*/
package jalview.analysis;
-import static jalview.io.gff.GffConstants.CLINICAL_SIGNIFICANCE;
-
+import jalview.commands.RemoveGapColCommand;
import jalview.datamodel.AlignedCodon;
import jalview.datamodel.AlignedCodonFrame;
import jalview.datamodel.AlignedCodonFrame.SequenceToSequenceMapping;
import jalview.datamodel.SequenceGroup;
import jalview.datamodel.SequenceI;
import jalview.datamodel.features.SequenceFeatures;
-import jalview.io.gff.Gff3Helper;
import jalview.io.gff.SequenceOntologyI;
import jalview.schemes.ResidueProperties;
import jalview.util.Comparison;
import jalview.util.IntRangeComparator;
import jalview.util.MapList;
import jalview.util.MappingUtils;
-import jalview.util.StringUtils;
-import java.io.UnsupportedEncodingException;
-import java.net.URLEncoder;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
*/
public class AlignmentUtils
{
-
private static final int CODON_LENGTH = 3;
private static final String SEQUENCE_VARIANT = "sequence_variant:";
- private static final String ID = "ID";
+ /*
+ * the 'id' attribute is provided for variant features fetched from
+ * Ensembl using its REST service with JSON format
+ */
+ public static final String VARIANT_ID = "id";
/**
* A data model to hold the 'normal' base value at a position, and an optional
SequenceI prot = mapping.findAlignedSequence(dnaSeq, protein);
if (prot != null)
{
- Mapping seqMap = mapping.getMappingForSequence(dnaSeq, false);
+ Mapping seqMap = mapping.getMappingForSequence(dnaSeq);
addCodonPositions(dnaSeq, prot, protein.getGapCharacter(), seqMap,
alignedCodons);
unmappedProtein.remove(prot);
cdsSeqs.add(cdsSeq);
- if (!dataset.getSequences().contains(cdsSeqDss))
- {
- // check if this sequence is a newly created one
- // so needs adding to the dataset
- dataset.addSequence(cdsSeqDss);
- }
-
/*
- * add a mapping from CDS to the (unchanged) mapped to range
+ * build the mapping from CDS to protein
*/
- List<int[]> cdsRange = Collections.singletonList(new int[] { 1,
- cdsSeq.getLength() });
+ List<int[]> cdsRange = Collections
+ .singletonList(new int[]
+ { cdsSeq.getStart(),
+ cdsSeq.getLength() + cdsSeq.getStart() - 1 });
MapList cdsToProteinMap = new MapList(cdsRange,
mapList.getToRanges(), mapList.getFromRatio(),
mapList.getToRatio());
- AlignedCodonFrame cdsToProteinMapping = new AlignedCodonFrame();
- cdsToProteinMapping.addMap(cdsSeqDss, proteinProduct,
- cdsToProteinMap);
- /*
- * guard against duplicating the mapping if repeating this action
- */
- if (!mappings.contains(cdsToProteinMapping))
+ if (!dataset.getSequences().contains(cdsSeqDss))
{
- mappings.add(cdsToProteinMapping);
+ /*
+ * if this sequence is a newly created one, add it to the dataset
+ * and made a CDS to protein mapping (if sequence already exists,
+ * CDS-to-protein mapping _is_ the transcript-to-protein mapping)
+ */
+ dataset.addSequence(cdsSeqDss);
+ AlignedCodonFrame cdsToProteinMapping = new AlignedCodonFrame();
+ cdsToProteinMapping.addMap(cdsSeqDss, proteinProduct,
+ cdsToProteinMap);
+
+ /*
+ * guard against duplicating the mapping if repeating this action
+ */
+ if (!mappings.contains(cdsToProteinMapping))
+ {
+ mappings.add(cdsToProteinMapping);
+ }
}
propagateDBRefsToCDS(cdsSeqDss, dnaSeq.getDatasetSequence(),
return;
}
- MapList newMap = targetToFrom.traverse(fromLoci.getMap());
+ MapList newMap = targetToFrom.traverse(fromLoci.getMapping());
if (newMap != null)
{
static SequenceI makeCdsSequence(SequenceI seq, Mapping mapping,
AlignmentI dataset)
{
- char[] seqChars = seq.getSequence();
- List<int[]> fromRanges = mapping.getMap().getFromRanges();
- int cdsWidth = MappingUtils.getLength(fromRanges);
- char[] newSeqChars = new char[cdsWidth];
+ /*
+ * construct CDS sequence name as "CDS|" with 'from id' held in the mapping
+ * if set (e.g. EMBL protein_id), else sequence name appended
+ */
+ String mapFromId = mapping.getMappedFromId();
+ final String seqId = "CDS|"
+ + (mapFromId != null ? mapFromId : seq.getName());
+
+ SequenceI newSeq = null;
- int newPos = 0;
