import jalview.datamodel.Mapping;
import jalview.datamodel.SearchResults;
import jalview.datamodel.Sequence;
+import jalview.datamodel.SequenceFeature;
import jalview.datamodel.SequenceGroup;
import jalview.datamodel.SequenceI;
import jalview.schemes.ResidueProperties;
import java.util.HashSet;
import java.util.Iterator;
import java.util.LinkedHashMap;
-import java.util.LinkedHashSet;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
* @param cdnaAlignment
* @return
*/
- public static boolean mapProteinToCdna(final AlignmentI proteinAlignment,
- final AlignmentI cdnaAlignment)
+ public static boolean mapProteinAlignmentToCdna(
+ final AlignmentI proteinAlignment, final AlignmentI cdnaAlignment)
{
if (proteinAlignment == null || cdnaAlignment == null)
{
final AlignmentI cdnaAlignment, Set<SequenceI> mappedDna,
Set<SequenceI> mappedProtein, boolean xrefsOnly)
{
- boolean mappingPerformed = false;
+ boolean mappingExistsOrAdded = false;
List<SequenceI> thisSeqs = proteinAlignment.getSequences();
for (SequenceI aaSeq : thisSeqs)
{
{
continue;
}
- if (!mappingExists(proteinAlignment.getCodonFrames(),
+ if (mappingExists(proteinAlignment.getCodonFrames(),
aaSeq.getDatasetSequence(), cdnaSeq.getDatasetSequence()))
{
- MapList map = mapProteinToCdna(aaSeq, cdnaSeq);
+ mappingExistsOrAdded = true;
+ }
+ else
+ {
+ MapList map = mapProteinSequenceToCdna(aaSeq, cdnaSeq);
if (map != null)
{
acf.addMap(cdnaSeq, aaSeq, map);
- mappingPerformed = true;
+ mappingExistsOrAdded = true;
proteinMapped = true;
mappedDna.add(cdnaSeq);
mappedProtein.add(aaSeq);
proteinAlignment.addCodonFrame(acf);
}
}
- return mappingPerformed;
+ return mappingExistsOrAdded;
}
/**
* Answers true if the mappings include one between the given (dataset)
* sequences.
*/
- public static boolean mappingExists(Set<AlignedCodonFrame> set,
+ public static boolean mappingExists(List<AlignedCodonFrame> mappings,
SequenceI aaSeq, SequenceI cdnaSeq)
{
- if (set != null)
+ if (mappings != null)
{
- for (AlignedCodonFrame acf : set)
+ for (AlignedCodonFrame acf : mappings)
{
if (cdnaSeq == acf.getDnaForAaSeq(aaSeq))
{
* @param cdnaSeq
* @return
*/
- public static MapList mapProteinToCdna(SequenceI proteinSeq,
+ public static MapList mapProteinSequenceToCdna(SequenceI proteinSeq,
SequenceI cdnaSeq)
{
/*
*/
final int mappedLength = 3 * aaSeqChars.length;
int cdnaLength = cdnaSeqChars.length;
- int cdnaStart = 1;
- int cdnaEnd = cdnaLength;
- final int proteinStart = 1;
- final int proteinEnd = aaSeqChars.length;
+ int cdnaStart = cdnaSeq.getStart();
+ int cdnaEnd = cdnaSeq.getEnd();
+ final int proteinStart = proteinSeq.getStart();
+ final int proteinEnd = proteinSeq.getEnd();
/*
* If lengths don't match, try ignoring stop codon.
/*
* If lengths still don't match, try ignoring start codon.
*/
+ int startOffset = 0;
if (cdnaLength != mappedLength
&& cdnaLength > 2
&& String.valueOf(cdnaSeqChars, 0, 3).toUpperCase()
.equals(ResidueProperties.START))
{
+ startOffset += 3;
cdnaStart += 3;
cdnaLength -= 3;
}
{
return null;
}
- if (!translatesAs(cdnaSeqChars, cdnaStart - 1, aaSeqChars))
+ if (!translatesAs(cdnaSeqChars, startOffset, aaSeqChars))
{
return null;
}
/*
* Locate the aligned source sequence whose dataset sequence is mapped. We
- * just take the first match here (as we can't align cDNA like more than one
- * protein sequence).
+ * just take the first match here (as we can't align like more than one
+ * sequence).
