public void TestClusterWithCross()
{
// Create a list of Mum classes.
List <MaxUniqueMatch> matches = new List <MaxUniqueMatch>();
MaxUniqueMatch match = null;
match = new MaxUniqueMatch();
match.FirstSequenceStart = 0;
match.Length = 4;
match.SecondSequenceStart = 4;
matches.Add(match);
match = new MaxUniqueMatch();
match.FirstSequenceStart = 4;
match.Length = 3;
match.SecondSequenceStart = 0;
matches.Add(match);
match = new MaxUniqueMatch();
match.FirstSequenceStart = 10;
match.Length = 3;
match.SecondSequenceStart = 10;
matches.Add(match);
IClusterBuilder clusterBuilder = new ClusterBuilder();
clusterBuilder.MinimumScore = 2;
clusterBuilder.FixedSeparation = 0;
IList <Cluster> actualOutput = clusterBuilder.BuildClusters(matches);
IList <Cluster> expectedOutput = new List <Cluster>();
IList <MaxUniqueMatchExtension> clusterMatches = new List <MaxUniqueMatchExtension>();
match = new MaxUniqueMatch();
match.FirstSequenceStart = 0;
match.Length = 4;
match.SecondSequenceStart = 4;
clusterMatches.Add(new MaxUniqueMatchExtension(match));
expectedOutput.Add(new Cluster(clusterMatches));
clusterMatches = new List <MaxUniqueMatchExtension>();
match = new MaxUniqueMatch();
match.FirstSequenceStart = 4;
match.Length = 3;
match.SecondSequenceStart = 0;
clusterMatches.Add(new MaxUniqueMatchExtension(match));
expectedOutput.Add(new Cluster(clusterMatches));
clusterMatches = new List <MaxUniqueMatchExtension>();
match = new MaxUniqueMatch();
match.FirstSequenceStart = 10;
match.Length = 3;
match.SecondSequenceStart = 10;
clusterMatches.Add(new MaxUniqueMatchExtension(match));
expectedOutput.Add(new Cluster(clusterMatches));
Assert.IsTrue(this.CompareMumList(actualOutput, expectedOutput));
}
public void TestMUMmer3GetMUMsMultipleMum()
{
string reference = "ATGCGCATCCCCTT";
string search = "GCGCCCCCTA";
Sequence referenceSeq = null;
Sequence querySeq = null;
referenceSeq = new Sequence(Alphabets.DNA, reference);
querySeq = new Sequence(Alphabets.DNA, search);
List <ISequence> querySeqs = new List <ISequence>();
querySeqs.Add(querySeq);
MUMmer mummer = new MUMmer3();
mummer.LengthOfMUM = 4;
var result = mummer.GetMUMs(referenceSeq, querySeqs);
// Check if output is not null
Assert.AreNotEqual(null, result);
IDictionary <ISequence, IList <MaxUniqueMatch> > expectedOutput = new Dictionary <ISequence, IList <MaxUniqueMatch> >();
MaxUniqueMatch mum1 = new MaxUniqueMatch();
mum1.FirstSequenceMumOrder = 1;
mum1.FirstSequenceStart = 2;
mum1.SecondSequenceMumOrder = 1;
mum1.SecondSequenceStart = 0;
mum1.Length = 4;
mum1.Query = querySeq;
MaxUniqueMatch mum2 = new MaxUniqueMatch();
mum2.FirstSequenceMumOrder = 2;
mum2.FirstSequenceStart = 8;
mum2.SecondSequenceMumOrder = 2;
mum2.SecondSequenceStart = 3;
mum2.Length = 4;
mum2.Query = querySeq;
MaxUniqueMatch mum3 = new MaxUniqueMatch();
mum3.FirstSequenceMumOrder = 3;
mum3.FirstSequenceStart = 8;
mum3.SecondSequenceMumOrder = 3;
mum3.SecondSequenceStart = 4;
mum3.Length = 5;
mum3.Query = querySeq;
expectedOutput.Add(querySeq, new List <MaxUniqueMatch> {
mum1, mum2, mum3
});
Assert.IsTrue(CompareMUMs(result, expectedOutput));
}
/// <summary>
/// Copy the content to MUM
/// </summary>
/// <param name="match">Maximun unique match</param>
public void CopyTo(MaxUniqueMatch match)
{
match.FirstSequenceMumOrder = FirstSequenceMumOrder;
match.FirstSequenceStart = FirstSequenceStart;
match.SecondSequenceMumOrder = SecondSequenceMumOrder;
match.SecondSequenceStart = SecondSequenceStart;
match.Length = Length;
match.Query = Query;
}
public void TestLISWithCross1()
{
// Create a list of Mum classes.
