/**
* Returns an AFPChain corresponding to the alignment between {@code structure1} and {@code
* structure2}, which is given by the gapped protein sequences {@code sequence1} and {@code
* sequence2}. The sequences need not correspond to the entire structures, since local alignment
* is performed to match the sequences to structures. Assumes that a residue is aligned if and
* only if it is given by an uppercase letter.
*
* @param sequence1 <em>Must</em> have {@link ProteinSequence#getUserCollection()} set to document
* upper- and lower-case as aligned and unaligned; see {@link
* #getAlignedUserCollection(String)}
* @throws StructureException
*/
public static AFPChain fastaToAfpChain(
ProteinSequence sequence1,
ProteinSequence sequence2,
Structure structure1,
Structure structure2)
throws StructureException {
if (structure1 == null || structure2 == null) {
throw new IllegalArgumentException("A structure is null");
}
if (sequence1 == null || sequence2 == null) {
throw new IllegalArgumentException("A sequence is null");
}
Atom[] ca1 = StructureTools.getRepresentativeAtomArray(structure1);
Atom[] ca2 = StructureTools.getRepresentativeAtomArray(structure2);
ResidueNumber[] residues1 =
StructureSequenceMatcher.matchSequenceToStructure(sequence1, structure1);
ResidueNumber[] residues2 =
StructureSequenceMatcher.matchSequenceToStructure(sequence2, structure2);
// nullify ResidueNumbers that have a lowercase sequence character
if (sequence1.getUserCollection() != null) {
CasePreservingProteinSequenceCreator.setLowercaseToNull(sequence1, residues1);
}
if (sequence2.getUserCollection() != null) {
CasePreservingProteinSequenceCreator.setLowercaseToNull(sequence2, residues2);
}
return buildAlignment(ca1, ca2, residues1, residues2);
}
/**
* Prints out the XML representation of an AFPChain from a file containing exactly two FASTA
* sequences.
*
* @param args A String array of fasta-file structure-1-name structure-2-name
* @throws StructureException
* @throws IOException
*/
public static void main(String[] args) throws StructureException, IOException {
if (args.length != 3) {
System.err.println(
"Usage: FastaAFPChainConverter fasta-file structure-1-name structure-2-name");
return;
}
File fasta = new File(args[0]);
Structure structure1 = StructureTools.getStructure(args[1]);
Structure structure2 = StructureTools.getStructure(args[2]);
if (structure1 == null)
throw new IllegalArgumentException("No structure for " + args[1] + " was found");
if (structure2 == null)
throw new IllegalArgumentException("No structure for " + args[2] + " was found");
AFPChain afpChain = fastaFileToAfpChain(fasta, structure1, structure2);
String xml = AFPChainXMLConverter.toXML(afpChain);
System.out.println(xml);
}
/**
* It replaces an optimal alignment of an AFPChain and calculates all the new alignment scores and
* variables.