*/
SequenceI alignFrom = null;
AlignedCodonFrame mapping = null;
/**
* Align sequence 'alignTo' the same way as 'alignFrom', using the mapping to
* match residues and codons. Flags control whether existing gaps in unmapped
- * (intron) and mapped (exon) regions are preserved or not. Gaps linking intro
- * and exon are only retained if both flags are set.
+ * (intron) and mapped (exon) regions are preserved or not. Gaps between
+ * intron and exon are only retained if both flags are set.
*
* @param alignTo
* @param alignFrom
boolean preserveUnmappedGaps)
{
// TODO generalise to work for Protein-Protein, dna-dna, dna-protein
- final char[] thisSeq = alignTo.getSequence();
- final char[] thatAligned = alignFrom.getSequence();
- StringBuilder thisAligned = new StringBuilder(2 * thisSeq.length);
// aligned and dataset sequence positions, all base zero
int thisSeqPos = 0;
char myGapChar = myGap.charAt(0);
int ratio = myGap.length();
- /*
- * Traverse the aligned protein sequence.
- */
+ int fromOffset = alignFrom.getStart() - 1;
+ int toOffset = alignTo.getStart() - 1;
int sourceGapMappedLength = 0;
boolean inExon = false;
+ final char[] thisSeq = alignTo.getSequence();
+ final char[] thatAligned = alignFrom.getSequence();
+ StringBuilder thisAligned = new StringBuilder(2 * thisSeq.length);
+
+ /*
+ * Traverse the 'model' aligned sequence
+ */
for (char sourceChar : thatAligned)
{
if (sourceChar == sourceGap)
}
/*
- * Found a residue. Locate its mapped codon (start) position.
+ * Found a non-gap character. Locate its mapped region if any.
*/
sourceDsPos++;
// Note mapping positions are base 1, our sequence positions base 0
int[] mappedPos = mapping.getMappedRegion(alignTo, alignFrom,
- sourceDsPos);
+ sourceDsPos + fromOffset);
if (mappedPos == null)
{
/*
- * Abort realignment if unmapped protein. Or could ignore it??
+ * unmapped position; treat like a gap
*/
- System.err.println("Can't align: no codon mapping to residue "
- + sourceDsPos + "(" + sourceChar + ")");
- return;
+ sourceGapMappedLength += ratio;
+ // System.err.println("Can't align: no codon mapping to residue "
+ // + sourceDsPos + "(" + sourceChar + ")");
+ // return;
+ continue;
}
int mappedCodonStart = mappedPos[0]; // position (1...) of codon start
* But then 'align dna as protein' doesn't make much sense otherwise.
*/
int intronLength = 0;
- while (basesWritten < mappedCodonEnd && thisSeqPos < thisSeq.length)
+ while (basesWritten + toOffset < mappedCodonEnd
+ && thisSeqPos < thisSeq.length)
{
final char c = thisSeq[thisSeqPos++];
if (c != myGapChar)
{
basesWritten++;
-
- if (basesWritten < mappedCodonStart)
+ int sourcePosition = basesWritten + toOffset;
+ if (sourcePosition < mappedCodonStart)
{
/*
* Found an unmapped (intron) base. First add in any preceding gaps
}
else
{
- final boolean startOfCodon = basesWritten == mappedCodonStart;
+ final boolean startOfCodon = sourcePosition == mappedCodonStart;
int gapsToAdd = calculateGapsToInsert(preserveMappedGaps,
preserveUnmappedGaps, sourceGapMappedLength, inExon,
trailingCopiedGap.length(), intronLength, startOfCodon);
}
/*
- * At end of protein sequence. Copy any remaining dna sequence, optionally
- * including (intron) gaps. We do not copy trailing gaps in protein.
+ * At end of model aligned sequence. Copy any remaining target sequence, optionally
+ * including (intron) gaps.
