalview.viewmodel; import java.util.Vector; import jalview.analysis.PCA; import jalview.datamodel.AlignmentView; import jalview.datamodel.SequenceI; import jalview.datamodel.SequencePoint; import jalview.api.RotatableCanvasI; public class PCAModel { public PCAModel(AlignmentView seqstrings2, SequenceI[] seqs2, boolean nucleotide2) { seqstrings=seqstrings2; seqs=seqs2; nucleotide=nucleotide2; } PCA pca; int top; AlignmentView seqstrings; SequenceI[] seqs; /** * use the identity matrix for calculating similarity between sequences. */ private boolean nucleotide=false; private Vector points; public void run() { pca = new PCA(seqstrings.getSequenceStrings(' '), nucleotide); pca.run(); // Now find the component coordinates int ii = 0; while ((ii < seqs.length) && (seqs[ii] != null)) { ii++; } double[][] comps = new double[ii][ii]; for (int i = 0; i < ii; i++) { if (pca.getEigenvalue(i) > 1e-4) { comps[i] = pca.component(i); } } top = pca.getM().rows - 1; points = new Vector(); float[][] scores = pca.getComponents(top - 1, top - 2, top - 3, 100); for (int i = 0; i < pca.getM().rows; i++) { SequencePoint sp = new SequencePoint(seqs[i], scores[i]); points.addElement(sp); } } public void updateRc(RotatableCanvasI rc) { rc.setPoints(points, pca.getM().rows); } public boolean isNucleotide() { return nucleotide; } public void setNucleotide(boolean nucleotide) { this.nucleotide=nucleotide; } /** * * * @return index of principle dimension of PCA */ public int getTop() { return top; } /** * update the 2d coordinates for the list of points to the given dimensions * Principal dimension is getTop(). Next greatest eigenvector is getTop()-1. * Note - pca.getComponents starts counting the spectrum from rank-2 to zero, rather than rank-1, so getComponents(dimN ...) == updateRcView(dimN+1 ..) * @param dim1 * @param dim2 * @param dim3 */ public void updateRcView(int dim1, int dim2, int dim3) { // note: actual indices for components are dim1-1, etc (patch for JAL-1123) float[][] scores = pca.getComponents(dim1-1, dim2-1, dim3-1, 100); for (int i = 0; i < pca.getM().rows; i++) { ((SequencePoint) points.elementAt(i)).coord = scores[i]; } } public String getDetails() { return pca.getDetails(); } public AlignmentView getSeqtrings() { return seqstrings; } public String getPointsasCsv(boolean transformed, int xdim, int ydim, int zdim) { StringBuffer csv = new StringBuffer(); csv.append("\"Sequence\""); if (transformed) { csv.append(","); csv.append(xdim); csv.append(","); csv.append(ydim); csv.append(","); csv.append(zdim); } else { for (int d = 1, dmax = pca.component(1).length; d <= dmax; d++) { csv.append("," + d); } } csv.append("\n"); for (int s = 0; s < seqs.length; s++) { csv.append("\"" + seqs[s].getName() + "\""); double fl[]; if (!transformed) { // output pca in correct order fl = pca.component(s); for (int d = fl.length - 1; d >= 0; d--) { csv.append(","); csv.append(fl[d]); } } else { // output current x,y,z coords for points fl = getPointPosition(s); for (int d = 0; d < fl.length; d++) { csv.append(","); csv.append(fl[d]); } } csv.append("\n"); } return csv.toString(); } /** * * @return x,y,z positions of point s (index into points) under current * transform. */ public double[] getPointPosition(int s) { double pts[] = new double[3]; float[] p = points.elementAt(s).coord; pts[0] = p[0]; pts[1] = p[1]; pts[2] = p[2]; return pts; } }