- for (int[] range : fromRanges)
+ final MapList maplist = mapping.getMap();
+ if (maplist.isContiguous() && maplist.isFromForwardStrand())
{
- if (range[0] <= range[1])
- {
- // forward strand mapping - just copy the range
- int length = range[1] - range[0] + 1;
- System.arraycopy(seqChars, range[0] - 1, newSeqChars, newPos,
- length);
- newPos += length;
- }
- else
+ /*
+ * just a subsequence, keep same dataset sequence
+ */
+ int start = maplist.getFromLowest();
+ int end = maplist.getFromHighest();
+ newSeq = seq.getSubSequence(start - 1, end);
+ newSeq.setName(seqId);
+ }
+ else
+ {
+ /*
+ * construct by splicing mapped from ranges
+ */
+ char[] seqChars = seq.getSequence();
+ List<int[]> fromRanges = maplist.getFromRanges();
+ int cdsWidth = MappingUtils.getLength(fromRanges);
+ char[] newSeqChars = new char[cdsWidth];
+
+ int newPos = 0;
+ for (int[] range : fromRanges)
{
- // reverse strand mapping - copy and complement one by one
- for (int i = range[0]; i >= range[1]; i--)
+ if (range[0] <= range[1])
{
- newSeqChars[newPos++] = Dna.getComplement(seqChars[i - 1]);
+ // forward strand mapping - just copy the range
+ int length = range[1] - range[0] + 1;
+ System.arraycopy(seqChars, range[0] - 1, newSeqChars, newPos,
+ length);
+ newPos += length;
+ }
+ else
+ {
+ // reverse strand mapping - copy and complement one by one
+ for (int i = range[0]; i >= range[1]; i--)
+ {
+ newSeqChars[newPos++] = Dna.getComplement(seqChars[i - 1]);
+ }
}
}
+
+ newSeq = new Sequence(seqId, newSeqChars, 1, newPos);
}
- /*
- * assign 'from id' held in the mapping if set (e.g. EMBL protein_id),
- * else generate a sequence name
- */
- String mapFromId = mapping.getMappedFromId();
- String seqId = "CDS|" + (mapFromId != null ? mapFromId : seq.getName());
- SequenceI newSeq = new Sequence(seqId, newSeqChars, 1, newPos);
if (dataset != null)
{
SequenceI[] matches = dataset.findSequenceMatch(newSeq.getName());
{
copyTo = copyTo.getDatasetSequence();
}
+ if (fromSeq == copyTo || fromSeq.getDatasetSequence() == copyTo)
+ {
+ return 0; // shared dataset sequence
+ }
/*
* get features, optionally restricted by an ontology term
}
/**
- * Maps exon features from dna to protein, and computes variants in peptide
- * product generated by variants in dna, and adds them as sequence_variant
- * features on the protein sequence. Returns the number of variant features
- * added.
- *
- * @param dnaSeq
- * @param peptide
- * @param dnaToProtein
- */
- public static int computeProteinFeatures(SequenceI dnaSeq,
- SequenceI peptide, MapList dnaToProtein)
- {
- while (dnaSeq.getDatasetSequence() != null)
- {
- dnaSeq = dnaSeq.getDatasetSequence();
- }
- while (peptide.getDatasetSequence() != null)
- {
- peptide = peptide.getDatasetSequence();
- }
-
- transferFeatures(dnaSeq, peptide, dnaToProtein, SequenceOntologyI.EXON);
-
- /*
- * compute protein variants from dna variants and codon mappings;
- * NB - alternatively we could retrieve this using the REST service e.g.
- * http://rest.ensembl.org/overlap/translation
- * /ENSP00000288602?feature=transcript_variation;content-type=text/xml
- * which would be a bit slower but possibly more reliable
- */
-
- /*
- * build a map with codon variations for each potentially varying peptide
- */
- LinkedHashMap<Integer, List<DnaVariant>[]> variants = buildDnaVariantsMap(
- dnaSeq, dnaToProtein);
-
- /*
- * scan codon variations, compute peptide variants and add to peptide sequence
- */
- int count = 0;
- for (Entry<Integer, List<DnaVariant>[]> variant : variants.entrySet())
- {
- int peptidePos = variant.getKey();
- List<DnaVariant>[] codonVariants = variant.getValue();
- count += computePeptideVariants(peptide, peptidePos, codonVariants);
- }
-
- return count;
- }
-
- /**
- * Computes non-synonymous peptide variants from codon variants and adds them
- * as sequence_variant features on the protein sequence (one feature per
- * allele variant). Selected attributes (variant id, clinical significance)
- * are copied over to the new features.