*/
public static AFPChain replaceOptAln(int[][][] newAlgn, AFPChain afpChain, Atom[] ca1, Atom[] ca2)
throws StructureException {
// The order is the number of groups in the newAlgn
int order = newAlgn.length;
// Calculate the alignment length from all the subunits lengths
int[] optLens = new int[order];
for (int s = 0; s < order; s++) {
optLens[s] = newAlgn[s][0].length;
}
int optLength = 0;
for (int s = 0; s < order; s++) {
optLength += optLens[s];
}
// Create a copy of the original AFPChain and set everything needed for the structure update
AFPChain copyAFP = (AFPChain) afpChain.clone();
// Set the new parameters of the optimal alignment
copyAFP.setOptLength(optLength);
copyAFP.setOptLen(optLens);
copyAFP.setOptAln(newAlgn);
// Set the block information of the new alignment
copyAFP.setBlockNum(order);
copyAFP.setBlockSize(optLens);
copyAFP.setBlockResList(newAlgn);
copyAFP.setBlockResSize(optLens);
copyAFP.setBlockGap(calculateBlockGap(newAlgn));
// Recalculate properties: superposition, tm-score, etc
Atom[] ca2clone = StructureTools.cloneAtomArray(ca2); // don't modify ca2 positions
AlignmentTools.updateSuperposition(copyAFP, ca1, ca2clone);
// It re-does the sequence alignment strings from the OptAlgn information only
copyAFP.setAlnsymb(null);
AFPAlignmentDisplay.getAlign(copyAFP, ca1, ca2clone);
return copyAFP;
}
/**
* @param afpChain Input afpchain. UNMODIFIED
* @param ca1
* @param ca2
* @param optLens
* @param optAln
* @return A NEW AfpChain based off the input but with the optAln modified
* @throws StructureException if an error occured during superposition
*/
public static AFPChain replaceOptAln(
AFPChain afpChain, Atom[] ca1, Atom[] ca2, int blockNum, int[] optLens, int[][][] optAln)
throws StructureException {
int optLength = 0;
for (int blk = 0; blk < blockNum; blk++) {
optLength += optLens[blk];
}
// set everything
AFPChain refinedAFP = (AFPChain) afpChain.clone();
refinedAFP.setOptLength(optLength);
refinedAFP.setBlockSize(optLens);
refinedAFP.setOptLen(optLens);
refinedAFP.setOptAln(optAln);
refinedAFP.setBlockNum(blockNum);
// TODO recalculate properties: superposition, tm-score, etc
Atom[] ca2clone = StructureTools.cloneAtomArray(ca2); // don't modify ca2 positions
AlignmentTools.updateSuperposition(refinedAFP, ca1, ca2clone);
AFPAlignmentDisplay.getAlign(refinedAFP, ca1, ca2clone);
return refinedAFP;
}
/**
* Takes a structure and sequence corresponding to an alignment between a structure or sequence
* and itself (or even a structure with a sequence), where the result has a circular permutation
* site {@link cpSite} residues to the right.
*
* @param first The unpermuted sequence
* @param second The sequence permuted by cpSite
* @param cpSite The number of residues from the beginning of the sequence at which the circular
* permutation site occurs; can be positive or negative; values greater than the length of the
* sequence are acceptable
* @throws StructureException
*/
public static AFPChain cpFastaToAfpChain(
ProteinSequence first, ProteinSequence second, Structure structure, int cpSite)
throws StructureException {
if (structure == null) {
throw new IllegalArgumentException("The structure is null");
}
if (first == null) {
throw new IllegalArgumentException("The sequence is null");
}
// we need to find the ungapped CP site
int gappedCpShift = 0;
int ungappedCpShift = 0;
while (ungappedCpShift < Math.abs(cpSite)) {
char c;
try {
if (cpSite <= 0) {
c = second.getSequenceAsString().charAt(gappedCpShift);
} else {
c = second.getSequenceAsString().charAt(first.getLength() - 1 - gappedCpShift);
}
} catch (StringIndexOutOfBoundsException e) {
throw new IllegalArgumentException("CP site of " + cpSite + " is wrong");
}
if (c != '-') {
ungappedCpShift++;
}
gappedCpShift++;
}
Atom[] ca1 = StructureTools.getRepresentativeAtomArray(structure);
Atom[] ca2 =
StructureTools.getRepresentativeAtomArray(
structure); // can't use cloneCAArray because it doesn't set parent
// group.chain.structure
ProteinSequence antipermuted = null;
try {
antipermuted =
