/**
* identify additional groups that are not directly attached to amino acids.
*
* @param mc {@link ModifiedCompound}.
* @param chain a {@link Chain}.
* @return a list of added groups.
*/
private void identifyAdditionalAttachments(
ModifiedCompound mc, List<Group> ligands, Map<String, Chain> mapChainIdChain) {
if (ligands.isEmpty()) {
return;
}
// TODO: should the additional groups only be allowed to the identified
// ligands or both amino acids and ligands? Currently only on ligands
// ligands to amino acid bonds for same modification of unknown category
// will be combined in mergeModComps()
// TODO: how about chain-chain links?
List<Group> identifiedGroups = new ArrayList<Group>();
for (StructureGroup num : mc.getGroups(false)) {
Group group;
try {
// String numIns = "" + num.getResidueNumber();
// if (num.getInsCode() != null) {
// numIns += num.getInsCode();
// }
ResidueNumber resNum = new ResidueNumber();
resNum.setChainId(num.getChainId());
resNum.setSeqNum(num.getResidueNumber());
resNum.setInsCode(num.getInsCode());
// group = chain.getGroupByPDB(numIns);
group = mapChainIdChain.get(num.getChainId()).getGroupByPDB(resNum);
} catch (StructureException e) {
logger.error("Exception: ", e);
// should not happen
continue;
}
identifiedGroups.add(group);
}
int start = 0;
int n = identifiedGroups.size();
while (n > start) {
for (Group group1 : ligands) {
for (int i = start; i < n; i++) {
Group group2 = identifiedGroups.get(i);
if (!identifiedGroups.contains(group1)) {
List<Atom[]> linkedAtoms =
StructureUtil.findAtomLinkages(group1, group2, false, bondLengthTolerance);
if (!linkedAtoms.isEmpty()) {
for (Atom[] atoms : linkedAtoms) {
mc.addAtomLinkage(
StructureUtil.getStructureAtomLinkage(atoms[0], false, atoms[1], false));
}
identifiedGroups.add(group1);
break;
}
}
}
}
start = n;
n = identifiedGroups.size();
}
}
/**
* Record unidentifiable atom linkages in a chain. Only linkages between two residues or one
* residue and one ligand will be recorded.
*/
private void recordUnidentifiableAtomLinkages(
List<ModifiedCompound> modComps, List<Group> ligands) {
// first put identified linkages in a map for fast query
Set<StructureAtomLinkage> identifiedLinkages = new HashSet<StructureAtomLinkage>();
for (ModifiedCompound mc : modComps) {
identifiedLinkages.addAll(mc.getAtomLinkages());
}
// record
// cross link
int nRes = residues.size();
for (int i = 0; i < nRes - 1; i++) {
Group group1 = residues.get(i);
for (int j = i + 1; j < nRes; j++) {
Group group2 = residues.get(j);
List<Atom[]> linkages =
StructureUtil.findAtomLinkages(group1, group2, true, bondLengthTolerance);
for (Atom[] atoms : linkages) {
StructureAtomLinkage link =
StructureUtil.getStructureAtomLinkage(atoms[0], true, atoms[1], true);
unidentifiableAtomLinkages.add(link);
}
}
}
// attachment
int nLig = ligands.size();
for (int i = 0; i < nRes; i++) {
Group group1 = residues.get(i);
for (int j = 0; j < nLig; j++) {
Group group2 = ligands.get(j);
if (group1.equals(group2)) { // overlap between residues and ligands
continue;
}
List<Atom[]> linkages =
StructureUtil.findAtomLinkages(group1, group2, false, bondLengthTolerance);
for (Atom[] atoms : linkages) {
StructureAtomLinkage link =
StructureUtil.getStructureAtomLinkage(atoms[0], true, atoms[1], false);
unidentifiableAtomLinkages.add(link);
}
}
}
}
/**
* Assembly the matched linkages.
*
* @param matchedAtomsOfLinkages
* @param mod
* @param condition
* @param ret ModifiedCompound will be stored here.