- *
- * @param peptide
- * the protein sequence
- * @param peptidePos
- * the position to compute peptide variants for
- * @param codonVariants
- * a list of dna variants per codon position
- * @return the number of features added
- */
- static int computePeptideVariants(SequenceI peptide, int peptidePos,
- List<DnaVariant>[] codonVariants)
- {
- String residue = String
- .valueOf(peptide.getCharAt(peptidePos - peptide.getStart()));
- int count = 0;
- String base1 = codonVariants[0].get(0).base;
- String base2 = codonVariants[1].get(0).base;
- String base3 = codonVariants[2].get(0).base;
-
- /*
- * variants in first codon base
- */
- for (DnaVariant var : codonVariants[0])
- {
- if (var.variant != null)
- {
- String alleles = (String) var.variant.getValue(Gff3Helper.ALLELES);
- if (alleles != null)
- {
- for (String base : alleles.split(","))
- {
- if (!base1.equalsIgnoreCase(base))
- {
- String codon = base.toUpperCase() + base2.toLowerCase()
- + base3.toLowerCase();
- String canonical = base1.toUpperCase() + base2.toLowerCase()
- + base3.toLowerCase();
- if (addPeptideVariant(peptide, peptidePos, residue, var,
- codon, canonical))
- {
- count++;
- }
- }
- }
- }
- }
- }
-
- /*
- * variants in second codon base
- */
- for (DnaVariant var : codonVariants[1])
- {
- if (var.variant != null)
- {
- String alleles = (String) var.variant.getValue(Gff3Helper.ALLELES);
- if (alleles != null)
- {
- for (String base : alleles.split(","))
- {
- if (!base2.equalsIgnoreCase(base))
- {
- String codon = base1.toLowerCase() + base.toUpperCase()
- + base3.toLowerCase();
- String canonical = base1.toLowerCase() + base2.toUpperCase()
- + base3.toLowerCase();
- if (addPeptideVariant(peptide, peptidePos, residue, var,
- codon, canonical))
- {
- count++;
- }
- }
- }
- }
- }
- }
-
- /*
- * variants in third codon base
- */
- for (DnaVariant var : codonVariants[2])
- {
- if (var.variant != null)
- {
- String alleles = (String) var.variant.getValue(Gff3Helper.ALLELES);
- if (alleles != null)
- {
- for (String base : alleles.split(","))
- {
- if (!base3.equalsIgnoreCase(base))
- {
- String codon = base1.toLowerCase() + base2.toLowerCase()
- + base.toUpperCase();
- String canonical = base1.toLowerCase() + base2.toLowerCase()
- + base3.toUpperCase();
- if (addPeptideVariant(peptide, peptidePos, residue, var,
- codon, canonical))
- {
- count++;
- }
- }
- }
- }
- }
- }
-
- return count;
- }
-
- /**
- * Helper method that adds a peptide variant feature. ID and
- * clinical_significance attributes of the dna variant (if present) are copied
- * to the new feature.