new ProteinSequence(
SequenceTools.permuteCyclic(second.getSequenceAsString(), gappedCpShift));
} catch (CompoundNotFoundException e) {
// this can't happen, the original sequence comes from a ProteinSequence
logger.error(
"Unexpected error while creating protein sequence: {}. This is most likely a bug.",
e.getMessage());
}
ResidueNumber[] residues = StructureSequenceMatcher.matchSequenceToStructure(first, structure);
ResidueNumber[] antipermutedResidues =
StructureSequenceMatcher.matchSequenceToStructure(antipermuted, structure);
ResidueNumber[] nonpermutedResidues = new ResidueNumber[antipermutedResidues.length];
SequenceTools.permuteCyclic(antipermutedResidues, nonpermutedResidues, -gappedCpShift);
// nullify ResidueNumbers that have a lowercase sequence character
if (first.getUserCollection() != null) {
CasePreservingProteinSequenceCreator.setLowercaseToNull(first, residues);
}
if (second.getUserCollection() != null) {
CasePreservingProteinSequenceCreator.setLowercaseToNull(second, nonpermutedResidues);
}
// for (int i = 0; i < residues.length; i++) {
// if (residues[i] == null) {
// System.out.print("=");
// } else {
// System.out.print(sequence.getSequenceAsString().charAt(i));
// }
// }
// System.out.println();
// for (int i = 0; i < residues.length; i++) {
// if (nonpermutedResidues[i] == null) {
// System.out.print("=");
// } else {
// System.out.print(second.getSequenceAsString().charAt(i));
// }
// }
// System.out.println();
return buildAlignment(ca1, ca2, residues, nonpermutedResidues);
}
/**
* Identify a set of modifications in a a list of chains.
*
* @param chains query {@link Chain}s.
* @param potentialModifications query {@link ProteinModification}s.
*/
public void identify(
final List<Chain> chains, final Set<ProteinModification> potentialModifications) {
if (chains == null) {
throw new IllegalArgumentException("Null structure.");
}
if (potentialModifications == null) {
throw new IllegalArgumentException("Null potentialModifications.");
}
reset();
if (potentialModifications.isEmpty()) {
return;
}
Map<String, Chain> mapChainIdChain = new HashMap<String, Chain>(chains.size());
residues = new ArrayList<Group>();
List<Group> ligands = new ArrayList<Group>();
Map<Component, Set<Group>> mapCompGroups = new HashMap<Component, Set<Group>>();
for (Chain chain : chains) {
mapChainIdChain.put(chain.getChainID(), chain);
List<Group> ress = StructureUtil.getAminoAcids(chain);
// List<Group> ligs = chain.getAtomLigands();
List<Group> ligs = StructureTools.filterLigands(chain.getAtomGroups());
residues.addAll(ress);
residues.removeAll(ligs);
ligands.addAll(ligs);
addModificationGroups(potentialModifications, ress, ligs, mapCompGroups);
}
if (residues.isEmpty()) {
String pdbId = "?";
if (chains.size() > 0) {
Structure struc = chains.get(0).getParent();
if (struc != null) pdbId = struc.getPDBCode();
}
logger.warn(
"No amino acids found for {}. Either you did not parse the PDB file with alignSEQRES records, or this record does not contain any amino acids.",
pdbId);
}
List<ModifiedCompound> modComps = new ArrayList<ModifiedCompound>();
for (ProteinModification mod : potentialModifications) {
ModificationCondition condition = mod.getCondition();
List<Component> components = condition.getComponents();
if (!mapCompGroups.keySet().containsAll(components)) {
// not all components exist for this mod.
continue;
}
int sizeComps = components.size();
if (sizeComps == 1) {
processCrosslink1(mapCompGroups, modComps, mod, components);
} else {
processMultiCrosslink(mapCompGroups, modComps, mod, condition);
}
}
if (recordAdditionalAttachments) {
// identify additional groups that are not directly attached to amino acids.
for (ModifiedCompound mc : modComps) {
identifyAdditionalAttachments(mc, ligands, mapChainIdChain);
}
}
mergeModComps(modComps);
identifiedModifiedCompounds.addAll(modComps);
// record unidentifiable linkage
if (recordUnidentifiableModifiedCompounds) {
recordUnidentifiableAtomLinkages(modComps, ligands);
recordUnidentifiableModifiedResidues(modComps);
}
}
