--- /dev/null
+<html>\r
+<head><title>Alignment Quality Annotation</title></head>\r
+<body>\r
+<p><strong>Alignment Quality Annotation</strong></p>\r
+<p>Alignment Quality is one of the automatically calculated\r
+quantitative alignment\r
+annotations displayed below the columns of a multiple sequence\r
+alignment (and can be used to shade the alignment). It is an ad-hoc\r
+measure of the likelihood of observing the mutations (if any) in a\r
+particular column of the alignment.</p>\r
+<p>\r
+More precisely, the quality score is inversely proportional to the\r
+average cost of all pairs of mutations oberved in a particular column\r
+of the alignment - a high alignment quality score for a column would\r
+suggest that there are no mutations, or most mutations observed are\r
+favourable.\r
+</p>\r
+\r
+<p><em>The Algorithm</em><br>\r
+The quality score is calculated for each column in an alignment by\r
+summing, for all mutations, the ratio of the two BLOSUM 62 scores for\r
+a mutation pair and each residue's conservered BLOSUM62 score (which\r
+is higher). This valueis normalised for each column, and then plotted\r
+on a scale from 0 to 1.\r
+</p>\r
+<p>\r
+Multiple alignment algorithms using the BLOSUM 62 substition matrices\r
+should, in theory, maximise alignment quality for an un-gapped\r
+alignment, and locally maximise quality for gapped alignments.\r
+</p>\r
+</body>\r
+</html>\r
git://source.jalview.org / jalview.git / blobdiff