/// <summary>
/// Assemble the list of sequence reads
/// </summary>
/// <param name="inputSequences">List of input sequences</param>
/// <returns>Assembled output</returns>
public IDeNovoAssembly Assemble(IList <ISequence> inputSequences)
{
ValidateAllSequences(inputSequences);
// Remove ambiguous reads and set up fields for assembler process
Initialize(inputSequences);
// Step 1, 2: Create k-mers from reads and build de bruijn graph
CreateGraph();
// Estimate and set default value for erosion and coverage thresholds
EstimateDefaultThresholds();
// Step 3: Remove dangling links from graph
UnDangleGraph();
// Step 4: Remove redundant paths from graph
RemoveRedundancy();
// Perform dangling link purger step once more.
// This is done to remove any links created by redundant paths purger.
UnDangleGraph();
// Step 5: Build Contigs
IList <ISequence> contigSequences = BuildContigs();
PaDeNAAssembly result = new PaDeNAAssembly();
result.AddContigs(contigSequences);
_graph.Dispose();
return(result);
}
/// <summary>
/// Generate sequences from list of contig nodes.
/// </summary>
/// <param name="contigGraph">Contig Overlap Graph.</param>
/// <param name="paths">Scaffold paths.</param>
/// <returns>List of sequences of scaffolds.</returns>
protected IList <ISequence> GenerateScaffold(
DeBruijnGraph contigGraph,
IList <ScaffoldPath> paths)
{
if (contigGraph == null)
{
throw new ArgumentNullException("contigGraph");
}
if (paths == null)
{
throw new ArgumentNullException("paths");
}
List <ISequence> scaffolds = paths.AsParallel().Select(t => t.BuildSequenceFromPath(contigGraph, _kmerLength)).ToList();
IEnumerable <DeBruijnNode> visitedNodes = contigGraph.Nodes.AsParallel().Where(t => !t.IsMarked());
scaffolds.AddRange(visitedNodes.AsParallel().Select(t => contigGraph.GetNodeSequence(t)));
contigGraph.Dispose();
return(scaffolds);
}