- *
- * @param peptide
- * @param peptidePos
- * @param residue
- * @param var
- * @param codon
- * the variant codon e.g. aCg
- * @param canonical
- * the 'normal' codon e.g. aTg
- * @return true if a feature was added, else false
- */
- static boolean addPeptideVariant(SequenceI peptide, int peptidePos,
- String residue, DnaVariant var, String codon, String canonical)
- {
- /*
- * get peptide translation of codon e.g. GAT -> D
- * note that variants which are not single alleles,
- * e.g. multibase variants or HGMD_MUTATION etc
- * are currently ignored here
- */
- String trans = codon.contains("-") ? null
- : (codon.length() > CODON_LENGTH ? null
- : ResidueProperties.codonTranslate(codon));
- if (trans == null)
- {
- return false;
- }
- String desc = canonical + "/" + codon;
- String featureType = "";
- if (trans.equals(residue))
- {
- featureType = SequenceOntologyI.SYNONYMOUS_VARIANT;
- }
- else if (ResidueProperties.STOP.equals(trans))
- {
- featureType = SequenceOntologyI.STOP_GAINED;
- }
- else
- {
- String residue3Char = StringUtils
- .toSentenceCase(ResidueProperties.aa2Triplet.get(residue));
- String trans3Char = StringUtils
- .toSentenceCase(ResidueProperties.aa2Triplet.get(trans));
- desc = "p." + residue3Char + peptidePos + trans3Char;
- featureType = SequenceOntologyI.NONSYNONYMOUS_VARIANT;
- }
- SequenceFeature sf = new SequenceFeature(featureType, desc, peptidePos,
- peptidePos, var.getSource());
-
- StringBuilder attributes = new StringBuilder(32);
- String id = (String) var.variant.getValue(ID);
- if (id != null)
- {
- if (id.startsWith(SEQUENCE_VARIANT))
- {
- id = id.substring(SEQUENCE_VARIANT.length());
- }
- sf.setValue(ID, id);
- attributes.append(ID).append("=").append(id);
- // TODO handle other species variants JAL-2064
- StringBuilder link = new StringBuilder(32);
- try
- {
- link.append(desc).append(" ").append(id).append(
- "|http://www.ensembl.org/Homo_sapiens/Variation/Summary?v=")
- .append(URLEncoder.encode(id, "UTF-8"));
- sf.addLink(link.toString());
- } catch (UnsupportedEncodingException e)
- {
- // as if
- }
- }
- String clinSig = (String) var.variant.getValue(CLINICAL_SIGNIFICANCE);
- if (clinSig != null)
- {
- sf.setValue(CLINICAL_SIGNIFICANCE, clinSig);
- attributes.append(";").append(CLINICAL_SIGNIFICANCE).append("=")
- .append(clinSig);
- }
- peptide.addSequenceFeature(sf);
- if (attributes.length() > 0)
- {
- sf.setAttributes(attributes.toString());
- }
- return true;
- }
-
- /**
- * Builds a map whose key is position in the protein sequence, and value is a
- * list of the base and all variants for each corresponding codon position.
- * <p>
- * This depends on dna variants being held as a comma-separated list as
- * property "alleles" on variant features.
- *
- * @param dnaSeq
- * @param dnaToProtein
- * @return
- */
- @SuppressWarnings("unchecked")
- static LinkedHashMap<Integer, List<DnaVariant>[]> buildDnaVariantsMap(
- SequenceI dnaSeq, MapList dnaToProtein)
- {
- /*
- * map from peptide position to all variants of the codon which codes for it
- * LinkedHashMap ensures we keep the peptide features in sequence order
- */
- LinkedHashMap<Integer, List<DnaVariant>[]> variants = new LinkedHashMap<>();
-
- List<SequenceFeature> dnaFeatures = dnaSeq.getFeatures()
- .getFeaturesByOntology(SequenceOntologyI.SEQUENCE_VARIANT);
- if (dnaFeatures.isEmpty())
- {
- return variants;
- }
-
- int dnaStart = dnaSeq.getStart();
- int[] lastCodon = null;
- int lastPeptidePostion = 0;
-
- /*
- * build a map of codon variations for peptides
- */
- for (SequenceFeature sf : dnaFeatures)
- {
- int dnaCol = sf.getBegin();
- if (dnaCol != sf.getEnd())
- {
- // not handling multi-locus variant features
- continue;
- }
-
- /*
- * ignore variant if not a SNP
- */
- String alls = (String) sf.getValue(Gff3Helper.ALLELES);
- if (alls == null)
- {
- continue; // non-SNP VCF variant perhaps - can't process this
- }
-
- String[] alleles = alls.toUpperCase().split(",");
- boolean isSnp = true;
- for (String allele : alleles)
- {
- if (allele.trim().length() > 1)
- {
- isSnp = false;
- }
- }
- if (!isSnp)
- {
- continue;
- }
-
- int[] mapsTo = dnaToProtein.locateInTo(dnaCol, dnaCol);
- if (mapsTo == null)
- {