*/
while (thisSeqPos < thisSeq.length)
{
{
thisAligned.append(c);
}
+ sourceGapMappedLength--;
+ }
+
+ /*
+ * finally add gaps to pad for any trailing source gaps or
+ * unmapped characters
+ */
+ if (preserveUnmappedGaps)
+ {
+ while (sourceGapMappedLength > 0)
+ {
+ thisAligned.append(myGapChar);
+ sourceGapMappedLength--;
+ }
}
/*
List<SequenceI> unmappedProtein = new ArrayList<SequenceI>();
unmappedProtein.addAll(protein.getSequences());
- Set<AlignedCodonFrame> mappings = protein.getCodonFrames();
+ List<AlignedCodonFrame> mappings = protein.getCodonFrames();
/*
* Map will hold, for each aligned codon position e.g. [3, 5, 6], a map of
}
AlignmentI dna = al1.isNucleotide() ? al1 : al2;
AlignmentI protein = dna == al1 ? al2 : al1;
- Set<AlignedCodonFrame> mappings = protein.getCodonFrames();
+ List<AlignedCodonFrame> mappings = protein.getCodonFrames();
for (SequenceI dnaSeq : dna.getSequences())
{
for (SequenceI proteinSeq : protein.getSequences())
* @return
*/
protected static boolean isMappable(SequenceI dnaSeq,
- SequenceI proteinSeq, Set<AlignedCodonFrame> mappings)
+ SequenceI proteinSeq, List<AlignedCodonFrame> mappings)
{
if (dnaSeq == null || proteinSeq == null)
{
SequenceI proteinDs = proteinSeq.getDatasetSequence() == null ? proteinSeq
: proteinSeq.getDatasetSequence();
- /*
- * Already mapped?
- */
for (AlignedCodonFrame mapping : mappings)
{
if (proteinDs == mapping.getAaForDnaSeq(dnaDs))
{
+ /*
+ * already mapped
+ */
return true;
}
}
* Just try to make a mapping (it is not yet stored), test whether
* successful.
*/
- return mapProteinToCdna(proteinDs, dnaDs) != null;
+ return mapProteinSequenceToCdna(proteinDs, dnaDs) != null;
}
/**
return false;
}
String name = seq2.getName();
- final DBRefEntry[] xrefs = seq1.getDBRef();
+ final DBRefEntry[] xrefs = seq1.getDBRefs();
if (xrefs != null)
{
for (DBRefEntry xref : xrefs)
}
/**
- * Constructs an alignment consisting of the mapped exon regions in the given
+ * Constructs an alignment consisting of the mapped cds regions in the given
* nucleotide sequences, and updates mappings to match.
*
* @param dna
* aligned dna sequences
* @param mappings
* from dna to protein; these are replaced with new mappings
- * @return an alignment whose sequences are the exon-only parts of the dna
- * sequences (or null if no exons are found)
+ * @return an alignment whose sequences are the cds-only parts of the dna
+ * sequences (or null if no cds are found)
*/
- public static AlignmentI makeExonAlignment(SequenceI[] dna,
- Set<AlignedCodonFrame> mappings)
+ public static AlignmentI makeCdsAlignment(SequenceI[] dna,
+ List<AlignedCodonFrame> mappings)
{
- Set<AlignedCodonFrame> newMappings = new LinkedHashSet<AlignedCodonFrame>();
- List<SequenceI> exonSequences = new ArrayList<SequenceI>();
+ List<AlignedCodonFrame> newMappings = new ArrayList<AlignedCodonFrame>();
+ List<SequenceI> cdsSequences = new ArrayList<SequenceI>();
for (SequenceI dnaSeq : dna)
{
for (AlignedCodonFrame acf : seqMappings)
{
AlignedCodonFrame newMapping = new AlignedCodonFrame();
- final List<SequenceI> mappedExons = makeExonSequences(ds, acf,
+ final List<SequenceI> mappedCds = makeCdsSequences(ds, acf,
newMapping);
- if (!mappedExons.isEmpty())
+ if (!mappedCds.isEmpty())
{
- exonSequences.addAll(mappedExons);
+ cdsSequences.addAll(mappedCds);
newMappings.add(newMapping);
}
}
}
AlignmentI al = new Alignment(
- exonSequences.toArray(new SequenceI[exonSequences.size()]));
+ cdsSequences.toArray(new SequenceI[cdsSequences.size()]));
al.setDataset(null);
/*
}
/**
- * Helper method to make exon-only sequences and populate their mappings to
+ * Helper method to make cds-only sequences and populate their mappings to
* protein products
* <p>
* For example, if ggCCaTTcGAg has mappings [3, 4, 6, 7, 9, 10] to protein
* then generate a sequence CCTTGA with mapping [1, 6] to the same protein
* residues
* <p>
- * Typically eukaryotic dna will include exons encoding for a single peptide
+ * Typically eukaryotic dna will include cds encoding for a single peptide
* sequence i.e. return a single result. Bacterial dna may have overlapping
- * exon mappings coding for multiple peptides so return multiple results
+ * cds mappings coding for multiple peptides so return multiple results
* (example EMBL KF591215).
