package jalview.analysis.scoremodels;
import jalview.api.analysis.PairwiseScoreModelI;
+import jalview.api.analysis.SimilarityParamsI;
import jalview.api.analysis.SimilarityScoreModelI;
import jalview.datamodel.AlignmentView;
import jalview.math.Matrix;
import jalview.math.MatrixI;
+import jalview.util.Comparison;
import java.util.Arrays;
*/
private boolean peptide;
- private boolean symmetric;
-
/**
* Constructor given a name, symbol alphabet, and matrix of scores for pairs
* of symbols. The matrix should be square and of the same size as the
* </ul>
*/
@Override
- public MatrixI findSimilarities(AlignmentView seqstrings)
+ public MatrixI findSimilarities(AlignmentView seqstrings,
+ SimilarityParamsI options)
{
char gapChar = scoreGapAsAny ? (seqstrings.isNa() ? 'N' : 'X') : ' ';
String[] seqs = seqstrings.getSequenceStrings(gapChar);
- return findSimilarities(seqs);
+ return findSimilarities(seqs, options);
}
/**
+ * Computes pairwise similarities of a set of sequences using the given
+ * parameters
+ *
* @param seqs
+ * @param params
* @return
*/
- protected MatrixI findSimilarities(String[] seqs)
+ protected MatrixI findSimilarities(String[] seqs, SimilarityParamsI params)
{
double[][] values = new double[seqs.length][];
for (int row = 0; row < seqs.length; row++)
values[row] = new double[seqs.length];
for (int col = 0; col < seqs.length; col++)
{
- int total = 0;
- int width = Math.min(seqs[row].length(), seqs[col].length());
- for (int i = 0; i < width; i++)
- {
- char c1 = seqs[row].charAt(i);
- char c2 = seqs[col].charAt(i);
- float score = getPairwiseScore(c1, c2);
- total += score;
- }
+ double total = computeSimilarity(seqs[row], seqs[col], params);
values[row][col] = total;
}
}
}
/**
+ * Calculates the pairwise similarity of two strings using the given
+ * calculation parameters
+ *
+ * @param seq1
+ * @param seq2
+ * @param params
+ * @return
+ */
+ protected double computeSimilarity(String seq1, String seq2,
+ SimilarityParamsI params)
+ {
+ int len1 = seq1.length();
+ int len2 = seq2.length();
+ double total = 0;
+
+ int width = Math.max(len1, len2);
+ for (int i = 0; i < width; i++)
+ {
+ if (i >= len1 || i >= len2)
+ {
+ /*
+ * off the end of one sequence; stop if we are only matching
+ * on the shorter sequence length, else treat as trailing gap
+ */
+ if (params.denominateByShortestLength())
+ {
+ break;
+ }
+ }
+ // Change GAP_SPACE to GAP_DASH if we adopt - for gap in matrices
+ char c1 = i >= len1 ? Comparison.GAP_SPACE : seq1.charAt(i);
+ char c2 = i >= len2 ? Comparison.GAP_SPACE : seq2.charAt(i);
+ boolean gap1 = Comparison.isGap(c1);
+ boolean gap2 = Comparison.isGap(c2);
+
+ if (gap1 && gap2)
+ {
+ /*
+ * gap-gap: include if options say so, else ignore
+ */
+ if (!params.includeGappedColumns())
+ {
+ continue;
+ }
+ }
+ else if (gap1 || gap2)
+ {
+ /*
+ * gap-residue: score if options say so
+ */
+ if (!params.includesGaps())
+ {
+ continue;
+ }
+ }
+ float score = getPairwiseScore(c1, c2);
+ total += score;
+ }
+ return total;
+ }
+
+ /**
* Answers a hashcode computed from the symbol alphabet and the matrix score
* values
*/
return false;
}
- public boolean isSymmetric()
- {
- return symmetric;
- }
-
/**
* Returns the alphabet the matrix scores for, as a string of characters
*