List <MaxUniqueMatch> MUM = new List <MaxUniqueMatch>();
MaxUniqueMatch mum = null;
mum = new MaxUniqueMatch();
mum.FirstSequenceStart = 0;
mum.FirstSequenceMumOrder = 1;
mum.Length = 4;
mum.SecondSequenceStart = 4;
mum.SecondSequenceMumOrder = 1;
MUM.Add(mum);
mum = new MaxUniqueMatch();
mum.FirstSequenceStart = 4;
mum.FirstSequenceMumOrder = 2;
mum.Length = 3;
mum.SecondSequenceStart = 0;
mum.SecondSequenceMumOrder = 2;
MUM.Add(mum);
mum = new MaxUniqueMatch();
mum.FirstSequenceStart = 10;
mum.FirstSequenceMumOrder = 3;
mum.Length = 3;
mum.SecondSequenceStart = 10;
mum.SecondSequenceMumOrder = 3;
MUM.Add(mum);
ILongestIncreasingSubsequence lis = new LongestIncreasingSubsequence();
IList <MaxUniqueMatch> lisList = lis.GetLongestSequence(MUM);
List <MaxUniqueMatch> expectedOutput = new List <MaxUniqueMatch>();
mum = new MaxUniqueMatch();
mum.FirstSequenceStart = 0;
mum.FirstSequenceMumOrder = 1;
mum.Length = 4;
mum.SecondSequenceStart = 4;
mum.SecondSequenceMumOrder = 1;
expectedOutput.Add(mum);
mum = new MaxUniqueMatch();
mum.FirstSequenceStart = 10;
mum.FirstSequenceMumOrder = 3;
mum.Length = 3;
mum.SecondSequenceStart = 10;
mum.SecondSequenceMumOrder = 3;
expectedOutput.Add(mum);
Assert.IsTrue(this.CompareMumList(lisList, expectedOutput));
}
/// <summary>
/// Finds the MUMs for suffix in given interval of query sequence
/// </summary>
/// <param name="startIndex">startindex of interval</param>
/// <param name="interval">length of interval</param>
/// <returns></returns>
private List <MaxUniqueMatch> FindMUMs(int startIndex, int interval)
{
int secondSequenceStart = 0;
int secondSequenceEnd = 0;
List <MaxUniqueMatch> mumList = new List <MaxUniqueMatch>();
MaxUniqueMatch match = null;
for (int index = startIndex; index < startIndex + interval && index < _searchSequence.Count; index++)
{
// loop through each suffix of search sequence and find the MUM in suffixTree
match = Search(index);
if (null != match)
{
// Make sure the mum found does not already exists in query sequence
if ((match.SecondSequenceStart >= secondSequenceStart &&
match.SecondSequenceStart <= secondSequenceEnd) &&
(match.SecondSequenceStart + match.Length >= secondSequenceStart &&
match.SecondSequenceStart + match.Length <= secondSequenceEnd))
{
continue;
}
mumList.Add(match);
secondSequenceStart = match.SecondSequenceStart;
secondSequenceEnd = match.SecondSequenceStart + match.Length;
if (_lastEdge.IsLeaf &&
match.SecondSequenceStart + match.Length == _searchSequence.Count)
{
// At index, we have found a MUM, such that there cannot be
// another MUM (till index + Current MUM Length) who length is
// greater then Current MUM
index += _lastMatch.Length - 1;