*
* @param dnaSeq
* a dna dataset sequence
* @param mapping
* containing one or more mappings of the sequence to protein
- * @param newMapping
- * the new mapping to populate, from the exon-only sequences to their
+ * @param newMappings
+ * the new mapping to populate, from the cds-only sequences to their
* mapped protein sequences
* @return
*/
- protected static List<SequenceI> makeExonSequences(SequenceI dnaSeq,
- AlignedCodonFrame mapping, AlignedCodonFrame newMapping)
+ protected static List<SequenceI> makeCdsSequences(SequenceI dnaSeq,
+ AlignedCodonFrame mapping, AlignedCodonFrame newMappings)
{
- List<SequenceI> exonSequences = new ArrayList<SequenceI>();
+ List<SequenceI> cdsSequences = new ArrayList<SequenceI>();
List<Mapping> seqMappings = mapping.getMappingsForSequence(dnaSeq);
- final char[] dna = dnaSeq.getSequence();
+
for (Mapping seqMapping : seqMappings)
{
- StringBuilder newSequence = new StringBuilder(dnaSeq.getLength());
+ SequenceI cds = makeCdsSequence(dnaSeq, seqMapping);
+ cdsSequences.add(cds);
/*
- * Get the codon regions as { [2, 5], [7, 12], [14, 14] etc }
+ * add new mappings, from dna to cds, and from cds to peptide
*/
- final List<int[]> dnaExonRanges = seqMapping.getMap().getFromRanges();
- for (int[] range : dnaExonRanges)
+ MapList dnaToCds = addCdsMappings(dnaSeq, cds, seqMapping,
+ newMappings);
+
+ /*
+ * transfer any features on dna that overlap the CDS
+ */
+ transferFeatures(dnaSeq, cds, dnaToCds, "CDS" /* SequenceOntology.CDS */);
+ }
+ return cdsSequences;
+ }
+
+ /**
+ * Transfers any co-located features on 'fromSeq' to 'toSeq', adjusting the
+ * feature start/end ranges, optionally omitting specified feature types.
+ *
+ * @param fromSeq
+ * @param toSeq
+ * @param mapping
+ * the mapping from 'fromSeq' to 'toSeq'
+ * @param omitting
+ */
+ protected static void transferFeatures(SequenceI fromSeq,
+ SequenceI toSeq, MapList mapping, String... omitting)
+ {
+ SequenceI copyTo = toSeq;
+ while (copyTo.getDatasetSequence() != null)
+ {
+ copyTo = copyTo.getDatasetSequence();
+ }
+
+ SequenceFeature[] sfs = fromSeq.getSequenceFeatures();
+ if (sfs != null)
+ {
+ for (SequenceFeature sf : sfs)
{
- for (int pos = range[0]; pos <= range[1]; pos++)
+ String type = sf.getType();
+ boolean omit = false;
+ for (String toOmit : omitting)
{
- newSequence.append(dna[pos - 1]);
+ if (type.equals(toOmit))
+ {
+ omit = true;
+ }
+ }
+ if (omit)
+ {
+ continue;
+ }
+
+ /*
+ * locate the mapped range - null if either start or end is
+ * not mapped (no partial overlaps are calculated)
+ */
+ int[] mappedTo = mapping.locateInTo(sf.getBegin(), sf.getEnd());
+ if (mappedTo != null)
+ {
+ SequenceFeature copy = new SequenceFeature(sf);
+ copy.setBegin(Math.min(mappedTo[0], mappedTo[1]));
+ copy.setEnd(Math.max(mappedTo[0], mappedTo[1]));
+ copyTo.addSequenceFeature(copy);
}
}
+ }
+ }
+
+ /**
+ * Creates and adds mappings
+ * <ul>
+ * <li>from cds to peptide</li>
+ * <li>from dna to cds</li>
+ * </ul>
+ * and returns the dna-to-cds mapping
+ *
+ * @param dnaSeq
+ * @param cdsSeq
+ * @param dnaMapping
+ * @param newMappings
+ * @return
+ */
+ protected static MapList addCdsMappings(SequenceI dnaSeq,
+ SequenceI cdsSeq,
+ Mapping dnaMapping, AlignedCodonFrame newMappings)
+ {
+ cdsSeq.createDatasetSequence();
- SequenceI exon = new Sequence(dnaSeq.getName(),
- newSequence.toString());
+ /*
+ * CDS to peptide is just a contiguous 3:1 mapping, with
+ * the peptide ranges taken unchanged from the dna mapping
+ */
+ List<int[]> cdsRanges = new ArrayList<int[]>();
+ cdsRanges.add(new int[] { 1, cdsSeq.getLength() });
+ MapList cdsToPeptide = new MapList(cdsRanges, dnaMapping.getMap()
+ .getToRanges(), 3, 1);
+ newMappings.addMap(cdsSeq.getDatasetSequence(), dnaMapping.getTo(),
+ cdsToPeptide);
- /*
- * Locate any xrefs to CDS database on the protein product and attach to
- * the CDS sequence. Also add as a sub-token of the sequence name.
