/// <summary>
/// ExecuteSearch
/// </summary>
/// <param name="queryMolfile"></param>
/// <param name="databases"></param>
/// <param name="minimumSimilarity"></param>
/// <param name="maxHits"></param>
/// <returns></returns>
public List <StructSearchMatch> ExecuteSearch(
string queryMolfile,
string databases,
FingerprintType fpType,
double minimumSimilarity,
int maxHits)
{
int fragCnt;
IAtomContainer query = CdkMol.MolfileToAtomContainer(queryMolfile);
IAtomContainer largestFrag = CdkMol.GetLargestMoleculeFragment(query, out fragCnt);
GetCorpSim = Lex.Contains(databases, "corp");
GetChemblSim = Lex.Contains(databases, "chembl");
if (!GetCorpSim && !GetChemblSim)
{
throw new Exception("Invalid database: " + databases);
}
FingerprintType = fpType;
if (FingerprintType != FingerprintType.MACCS && FingerprintType != FingerprintType.Circular)
{
throw new Exception("Invalid FingerprintType: " + FingerprintType);
}
MinimumSimilarity = minimumSimilarity;
MaxHits = maxHits;
List <StructSearchMatch> matches = ExecuteSearch(largestFrag);
return(matches);
}
// # Get list of bits where at least one must be in result fp. Use least popular bits if possible.
// if db.mfp_counts:
// reqbits = [count['_id'] for count in db.mfp_counts.find({'_id': {'$in': qfp
// }
//}).sort('count', 1).limit(ncommon)]
// else:
// reqbits = qfp[:ncommon]
// results = []
// for fp in db.molecules.find({'mfp.bits': {'$in': reqbits}, 'mfp.count': {'$gte': qmin, '$lte': qmax}}):
// intersection = len(set(qfp) & set(fp['mfp']['bits']))
// pn = fp['mfp']['count']
// tanimoto = float(intersection) / (pn + qn - intersection)
// if tanimoto >= threshold:
// results.append((tanimoto, fp['chembl_id'], fp['smiles']))
// return results
/// <summary>
/// Test to calculate similarity between pairs of structures (Alt syntax: Call SimSearchMx.Test [cid1] [cid2]
/// Tautomers:
/// Isotopes:
/// Neg Counterion, (Cl-) with 2 quatternary N+ with H attached in main frag:
/// Neg Counterion, (I-) with quatternary N+ (no attached H) in main frag:
/// Pos Counterion (Li+) with O- in main frag:
/// Benzene, cyclohexane:
/// StereoIsomers:
/// StereoIsomers:
/// </summary>
/// <param name="args"></param>
/// <returns></returns>
public static string Test(string args)
{
string[] sa = args.Split(' ');
string cid1 = CompoundId.Normalize(sa[0]);
string cid2 = CompoundId.Normalize(sa[1]);
int smiLen = 40;
MoleculeMx s1 = MoleculeMx.SelectMoleculeForCid(cid1);
IAtomContainer mol = CdkMol.MolfileToAtomContainer(s1.GetMolfileString());
UniChemData ucd1 = UniChemUtil.BuildUniChemData(mol);
List <FingerprintMx> fps1 = UniChemDataToFingerprintMxList(ucd1);
MoleculeMx s2 = MoleculeMx.SelectMoleculeForCid(cid2);
IAtomContainer mol2 = CdkMol.MolfileToAtomContainer(s2.GetMolfileString());
UniChemData ucd2 = UniChemUtil.BuildUniChemData(mol2);
List <FingerprintMx> fps2 = UniChemDataToFingerprintMxList(ucd2);
string fps2Smiles = (sa[0] + " / " + sa[1]).PadRight(smiLen);
string scores = "";
for (int i1 = 0; i1 < fps1.Count; i1++)
{
FingerprintMx fp1 = fps1[i1];
for (int i2 = 0; i2 < fps2.Count; i2++)
{
FingerprintMx fp2 = fps2[i2];
if (i1 == 0) // build smiles headers of cid2 frags if first cid1 frag
{
fps2Smiles += "\t" + fp2.CanonSmiles.PadRight(smiLen);
}
if (i2 == 0)
{
scores += "\r\n" + fp1.CanonSmiles.PadRight(smiLen); // include smiles at start of each line
}
float simScore = CalculateFingerprintPairSimilarityScore(fp1, fp2);
scores += "\t" + string.Format("{0:0.00}", simScore).PadRight(smiLen);
}
}
scores = fps2Smiles + scores;
FileUtil.WriteAndOpenTextDocument("SimilarityScores", scores);
return("");
}
