/// <summary>
/// Intersect bins with the targeted regions defined in callset.Manifest.
/// Assumes that the targeted regions don't intersect, the bins are sorted by genomic location and the bins don't intersect.
/// </summary>
/// <param name="callset"></param>
/// <param name="partitionedPath">Output of CanvasPartition. Bins are assumed to be sorted</param>
/// <returns></returns>
private IFileLocation IntersectBinsWithTargetedRegions(CanvasCallset callset, IFileLocation partitionedPath)
{
if (!partitionedPath.Exists) { return partitionedPath; }
var rawPartitionedPath = partitionedPath.AppendName(".raw");
if (rawPartitionedPath.Exists) { rawPartitionedPath.Delete(); }
partitionedPath.MoveTo(rawPartitionedPath);
//callset.Manifest
Dictionary<string, List<NexteraManifest.ManifestRegion>> manifestRegionsByChrom = callset.Manifest.GetManifestRegionsByChromosome();
// CanvasPartition output file is in the BED format
// start: 0-based, inclusive
// end: 0-based, exclusive
// Manifest
// start: 1-based, inclusive
// end: 1-based, inclusive
using (GzipReader reader = new GzipReader(rawPartitionedPath.FullName))
using (GzipWriter writer = new GzipWriter(partitionedPath.FullName))
{
string currentChrom = null;
int manifestRegionIdx = 0;
string line;
string[] toks;
while ((line = reader.ReadLine()) != null)
{
toks = line.Split('\t');
string chrom = toks[0];
int start = int.Parse(toks[1]) + 1; // 1-based, inclusive
int end = int.Parse(toks[2]); // 1-based, inclusive
if (chrom != currentChrom)
{
currentChrom = chrom;
manifestRegionIdx = 0;
}
if (!manifestRegionsByChrom.ContainsKey(currentChrom)) { continue; }
while (manifestRegionIdx < manifestRegionsByChrom[currentChrom].Count
&& manifestRegionsByChrom[currentChrom][manifestRegionIdx].End < start) // |- manifest region -| |- bin -|
{
manifestRegionIdx++;
}
if (manifestRegionIdx >= manifestRegionsByChrom[currentChrom].Count || // |- last manifest region -| |- bin -|
end < manifestRegionsByChrom[currentChrom][manifestRegionIdx].Start) // |- bin -| |- manifest region -|
{
continue; // skip bin
}
// |- bin -|
// |- manifest region -|
while (manifestRegionIdx < manifestRegionsByChrom[currentChrom].Count &&
end >= manifestRegionsByChrom[currentChrom][manifestRegionIdx].Start)
{
// calculate intersection
int intersectionStart = Math.Max(start, manifestRegionsByChrom[currentChrom][manifestRegionIdx].Start); // 1-based, inclusive
int intersectionEnd = Math.Min(end, manifestRegionsByChrom[currentChrom][manifestRegionIdx].End); // 1-based, inclusive
// start/end in BED format
toks[1] = String.Format("{0}", intersectionStart - 1); // 0-based, inclusive
toks[2] = String.Format("{0}", intersectionEnd); // 0-based, exclusive
// write intersected bin
writer.WriteLine(String.Join("\t", toks));
manifestRegionIdx++;
}
}
}
return partitionedPath;
}
/// <summary>
/// Not making this an extension method so that it doesn't collide with the instance method
/// We still want to be able to refer to it statically though
/// </summary>
/// <param name="source"></param>
/// <param name="destination"></param>
public static void MoveTo(IFileLocation source, IFileLocation destination)
{
source.MoveTo(destination);
}
/// <summary>
/// Intersect bins with the targeted regions defined in callset.Manifest.
/// Assumes that the targeted regions don't intersect, the bins are sorted by genomic location and the bins don't intersect.
/// </summary>
/// <param name="callset"></param>
/// <param name="partitionedPath">Output of CanvasPartition. Bins are assumed to be sorted</param>
/// <returns></returns>
private IFileLocation IntersectBinsWithTargetedRegions(CanvasCallset callset, IFileLocation partitionedPath)
{
if (!partitionedPath.Exists)
{
return(partitionedPath);
}
var rawPartitionedPath = partitionedPath.AppendName(".raw");
if (rawPartitionedPath.Exists)
{
rawPartitionedPath.Delete();
}
partitionedPath.MoveTo(rawPartitionedPath);
//callset.Manifest
Dictionary <string, List <NexteraManifest.ManifestRegion> > manifestRegionsByChrom = callset.Manifest.GetManifestRegionsByChromosome();
// CanvasPartition output file is in the BED format
// start: 0-based, inclusive
// end: 0-based, exclusive
// Manifest
// start: 1-based, inclusive
// end: 1-based, inclusive
using (GzipReader reader = new GzipReader(rawPartitionedPath.FullName))
using (GzipWriter writer = new GzipWriter(partitionedPath.FullName))
{
string currentChrom = null;
int manifestRegionIdx = 0;
string line;
string[] toks;
while ((line = reader.ReadLine()) != null)
{
toks = line.Split('\t');
string chrom = toks[0];
int start = int.Parse(toks[1]) + 1; // 1-based, inclusive
int end = int.Parse(toks[2]); // 1-based, inclusive
if (chrom != currentChrom)
{
currentChrom = chrom;
manifestRegionIdx = 0;
}
if (!manifestRegionsByChrom.ContainsKey(currentChrom))
{
continue;
}
while (manifestRegionIdx < manifestRegionsByChrom[currentChrom].Count &&
manifestRegionsByChrom[currentChrom][manifestRegionIdx].End < start) // |- manifest region -| |- bin -|
{
manifestRegionIdx++;
}
if (manifestRegionIdx >= manifestRegionsByChrom[currentChrom].Count || // |- last manifest region -| |- bin -|
end < manifestRegionsByChrom[currentChrom][manifestRegionIdx].Start) // |- bin -| |- manifest region -|
{
continue; // skip bin
}
// |- bin -|
// |- manifest region -|
while (manifestRegionIdx < manifestRegionsByChrom[currentChrom].Count &&
end >= manifestRegionsByChrom[currentChrom][manifestRegionIdx].Start)
{
// calculate intersection
int intersectionStart = Math.Max(start, manifestRegionsByChrom[currentChrom][manifestRegionIdx].Start); // 1-based, inclusive
int intersectionEnd = Math.Min(end, manifestRegionsByChrom[currentChrom][manifestRegionIdx].End); // 1-based, inclusive
// start/end in BED format
toks[1] = String.Format("{0}", intersectionStart - 1); // 0-based, inclusive
toks[2] = String.Format("{0}", intersectionEnd); // 0-based, exclusive
// write intersected bin
writer.WriteLine(String.Join("\t", toks));
manifestRegionIdx++;
}
}
}
return(partitionedPath);
}