}
}
}
// Order the mum list with query sequence order
for (int index = 0; index < mumList.Count; index++)
{
mumList[index].FirstSequenceMumOrder = index + 1;
mumList[index].SecondSequenceMumOrder = index + 1;
}
return(mumList);
}
public void TestMUMmer3GetMUMsMaxMatch()
{
string reference = "TTTTAATTTTAG";
string search = "ACTTTTGGA";
Sequence referenceSeq = null;
Sequence querySeq = null;
List <ISequence> querySeqs = null;
referenceSeq = new Sequence(Alphabets.DNA, reference);
querySeq = new Sequence(Alphabets.DNA, search);
querySeqs = new List <ISequence>();
querySeqs.Add(querySeq);
MUMmer mummer = new MUMmer3();
mummer.LengthOfMUM = 3;
var result = mummer.GetMUMs(referenceSeq, querySeqs);
// Check if output is not null
Assert.AreNotEqual(null, result);
// Check the mums count.
Assert.AreEqual(1, result.Count);
Assert.AreEqual(0, result.Values.First().Count);
mummer.MaximumMatchEnabled = true;
result = mummer.GetMUMs(referenceSeq, querySeqs);
IDictionary <ISequence, IList <MaxUniqueMatch> > expectedOutput = new Dictionary <ISequence, IList <MaxUniqueMatch> >();
MaxUniqueMatch mum = new MaxUniqueMatch();
mum.FirstSequenceMumOrder = 1;
mum.FirstSequenceStart = 0;
mum.SecondSequenceMumOrder = 1;
mum.SecondSequenceStart = 2;
mum.Length = 4;
mum.Query = querySeq;
expectedOutput.Add(querySeq, new List <MaxUniqueMatch> {
mum
});
Assert.IsTrue(CompareMUMs(result, expectedOutput));
}
/// <summary>
/// Validate following conditions and create MUM only if valid
/// 1. Make sure there is no split edge in reference sequence (this
/// represent duplicate in reference sequence)
/// 2. Validate required length of MUM
/// </summary>
/// <param name="referenceEndIndex">End index of string found in reference sequence</param>
/// <param name="queryStartIndex">Start index of string found in query sequence</param>
/// <param name="length">Length of match</param>
/// <returns>Maximum Unique Match</returns>
private MaxUniqueMatch CreateMUM(
int referenceEndIndex,
int queryStartIndex,
int length)
{
MaxUniqueMatch newMUM = null;
if (length >= _minimumLengthOfMUM)
{
newMUM = new MaxUniqueMatch();
newMUM.Query = _searchSequence;
newMUM.FirstSequenceStart = referenceEndIndex - (length - 1);
newMUM.SecondSequenceStart = queryStartIndex;
newMUM.Length = length;
}
return(newMUM);
}
public void TestMUMmer3GetFinalMUMsWithRNASingleMum()
{
string reference = "AUGCSWRYKMBVHDN";
string search = "UAUASWRYBB";
Sequence referenceSeq = null;
Sequence querySeq = null;
List <ISequence> querySeqs = null;
referenceSeq = new Sequence(Alphabets.RNA, reference);
querySeq = new Sequence(Alphabets.RNA, search);
querySeqs = new List <ISequence>();
