+ /**
+ * Returns a mapping given list of one or more Align descriptors (exonerate
+ * format)
+ *
+ * @param alignedRegions
+ * a list of "Align fromStart toStart fromCount"
+ * @param mapIsFromCdna
+ * if true, 'from' is dna, else 'from' is protein
+ * @param strand
+ * either 1 (forward) or -1 (reverse)
+ * @return
+ * @throws IOException
+ */
+ protected MapList constructCodonMappingFromAlign(
+ List<String> alignedRegions, boolean mapIsFromCdna, int strand)
+ throws IOException
+ {
+ if (strand == 0)
+ {
+ throw new IOException(
+ "Invalid strand for a codon mapping (cannot be 0)");
+ }
+ int regions = alignedRegions.size();
+ // arrays to hold [start, end] for each aligned region
+ int[] fromRanges = new int[regions * 2]; // from dna
+ int[] toRanges = new int[regions * 2]; // to protein
+ int fromRangesIndex = 0;
+ int toRangesIndex = 0;
+
+ for (String range : alignedRegions)
+ {
+ /*
+ * Align mapFromStart mapToStart mapFromCount
+ * e.g. if mapIsFromCdna
+ * Align 11270 143 120
+ * means:
+ * 120 bases from pos 11270 align to pos 143 in peptide
+ * if !mapIsFromCdna this would instead be
+ * Align 143 11270 40
+ */
+ String[] tokens = range.split(" ");
+ if (tokens.length != 3)
+ {
+ throw new IOException("Wrong number of fields for Align");
+ }
+ int fromStart = 0;
+ int toStart = 0;
+ int fromCount = 0;
+ try
+ {
+ fromStart = Integer.parseInt(tokens[0]);
+ toStart = Integer.parseInt(tokens[1]);
+ fromCount = Integer.parseInt(tokens[2]);
+ } catch (NumberFormatException nfe)
+ {
+ throw new IOException("Invalid number in Align field: "
+ + nfe.getMessage());
+ }
+
+ /*
+ * Jalview always models from dna to protein, so adjust values if the
+ * GFF mapping is from protein to dna
+ */
+ if (!mapIsFromCdna)
+ {
+ fromCount *= 3;
+ int temp = fromStart;
+ fromStart = toStart;
+ toStart = temp;
+ }
+ fromRanges[fromRangesIndex++] = fromStart;
+ fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1);
+
+ /*
+ * If a codon has an intron gap, there will be contiguous 'toRanges';
+ * this is handled for us by the MapList constructor.
+ * (It is not clear that exonerate ever generates this case)
+ */
+ toRanges[toRangesIndex++] = toStart;
+ toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3;
+ }
+
+ return new MapList(fromRanges, toRanges, 3, 1);
+ }
+
+ /**
+ * Parse a GFF format feature. This may include creating a 'dummy' sequence to
+ * hold the feature, or for its mapped sequence, or both, to be resolved
+ * either later in the GFF file (##FASTA section), or when the user loads
+ * additional sequences.
+ *
+ * @param gffColumns
+ * @param alignment
+ * @param relaxedIdMatching
+ * @param newseqs
+ * @return
+ */
+ protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment,
+ boolean relaxedIdMatching, List<SequenceI> newseqs)
+ {
+ /*
+ * GFF: seqid source type start end score strand phase [attributes]
+ */
+ if (gffColumns.length < 5)
+ {
+ System.err.println("Ignoring GFF feature line with too few columns ("
+ + gffColumns.length + ")");
+ return null;
+ }