- // feature doesn't lie within coding region
- continue;
- }
- int peptidePosition = mapsTo[0];
- List<DnaVariant>[] codonVariants = variants.get(peptidePosition);
- if (codonVariants == null)
- {
- codonVariants = new ArrayList[CODON_LENGTH];
- codonVariants[0] = new ArrayList<>();
- codonVariants[1] = new ArrayList<>();
- codonVariants[2] = new ArrayList<>();
- variants.put(peptidePosition, codonVariants);
- }
-
- /*
- * get this peptide's codon positions e.g. [3, 4, 5] or [4, 7, 10]
- */
- int[] codon = peptidePosition == lastPeptidePostion ? lastCodon
- : MappingUtils.flattenRanges(dnaToProtein.locateInFrom(
- peptidePosition, peptidePosition));
- lastPeptidePostion = peptidePosition;
- lastCodon = codon;
-
- /*
- * save nucleotide (and any variant) for each codon position
- */
- for (int codonPos = 0; codonPos < CODON_LENGTH; codonPos++)
- {
- String nucleotide = String.valueOf(
- dnaSeq.getCharAt(codon[codonPos] - dnaStart)).toUpperCase();
- List<DnaVariant> codonVariant = codonVariants[codonPos];
- if (codon[codonPos] == dnaCol)
- {
- if (!codonVariant.isEmpty()
- && codonVariant.get(0).variant == null)
- {
- /*
- * already recorded base value, add this variant
- */
- codonVariant.get(0).variant = sf;
- }
- else
- {
- /*
- * add variant with base value
- */
- codonVariant.add(new DnaVariant(nucleotide, sf));
- }
- }
- else if (codonVariant.isEmpty())
- {
- /*
- * record (possibly non-varying) base value
- */
- codonVariant.add(new DnaVariant(nucleotide));
- }
- }
- }
- return variants;
- }
-
- /**
* Makes an alignment with a copy of the given sequences, adding in any
* non-redundant sequences which are mapped to by the cross-referenced
* sequences.
* true; else returns false
*
* @param unaligned
- * - sequences to be aligned based on aligned
+ * - sequences to be aligned based on aligned
* @param aligned
- * - 'guide' alignment containing sequences derived from same dataset
- * as unaligned
+ * - 'guide' alignment containing sequences derived from same
+ * dataset as unaligned
* @return
*/
static boolean alignAsSameSequences(AlignmentI unaligned,
}
/*
- * first pass - check whether all sequences to be aligned share a dataset
- * sequence with an aligned sequence
+ * first pass - check whether all sequences to be aligned share a
+ * dataset sequence with an aligned sequence; also note the leftmost
+ * ungapped column from which to copy
*/
+ int leftmost = Integer.MAX_VALUE;
for (SequenceI seq : unaligned.getSequences())
{
- if (!alignedDatasets.containsKey(seq.getDatasetSequence()))
+ final SequenceI ds = seq.getDatasetSequence();
+ if (!alignedDatasets.containsKey(ds))
{
return false;
}
+ SequenceI alignedSeq = alignedDatasets.get(ds)
+ .get(0);
+ int startCol = alignedSeq.findIndex(seq.getStart()); // 1..
+ leftmost = Math.min(leftmost, startCol);
}
/*
* heuristic rule: pair off sequences in order for the case where
* more than one shares the same dataset sequence
*/
+ final char gapCharacter = aligned.getGapCharacter();
for (SequenceI seq : unaligned.getSequences())
{
List<SequenceI> alignedSequences = alignedDatasets
.get(seq.getDatasetSequence());
- // TODO: getSequenceAsString() will be deprecated in the future
- // TODO: need to leave to SequenceI implementor to update gaps
- seq.setSequence(alignedSequences.get(0).getSequenceAsString());
+ if (alignedSequences.isEmpty())
+ {
+ /*
+ * defensive check - shouldn't happen! (JAL-3536)
+ */
+ continue;
+ }
+ SequenceI alignedSeq = alignedSequences.get(0);
+
+ /*
+ * gap fill for leading (5') UTR if any
+ */
+ // TODO this copies intron columns - wrong!
+ int startCol = alignedSeq.findIndex(seq.getStart()); // 1..
+ int endCol = alignedSeq.findIndex(seq.getEnd());
+ char[] seqchars = new char[endCol - leftmost + 1];
+ Arrays.fill(seqchars, gapCharacter);
+ char[] toCopy = alignedSeq.getSequence(startCol - 1, endCol);
+ System.arraycopy(toCopy, 0, seqchars, startCol - leftmost,
+ toCopy.length);
+ seq.setSequence(String.valueOf(seqchars));
if (alignedSequences.size() > 0)
{
// pop off aligned sequences (except the last one)
}
}
+ /*
+ * finally remove gapped columns (e.g. introns)
+ */
+ new RemoveGapColCommand("", unaligned.getSequencesArray(), 0,
+ unaligned.getWidth() - 1, unaligned);
+
return true;
}