/// <summary>
/// Public method mapping Reads to Contigs.
/// </summary>
/// <param name="contigs">List of sequences of contigs.</param>
/// <param name="reads">List of input reads.</param>
/// <param name="kmerLength">Length of kmer.</param>
/// <returns>Contig Read Map.</returns>
public ReadContigMap Map(IList <ISequence> contigs, IEnumerable <ISequence> reads, int kmerLength)
{
KmerIndexerDictionary map = SequenceToKmerBuilder.BuildKmerDictionary(contigs, kmerLength);
ReadContigMap maps = new ReadContigMap();
Parallel.ForEach(reads, readSequence =>
{
IEnumerable <ISequence> kmers = SequenceToKmerBuilder.GetKmerSequences(readSequence, kmerLength);
ReadIndex read = new ReadIndex(readSequence);
foreach (ISequence kmer in kmers)
{
IList <KmerIndexer> positions;
if (map.TryGetValue(kmer, out positions) ||
map.TryGetValue(kmer.GetReverseComplementedSequence(), out positions))
{
read.ContigReadMatchIndexes.Add(positions);
}
}
IList <Task <IList <ReadMap> > > tasks =
new List <Task <IList <ReadMap> > >();
// Stores information about contigs for which tasks has been generated.
IList <long> visitedContigs = new List <long>();
// Creates Task for every read in nodes for a given contig.
for (int index = 0; index < read.ContigReadMatchIndexes.Count; index++)
{
int readPosition = index;
foreach (KmerIndexer kmer in read.ContigReadMatchIndexes[index])
{
long contigIndex = kmer.SequenceIndex;
if (!visitedContigs.Contains(contigIndex))
{
visitedContigs.Add(contigIndex);
tasks.Add(
Task <IList <ReadMap> > .Factory.StartNew(
t => MapRead(
readPosition,
read.ContigReadMatchIndexes,
contigIndex,
read.ReadSequence.Count,
kmerLength),
TaskCreationOptions.AttachedToParent));
}
}
}
var overlapMaps = new Dictionary <ISequence, IList <ReadMap> >();
for (int index = 0; index < visitedContigs.Count; index++)
{
overlapMaps.Add(contigs.ElementAt(visitedContigs[index]), tasks[index].Result);
}
lock (maps)
{
if (!maps.ContainsKey(read.ReadSequence.ID))
{
maps.Add(read.ReadSequence.ID, overlapMaps);
}
else
{
throw new ArgumentException(
string.Format(CultureInfo.CurrentCulture, Resource.DuplicatingReadIds, read.ReadSequence.ID));
}
}
});
return(maps);
}
/// <summary>
/// Aligns reads to contigs using kmer method of alignment.
/// </summary>
/// <param name="contigs">List of contig sequences.</param>
/// <param name="reads">List of read sequences.</param>
/// <param name="kmerLength">Kmer Length.</param>
/// <returns>List of Contig.</returns>
public static IList <Contig> ReadContigAlignment(IList <ISequence> contigs, IList <ISequence> reads, int kmerLength)
{
KmerIndexerDictionary map = SequenceToKmerBuilder.BuildKmerDictionary(reads, kmerLength);
IList <ContigIndex> contigDatas;
contigDatas = contigs.AsParallel().Select(contig =>
{
IEnumerable <ISequence> kmers = SequenceToKmerBuilder.GetKmerSequences(contig, kmerLength);
ContigIndex index = new ContigIndex(contig);
IList <KmerIndexer> positions;
foreach (ISequence kmer in kmers)
{
if (map.TryGetValue(kmer, out positions) ||
map.TryGetValue(kmer.GetReverseComplementedSequence(), out positions))
{
index.ContigReadMatchIndexes.Add(positions);
}
else
{
index.ContigReadMatchIndexes.Add(new List <KmerIndexer>());
}
}
return(index);
}).ToList();
return(contigDatas.Select(contigData =>
{
IList <Task <IList <ReadMap> > > tasks =
new List <Task <IList <ReadMap> > >();
// Stores information about contigs for which tasks has been generated.
IList <long> visitedReads = new List <long>();
// Creates Task for every read in nodes for a given contig.
for (int index = 0; index < contigData.ContigReadMatchIndexes.Count; index++)
{
int readPosition = index;
foreach (KmerIndexer kmer in contigData.ContigReadMatchIndexes[index])
{
long contigIndex = kmer.SequenceIndex;
if (!visitedReads.Contains(contigIndex))
{
visitedReads.Add(contigIndex);
tasks.Add(
Task <IList <ReadMap> > .Factory.StartNew(t => MapRead(readPosition, contigData.ContigReadMatchIndexes, contigIndex, kmerLength), TaskCreationOptions.AttachedToParent));
}
}
}
Contig contigOutputStructure = new Contig();
contigOutputStructure.Consensus = contigData.ContigSequence;
for (int index = 0; index < visitedReads.Count; index++)
{
foreach (ReadMap maps in tasks[index].Result)
{
Contig.AssembledSequence assembledSeq = new Contig.AssembledSequence()
{
Length = maps.Length,
Position = maps.StartPositionOfContig,
ReadPosition = maps.StartPositionOfRead,
Sequence = reads.ElementAt(visitedReads[index])
};
if (new string(
contigOutputStructure.Consensus.GetSubSequence(
assembledSeq.Position, assembledSeq.Length).Select(a => (char)a).ToArray()).
Equals(new string(assembledSeq.Sequence.GetSubSequence(assembledSeq.ReadPosition, assembledSeq.Length)
.Select(a => (char)a).ToArray())))
{
assembledSeq.IsComplemented = false;
assembledSeq.IsReversed = false;
}
else
{
assembledSeq.IsComplemented = true;
assembledSeq.IsReversed = true;
}
contigOutputStructure.Sequences.Add(assembledSeq);
}
}
return contigOutputStructure;
}).ToList());
}
public void PathPurger1()
{
const int KmerLength = 7;
ISequence sequence = new Sequence(Alphabets.DNA, "GATTCAAGGGCTGGGGG");
IList <ISequence> contigsSequence = SequenceToKmerBuilder.GetKmerSequences(sequence, KmerLength).ToList();
ContigGraph graph = new ContigGraph();
graph.BuildContigGraph(contigsSequence, KmerLength);
List <Node> contigs = graph.Nodes.ToList();
IList <ScaffoldPath> paths =
new List <ScaffoldPath>();
ScaffoldPath path = new ScaffoldPath();
foreach (Node node in contigs)
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(2, 5))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(3, 5))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(6, 5))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(0, 11))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(7, 4))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(11, 0))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(2, 9))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
path = new ScaffoldPath();
foreach (Node node in contigs.GetRange(1, 10))
{
path.Add(new KeyValuePair <Node, Edge>(node, null));
}
paths.Add(path);
PathPurger assembler = new PathPurger();
assembler.PurgePath(paths);
Assert.AreEqual(paths.Count, 1);
Assert.IsTrue(Compare(paths.First(), contigs));
}