- */
- // default to "CDS" if we can't locate an actual gene id
- String cdsAccId = FeatureProperties
- .getCodingFeature(DBRefSource.EMBL);
- DBRefEntry[] cdsRefs = DBRefUtils.selectRefs(seqMapping.getTo()
- .getDBRef(), DBRefSource.CODINGDBS);
- if (cdsRefs != null)
+ /*
+ * dna 'from' ranges map 1:1 to the contiguous extracted CDS
+ */
+ MapList dnaToCds = new MapList(
+ dnaMapping.getMap().getFromRanges(), cdsRanges, 1, 1);
+ newMappings.addMap(dnaSeq, cdsSeq.getDatasetSequence(), dnaToCds);
+ return dnaToCds;
+ }
+
+ /**
+ * Makes and returns a CDS-only sequence, where the CDS regions are identified
+ * as the 'from' ranges of the mapping on the dna.
+ *
+ * @param dnaSeq
+ * nucleotide sequence
+ * @param seqMapping
+ * mappings from CDS regions of nucleotide
+ * @return
+ */
+ protected static SequenceI makeCdsSequence(SequenceI dnaSeq,
+ Mapping seqMapping)
+ {
+ StringBuilder newSequence = new StringBuilder(dnaSeq.getLength());
+ final char[] dna = dnaSeq.getSequence();
+ int offset = dnaSeq.getStart() - 1;
+
+ /*
+ * Get the codon regions as { [2, 5], [7, 12], [14, 14] etc }
+ */
+ final List<int[]> dnaCdsRanges = seqMapping.getMap().getFromRanges();
+ for (int[] range : dnaCdsRanges)
+ {
+ // TODO handle reverse mapping as well (range[1] < range[0])
+ for (int pos = range[0]; pos <= range[1]; pos++)
{
- for (DBRefEntry cdsRef : cdsRefs)
- {
- exon.addDBRef(new DBRefEntry(cdsRef));
- cdsAccId = cdsRef.getAccessionId();
- }
+ newSequence.append(dna[pos - offset - 1]);
}
- exon.setName(exon.getName() + "|" + cdsAccId);
- exon.createDatasetSequence();
+ }
- /*
- * Build new mappings - from the same protein regions, but now to
- * contiguous exons
- */
- List<int[]> exonRange = new ArrayList<int[]>();
- exonRange.add(new int[] { 1, newSequence.length() });
- MapList map = new MapList(exonRange, seqMapping.getMap()
- .getToRanges(), 3, 1);
- newMapping.addMap(exon.getDatasetSequence(), seqMapping.getTo(), map);
- MapList cdsToDnaMap = new MapList(dnaExonRanges, exonRange, 1, 1);
- newMapping.addMap(dnaSeq, exon.getDatasetSequence(), cdsToDnaMap);
-
- exonSequences.add(exon);
+ SequenceI cds = new Sequence(dnaSeq.getName(),
+ newSequence.toString());
+
+ transferDbRefs(seqMapping.getTo(), cds);
+
+ return cds;
+ }
+
+ /**
+ * Locate any xrefs to CDS databases on the protein product and attach to the
+ * CDS sequence. Also add as a sub-token of the sequence name.
+ *
+ * @param from
+ * @param to
+ */
+ protected static void transferDbRefs(SequenceI from, SequenceI to)
+ {
+ String cdsAccId = FeatureProperties.getCodingFeature(DBRefSource.EMBL);
+ DBRefEntry[] cdsRefs = DBRefUtils.selectRefs(from.getDBRefs(),
+ DBRefSource.CODINGDBS);
+ if (cdsRefs != null)
+ {
+ for (DBRefEntry cdsRef : cdsRefs)
+ {
+ to.addDBRef(new DBRefEntry(cdsRef));
+ cdsAccId = cdsRef.getAccessionId();
+ }
+ }
+ if (!to.getName().contains(cdsAccId))
+ {
+ to.setName(to.getName() + "|" + cdsAccId);
}
- return exonSequences;
}
}
git://source.jalview.org / jalview.git / blobdiff