querySeqs.Add(querySeq);
MUMmer3 mummer = new MUMmer3();
mummer.LengthOfMUM = 3;
var result = mummer.GetMUMs(referenceSeq, querySeqs, true);
// Check if output is not null
Assert.AreNotEqual(null, result);
IDictionary <ISequence, IList <MaxUniqueMatch> > expectedOutput = new Dictionary <ISequence, IList <MaxUniqueMatch> >();
MaxUniqueMatch mum1 = new MaxUniqueMatch();
mum1.FirstSequenceMumOrder = 1;
mum1.FirstSequenceStart = 4;
mum1.SecondSequenceMumOrder = 1;
mum1.SecondSequenceStart = 4;
mum1.Length = 4;
mum1.Query = querySeq;
expectedOutput.Add(querySeq, new List <MaxUniqueMatch> {
mum1
});
Assert.IsTrue(CompareMUMs(result, expectedOutput));
}
/// <summary>
/// Search for a query sequence in give Suffix Tree for existence
/// </summary>
/// <param name="startIndex">Index of first suffix character in search sequence</param>
/// <returns>Does query sequence exists</returns>
private MaxUniqueMatch Search(int startIndex)
{
// if the input sequence is empty
if (0 == _queryString.Length)
{
return(null);
}
int hash;
Edge edge = null;
hash = Hash(0, _queryString[startIndex]);
// if edge that starts with start character does not exits
if (!_suffixTree.Edges.TryGetValue(hash, out edge))
{
return(null);
}
if (-1 == edge.StartNode)
{
return(null);
}
int queryIndex = startIndex;
int referenceIndex = 0;
MaxUniqueMatch match = null;
bool matchFound = false;
Edge nextEdge = null;
while (!matchFound)
{
for (referenceIndex = edge.StartIndex; referenceIndex <= edge.EndIndex; referenceIndex++)
{
if (queryIndex == _queryString.Length || referenceIndex == _referenceString.Length)
{
match = CreateMUM(
referenceIndex - 1,
startIndex,
queryIndex - startIndex);
matchFound = true;
break;
}
if (_referenceString[referenceIndex] != _queryString[queryIndex++])
{
match = CreateMUM(
referenceIndex - 1,
startIndex,
queryIndex - 1 - startIndex);
matchFound = true;
break;
}
}
if (!matchFound)
{
if (queryIndex < _queryString.Length)
{
hash = Hash(edge.EndNode, _queryString[queryIndex]);
// If the node exists in Edge list
// Find the edge and return it
while (_suffixTree.Edges.TryGetValue(hash, out nextEdge))
{
if (_queryString[queryIndex] ==
_referenceString[nextEdge.StartIndex])
{
break;
}
hash = ++hash % _maxHashKey;
}
if (null == nextEdge)
{
match = CreateMUM(
edge.EndIndex,
startIndex,
queryIndex - startIndex);
matchFound = true;
}
else
{
edge = nextEdge;
}
}
else
{
match = CreateMUM(
edge.EndIndex,
startIndex,
queryIndex - startIndex);
matchFound = true;
}
}
}
if (null == match)
{
return(null);
}
_lastEdge = edge;
_lastMatch = match;
// Make sure there is not split, if there is split, then this is a duplicate
// and should be ignored.
