// Public Methods ===============================================================================================//
/// <summary>
/// Preprocesses the expression.
/// </summary>
public void PreprocessExpression()
{
this.currentStage = Stage.Preprocess;
BedFile.ToFileBed6(this.LocusFile.Locations.Values.Where(x => !this.BadLocusSet.Contains(x.Name)).ToList(), this.ProcessedLociFilename);
BedFile.ToFileBed6(this.TranscriptLocations.Values.ToList(), this.ProcessedTranscriptLocationFilename);
using (TextWriter tw = Helpers.CreateStreamWriter(this.ProcessedExpressionFilename))
{
tw.WriteLine("TSS_ID\t" + string.Join("\t", this.TissueSources));
tw.WriteLine(string.Join(
"\n",
this.TranscriptExpression
.Select(x => x.Key + "\t" + string.Join(
"\t",
this.TissueSources.Select(c => x.Value.ContainsKey(c) ? x.Value[c].ToString() : "NaN")))));
}
using (TextWriter tw2 = Helpers.CreateStreamWriter(this.ProcessedHistoneFilename))
{
tw2.WriteLine("RE_Locus\t" + string.Join("\t", this.TissueSources));
tw2.WriteLine(string.Join(
"\n",
this.LocusData
.Select(x => x.Key + "\t" + string.Join(
"\t",
this.TissueSources.Select(c => x.Value.ContainsKey(c) ? x.Value[c].ToString() : "NaN")))));
}
}
/// <summary>
/// Gets the overlapping regions.
/// </summary>
/// <returns>The overlapping regions.</returns>
/// <param name="regions">Regions to be overlapped</param>
/// <param name="overlay">Regions to overlay</param>
private HashSet <string> GetOverlappingRegions(BedFile regions, BedFile overlay)
{
return(new HashSet <string>(regions.Locations
.Where(x => overlay.MaxLocationSize.ContainsKey(x.Value.Chromosome))
.Select(x => TRFScorer.GetOverlaps(
x.Value,
overlay.ChromosomeIndexedLocations,
BedFile.IndexSize,
overlay.MaxLocationSize[x.Value.Chromosome]))
.SelectMany(x => x)
.Select(x => x.Name)));
}
/// <summary>
/// Execute this instance.
/// </summary>
public void Execute()
{
var peaks = new BedFile(this.ChIPSeqFile, BedFile.Bed6Plus4Layout);
var motifMatches = new BedFile(this.MotifMatchFile, BedFile.Bed6Layout);
var segmentation = new BedFile(this.SegmentationFile, BedFile.Bed6Layout);
HashSet <string> peaksOverlappingMotifs = this.GetOverlappingRegions(motifMatches, peaks);
HashSet <string> peaksOverlappingSegment = this.GetOverlappingRegions(segmentation, peaks);
HashSet <string> motifsOverlappingSegment = this.GetOverlappingRegions(segmentation, motifMatches);
int peaksAtMotifs = peaksOverlappingMotifs.Count;
int peaksInSegment = peaksOverlappingSegment.Count(x => peaksOverlappingMotifs.Contains(x));
int motifsInSegment = motifsOverlappingSegment.Count;
int motifs = motifMatches.Locations.Count;
Console.WriteLine("First peak overlapping segment " + peaksOverlappingSegment.First());
Console.WriteLine("First peak overlapping motif " + peaksOverlappingMotifs.First());
Console.WriteLine("Peaks\tPeaksOverlappingMotif\tMotifs\tPeaksInSegment\tPeaksInSegmentOverlappingMotif\tMotifsInSegment");
Console.WriteLine(string.Join("\t", new List <int> {
peaks.Locations.Count, peaksAtMotifs, motifs, peaksOverlappingSegment.Count, peaksInSegment, motifsInSegment
}));
}
public void Convert()
{
var contacts = Helpers.GetFileDataLines(this.ContactFileName, true).Select(line =>
{
var fields = line.Split('\t');
var pair = fields[6];
var ids = pair.Split('_');
return(new
{
Pair = pair,
Id1 = ids[0],
Chr1 = fields[0],
Start1 = int.Parse(fields[1]),
End1 = int.Parse(fields[2]),
Id2 = ids[1],
Chr2 = fields[3],
Start2 = int.Parse(fields[4]),
End2 = int.Parse(fields[5]),
Score = double.Parse(fields[7]),
});
}).ToList();
var locations = contacts.Select(x => new Location[]
{
new Location
{
Name = x.Id1, Chromosome = x.Chr1, Start = x.Start1, End = x.End1, AlternateName = x.Id2, Score = x.Score
},
new Location
{
Name = x.Id2, Chromosome = x.Chr2, Start = x.Start2, End = x.End2, AlternateName = x.Id1, Score = x.Score
},
}).SelectMany(x => x)
.ToList();
var uniqueLocations = locations
.ToLookup(x => x.Name)
.ToDictionary(x => x.Key, x => x.First());
var contactMap = locations
.ToLookup(x => x.Name, x => x)
.ToDictionary(x => x.Key, x => x
.ToDictionary(y => y.AlternateName, y => uniqueLocations[y.AlternateName]));
var indexedLocations = new BedFile(uniqueLocations);
var LocusFile = new BedFile(this.LocusFileName, BedFile.Bed3Layout);
var annotationSet = new GtfExpressionFile(GtfExpressionFile.ExpressionType.Cage, this.AnnotationFileName);
var LocusOverlaps = indexedLocations.Locations.ToDictionary(
x => x.Key,
x => TRFScorer.GetOverlaps(
x.Value,
LocusFile.ChromosomeIndexedLocations,
BedFile.IndexSize,
LocusFile.MaxLocationSize[x.Value.Chromosome]));
var tssOverlaps = indexedLocations.Locations.ToDictionary(
x => x.Key,
x => TRFScorer.GetOverlaps(
x.Value,
annotationSet.ChromosomeIndexedLocations,
BedFile.IndexSize,
annotationSet.MaxLocationSize[x.Value.Chromosome],
Location.OverlapsDirectionalStart));
var links = contactMap
.Where(x => tssOverlaps[x.Key] != null)
.ToDictionary(x => x.Key, x => x.Value.Where(y => LocusOverlaps[y.Key] != null))
.Where(x => x.Value.Any())
.Select(x => tssOverlaps[x.Key].Select(y => new {
Id1 = x.Key,
Tss = y.Name,
LocusList = x.Value
.Select(z => LocusOverlaps[z.Key]
.Select(a => new { Name = a.Name, Id2 = z.Key }))
.SelectMany(z => z).ToList()
})
.Select(y => y.LocusList.Select(z =>
{
var LocusLocation = LocusFile.Locations[z.Name];
var tssStart = annotationSet.Transcripts[y.Tss].DirectionalStart;
var linkLength = LocusLocation.End < tssStart ? LocusLocation.End - tssStart : LocusLocation.Start - tssStart;
return(new MapLink
{
TranscriptName = y.Tss,
TssName = y.Tss,
LocusName = z.Name,
ConfidenceScore = contactMap[y.Id1][z.Id2].Score,
LinkLength = linkLength,
});
}))
.SelectMany(y => y))
.SelectMany(x => x)
.ToList();
var convertedMap = new TssRegulatoryMap(links);
NullMapBuilder.WriteMap(convertedMap, annotationSet.Locations, "Contact", this.MapFileName);
}