*/
private void assembleLinkages(
List<List<Atom[]>> matchedAtomsOfLinkages,
ProteinModification mod,
List<ModifiedCompound> ret) {
ModificationCondition condition = mod.getCondition();
List<ModificationLinkage> modLinks = condition.getLinkages();
int nLink = matchedAtomsOfLinkages.size();
int[] indices = new int[nLink];
Set<ModifiedCompound> identifiedCompounds = new HashSet<ModifiedCompound>();
while (indices[0] < matchedAtomsOfLinkages.get(0).size()) {
List<Atom[]> atomLinkages = new ArrayList<Atom[]>(nLink);
for (int iLink = 0; iLink < nLink; iLink++) {
Atom[] atoms = matchedAtomsOfLinkages.get(iLink).get(indices[iLink]);
atomLinkages.add(atoms);
}
if (matchLinkages(modLinks, atomLinkages)) {
// matched
int n = atomLinkages.size();
List<StructureAtomLinkage> linkages = new ArrayList<StructureAtomLinkage>(n);
for (int i = 0; i < n; i++) {
Atom[] linkage = atomLinkages.get(i);
StructureAtomLinkage link =
StructureUtil.getStructureAtomLinkage(
linkage[0], residues.contains(linkage[0].getGroup()),
linkage[1], residues.contains(linkage[1].getGroup()));
linkages.add(link);
}
ModifiedCompound mc = new ModifiedCompoundImpl(mod, linkages);
if (!identifiedCompounds.contains(mc)) {
ret.add(mc);
identifiedCompounds.add(mc);
}
}
// indices++ (e.g. [0,0,1]=>[0,0,2]=>[1,2,0])
int i = nLink - 1;
while (i >= 0) {
if (i == 0 || indices[i] < matchedAtomsOfLinkages.get(i).size() - 1) {
indices[i]++;
break;
} else {
indices[i] = 0;
i--;
}
}
}
}
private void recordUnidentifiableModifiedResidues(List<ModifiedCompound> modComps) {
Set<StructureGroup> identifiedComps = new HashSet<StructureGroup>();
for (ModifiedCompound mc : modComps) {
identifiedComps.addAll(mc.getGroups(true));
}
// TODO: use the ModifiedAminoAcid after Andreas add that.
for (Group group : residues) {
if (group.getType().equals(GroupType.HETATM)) {
StructureGroup strucGroup = StructureUtil.getStructureGroup(group, true);
if (!identifiedComps.contains(strucGroup)) {
unidentifiableModifiedResidues.add(strucGroup);
}
}
}
}
private void processCrosslink1(
Map<Component, Set<Group>> mapCompGroups,
List<ModifiedCompound> modComps,
ProteinModification mod,
List<Component> components) {
// modified residue
// TODO: is this the correct logic for CROSS_LINK_1?
Set<Group> modifiedResidues = mapCompGroups.get(components.get(0));
if (modifiedResidues != null) {
for (Group residue : modifiedResidues) {
StructureGroup strucGroup = StructureUtil.getStructureGroup(residue, true);
ModifiedCompound modRes = new ModifiedCompoundImpl(mod, strucGroup);
modComps.add(modRes);
}
}
}
/** Get matched atoms for all linkages. */
private List<List<Atom[]>> getMatchedAtomsOfLinkages(
ModificationCondition condition, Map<Component, Set<Group>> mapCompGroups) {
List<ModificationLinkage> linkages = condition.getLinkages();
int nLink = linkages.size();
List<List<Atom[]>> matchedAtomsOfLinkages = new ArrayList<List<Atom[]>>(nLink);
for (int iLink = 0; iLink < nLink; iLink++) {
ModificationLinkage linkage = linkages.get(iLink);
Component comp1 = linkage.getComponent1();
Component comp2 = linkage.getComponent2();
// boolean isAA1 = comp1.;
// boolean isAA2 = comp2.getType()==true;
Set<Group> groups1 = mapCompGroups.get(comp1);
Set<Group> groups2 = mapCompGroups.get(comp2);
List<Atom[]> list = new ArrayList<Atom[]>();
List<String> potentialNamesOfAtomOnGroup1 = linkage.getPDBNameOfPotentialAtomsOnComponent1();
for (String name : potentialNamesOfAtomOnGroup1) {
if (name.equals("*")) {
// wildcard
potentialNamesOfAtomOnGroup1 = null; // search all atoms
break;
}
}
List<String> potentialNamesOfAtomOnGroup2 = linkage.getPDBNameOfPotentialAtomsOnComponent2();
for (String name : potentialNamesOfAtomOnGroup2) {
if (name.equals("*")) {
// wildcard
potentialNamesOfAtomOnGroup2 = null; // search all atoms
break;
}
}
for (Group g1 : groups1) {
for (Group g2 : groups2) {
if (g1.equals(g2)) {
continue;
}
// only for wildcard match of two residues
boolean ignoreNCLinkage =
potentialNamesOfAtomOnGroup1 == null
&& potentialNamesOfAtomOnGroup2 == null
&& residues.contains(g1)
&& residues.contains(g2);
Atom[] atoms =
StructureUtil.findNearestAtomLinkage(
g1,
g2,
potentialNamesOfAtomOnGroup1,
potentialNamesOfAtomOnGroup2,
ignoreNCLinkage,
bondLengthTolerance);
if (atoms != null) {
list.add(atoms);
}
}
}
if (list.isEmpty()) {
// broken linkage
break;
}
matchedAtomsOfLinkages.add(list);
}
return matchedAtomsOfLinkages;
}
/**
* 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);
}
}