// And the length of match is greater then minimum required length
if (!edge.IsLeaf)
{
match = null;
}
return(match);
}
/// <summary>
/// Find the matches of sequence in suffix tree
/// </summary>
/// <param name="suffixTree">Suffix Tree</param>
/// <param name="searchSequence">Query searchSequence</param>
/// <param name="lengthOfMUM">Mininum length of MUM</param>
/// <returns>Matches found</returns>
public IList <MaxUniqueMatch> FindMatches(
SequenceSuffixTree suffixTree,
ISequence searchSequence,
long lengthOfMUM)
{
if (suffixTree == null)
{
throw new ArgumentNullException("suffixTree");
}
if (searchSequence == null)
{
throw new ArgumentNullException("searchSequence");
}
// Initialize
_referenceString = string.Empty;
_minimumLengthOfMUM = lengthOfMUM;
_suffixTree = suffixTree;
_searchSequence = searchSequence;
_queryString = _searchSequence.ToString();
SegmentedSequence referenceSequence = _suffixTree.Sequence as SegmentedSequence;
if (null != referenceSequence)
{
foreach (Sequence sequence in referenceSequence.Sequences)
{
_referenceString += sequence.ToString() + CONCATENATING_SYMBOL;
}
// remove the concatenating symbol form end and add terminating symbol
_referenceString = _referenceString.TrimEnd(CONCATENATING_SYMBOL);
_referenceString += TERMINATING_SYMBOL;
}
else
{
_referenceString = _suffixTree.Sequence.ToString() + TERMINATING_SYMBOL;
}
int interval = (int)(_queryString.Length - (_minimumLengthOfMUM - 1)) / Environment.ProcessorCount;
if (interval < 1)
{
interval = 1;
}
IList <Task <List <MaxUniqueMatch> > > result = new List <Task <List <MaxUniqueMatch> > >();
for (int index = 0; index < _queryString.Length - (_minimumLengthOfMUM - 1); index += interval)
{
int taskIndex = index;
result.Add(
Task.Factory.StartNew <List <MaxUniqueMatch> >(
o => FindMUMs(taskIndex, interval),
TaskCreationOptions.None));
}
List <MaxUniqueMatch> mergedList = new List <MaxUniqueMatch>();
foreach (List <MaxUniqueMatch> local in result.Select(l => l.Result))
{
// Check if there is overlap, last MUM of mergedList overlaps with first MUM of local
if (0 == mergedList.Count)
{
mergedList.AddRange(local.Select(m => m));
}
else
{
if (0 < local.Count)
{
MaxUniqueMatch previous = mergedList.Last();
MaxUniqueMatch current = local.First();
if ((current.SecondSequenceStart >= previous.SecondSequenceStart &&
current.SecondSequenceStart <= previous.SecondSequenceStart + previous.Length) &&
(current.SecondSequenceStart + current.Length >= previous.SecondSequenceStart &&
current.SecondSequenceStart + current.Length <= previous.SecondSequenceStart + previous.Length))
{
local.RemoveAt(0);
}
if (0 < local.Count)
{
mergedList.AddRange(local.Select(m => m));
}
}
}
}
// Order the mum list with query sequence order
for (int index = 0; index < mergedList.Count; index++)
{
mergedList[index].FirstSequenceMumOrder = index + 1;
mergedList[index].SecondSequenceMumOrder = index + 1;
}
return(mergedList);
}
/// <summary>
/// Initializes a new instance of the MaxUniqueMatchExtension class
/// </summary>
/// <param name="mum">Maximum Unique Match</param>
public MaxUniqueMatchExtension(MaxUniqueMatch mum)
{
mum.CopyTo(this);
IsGood = false;
IsTentative = false;
}
/// <summary>
/// Search for a query sequence in give Suffix Tree for existence
/// </summary>
/// <param name="startIndex">Index of first suffix character in search sequence</param>
/// <returns>Does query sequence exists</returns>
private MaxUniqueMatch Search(int startIndex)
{
// if the input sequence is empty
if (0 == _querySequence.Count)
{
return(null);
}
IEdge edge = _suffixTree.Find(_suffixTree.Root, GetQuerySymbol(startIndex));
// if edge that starts with start character does not exits
if (edge == null)
{
return(null);
}
int queryIndex = startIndex;
int referenceIndex = 0;
MaxUniqueMatch match = null;
bool matchFound = false;
IEdge nextEdge = null;
while (!matchFound)
{
for (referenceIndex = edge.StartIndex; referenceIndex <= edge.EndIndex; referenceIndex++)
{
if (queryIndex == _querySequence.Count || referenceIndex == ReferenceLength)
{
match = CreateMUM(
referenceIndex - 1,
startIndex,
queryIndex - startIndex);
matchFound = true;
break;
}
if (GetReferenceSymbol(referenceIndex) != GetQuerySymbol(queryIndex))
{
match = CreateMUM(
referenceIndex - 1,
startIndex,
queryIndex - startIndex);
matchFound = true;
break;
}
queryIndex++;
}
if (!matchFound)
{
if (queryIndex < _querySequence.Count)
{
nextEdge = _suffixTree.Find(edge, GetQuerySymbol(queryIndex));
if (null == nextEdge)
{
match = CreateMUM(
edge.EndIndex,
startIndex,
queryIndex - startIndex);
matchFound = true;
}
else
{
edge = nextEdge;
}
}
else
{
match = CreateMUM(
edge.EndIndex,
startIndex,
queryIndex - startIndex);
matchFound = true;
}
}
}
if (null == match)
{
return(null);
}
_lastEdge = edge;
_lastMatch = match;
// Make sure there is not split, if there is split, then this is a duplicate
// and should be ignored.
// And the length of match is greater then minimum required length
if (!edge.IsLeaf && !_findMaximumMatch)
{
match = null;
}
return(match);
}
/// <summary>
/// Find the matches of sequence in suffix tree
/// </summary>
/// <param name="suffixTree">Suffix tree to searh on</param>
/// <param name="searchSequence">query sequence to find matches</param>
/// <param name="lengthOfMUM">Minimum length of the match</param>
/// <returns>Matches found</returns>
private IList <MaxUniqueMatch> FindMatchWithOption(
ISuffixTree suffixTree,
ISequence searchSequence,
long lengthOfMUM)
{
if (suffixTree == null)
{
throw new ArgumentNullException("suffixTree");
}
if (searchSequence == null)
{
throw new ArgumentNullException("searchSequence");
}
IMultiWaySuffixTree mwSuffixTree = suffixTree as IMultiWaySuffixTree;
if (mwSuffixTree == null)
{
throw new ArgumentNullException("suffixTree");
}
ValidateSequence(suffixTree.Sequence, searchSequence);
// Initialize
_minimumLengthOfMUM = lengthOfMUM;
_suffixTree = mwSuffixTree;
InitializeReferenceSequence(suffixTree.Sequence);
InitializeQuerySequence(searchSequence);
int interval = (int)(_querySequence.Count - (_minimumLengthOfMUM - 1)) / Environment.ProcessorCount;
if (interval < 1)
{
interval = 1;
}
IList <Task <List <MaxUniqueMatch> > > result = new List <Task <List <MaxUniqueMatch> > >();
for (int index = 0; index < _querySequence.Count - (_minimumLengthOfMUM - 1); index += interval)
{
int taskIndex = index;
result.Add(
Task.Factory.StartNew <List <MaxUniqueMatch> >(
o => FindMUMs(taskIndex, interval),
TaskCreationOptions.None));
}
List <MaxUniqueMatch> mergedList = new List <MaxUniqueMatch>();
foreach (List <MaxUniqueMatch> local in result.Select(l => l.Result))
{
// Check if there is overlap, last MUM of mergedList overlaps with first MUM of local
if (0 == mergedList.Count)
{
mergedList.AddRange(local.Select(m => m));
}
else
{
if (0 < local.Count)
{
MaxUniqueMatch previous = mergedList.Last();
MaxUniqueMatch current = local.First();
if ((current.SecondSequenceStart >= previous.SecondSequenceStart &&
current.SecondSequenceStart <= previous.SecondSequenceStart + previous.Length) &&
(current.SecondSequenceStart + current.Length >= previous.SecondSequenceStart &&
current.SecondSequenceStart + current.Length <= previous.SecondSequenceStart + previous.Length))
{
local.RemoveAt(0);
}
if (0 < local.Count)
{
mergedList.AddRange(local.Select(m => m));
}
}
}
}
// Order the mum list with query sequence order
for (int index = 0; index < mergedList.Count; index++)
{
mergedList[index].FirstSequenceMumOrder = index + 1;
mergedList[index].SecondSequenceMumOrder = index + 1;
}
return(mergedList);
}