/// <summary>
/// Create CDS for transcripts in this gene model, based on the translation from CDS in another model
/// </summary>
/// <param name="referenceGeneModel"></param>
public void CreateCDSFromAnnotatedStartCodons(GeneModel referenceGeneModel)
{
referenceGeneModel.GenomeForest.Build(); // so we don't need to lock the IntervalTree if we end up parallelizing this method
foreach (Gene g in Genes)
{
bool hasSource = g.FeatureMetadata.SubItems.TryGetValue("source", out List <string> sourceish);
if (!referenceGeneModel.GenomeForest.Forest.TryGetValue(g.Chromosome.FriendlyName, out IntervalTree tree))
{
continue;
}
foreach (Transcript t in g.Transcripts)
{
List <Transcript> referenceTranscripts = tree.Query(t).OfType <Transcript>().ToList();
List <Transcript> referenceTranscriptsWithCDS = referenceTranscripts.Where(tt => tt.IsProteinCoding()).ToList();
foreach (Transcript tWithCds in referenceTranscriptsWithCDS)
{
lock (tWithCds)
{
if (t.CreateCDSFromAnnotatedStartCodons(tWithCds))
{
break;
} // for now, only use the first annotation found if any
}
}
}
}
}
public override string GetGtfAttributes()
{
var attributes = GeneModel.SplitAttributes(FeatureMetadata.FreeText);
List <Tuple <string, string> > attributeSubsections = new List <Tuple <string, string> >();
string exonIdLabel = "exon_id";
bool hasExonId = attributes.TryGetValue(exonIdLabel, out string exonId);
if (hasExonId)
{
attributeSubsections.Add(new Tuple <string, string>(exonIdLabel, exonId));
}
string exonVersionLabel = "exon_version";
bool hasExonVersion = attributes.TryGetValue(exonVersionLabel, out string exonVersion);
if (hasExonVersion)
{
attributeSubsections.Add(new Tuple <string, string>(exonVersionLabel, exonVersion));
}
string exonNumberLabel = "exon_number";
string exonNumber = (Parent as Transcript).Exons.Count(x => x.OneBasedStart <= OneBasedStart).ToString();
attributeSubsections.Add(new Tuple <string, string>(exonNumberLabel, exonNumber));
return(Parent.GetGtfAttributes() + " " + String.Join(" ", attributeSubsections.Select(x => x.Item1 + " \"" + x.Item2 + "\";")));
}
public override string GetGtfAttributes()
{
var attributes = GeneModel.SplitAttributes(FeatureMetadata.FreeText);
List <Tuple <string, string> > attributeSubsections = new List <Tuple <string, string> >();
string tIdLabel = "transcript_id";
bool hasTranscriptId = attributes.TryGetValue(tIdLabel, out string transcriptId);
if (hasTranscriptId)
{
attributeSubsections.Add(new Tuple <string, string>(tIdLabel, transcriptId));
}
string tVersionLabel = "transcript_version";
bool hasTranscriptVersion = attributes.TryGetValue(tVersionLabel, out string transcriptVersion);
if (hasTranscriptVersion)
{
attributeSubsections.Add(new Tuple <string, string>(tVersionLabel, transcriptVersion));
}
string tBiotypeLabel = "transcript_biotype";
bool hasTranscriptBiotype = attributes.TryGetValue(tBiotypeLabel, out string transcriptBiotype);
if (hasTranscriptVersion)
{
attributeSubsections.Add(new Tuple <string, string>(tBiotypeLabel, transcriptBiotype));
}
// Cufflinks-related, but not using Cufflinks much because stringtie is better
//bool hasNearestRef = attributes.TryGetValue("nearest_ref", out string nearestRef);
//bool hasClassCode = attributes.TryGetValue("class_code", out string classCode);
bool hasSource = FeatureMetadata.SubItems.TryGetValue("source", out List <string> sourceish);
bool hasStrand = FeatureMetadata.SubItems.TryGetValue("strand", out List <string> strandish);
bool hasFrame = FeatureMetadata.SubItems.TryGetValue("frame", out List <string> framey);
return(Parent.GetGtfAttributes() + " " + String.Join(" ", attributeSubsections.Select(x => x.Item1 + " \"" + x.Item2 + "\";")));
}
public override string GetGtfAttributes()
{
var attributes = GeneModel.SplitAttributes(FeatureMetadata.FreeText);
List <Tuple <string, string> > attributeSubsections = new List <Tuple <string, string> >();
string geneIdLabel = "gene_id";
bool hasGeneId = attributes.TryGetValue(geneIdLabel, out string geneId);
if (hasGeneId)
{
attributeSubsections.Add(new Tuple <string, string>(geneIdLabel, geneId));
}
string geneNameLabel = "gene_name";
bool hasGeneName = attributes.TryGetValue(geneNameLabel, out string geneName);
if (hasGeneName)
{
attributeSubsections.Add(new Tuple <string, string>(geneNameLabel, geneName));
}
string geneVersionLabel = "gene_version";
bool hasGeneVersion = attributes.TryGetValue(geneVersionLabel, out string geneVersion);
if (hasGeneVersion)
{
attributeSubsections.Add(new Tuple <string, string>(geneVersionLabel, geneVersion));
}
string geneBiotypeLabel = "gene_biotype";
bool hasGeneBiotype = attributes.TryGetValue(geneBiotypeLabel, out string geneBiotype);
if (hasGeneBiotype)
{
attributeSubsections.Add(new Tuple <string, string>(geneBiotypeLabel, geneBiotype));
}
return(String.Join(" ", attributeSubsections.Select(x => x.Item1 + " \"" + x.Item2 + "\";")));
}
} = new Dictionary <string, int>(); // key: mappingStrand + strandFromGene
/// <summary>
/// Given a BAM file, try to guess the RNA-Seq experiment:
/// 1) single-end or pair-end
/// 2) strand_specific or not
/// 3) if it is strand-specific, what's the strand_ness of the protocol
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="bamPath"></param>
/// <param name="geneModelPath"></param>
/// <param name="minFractionStrandSpecific"></param>
/// <returns></returns>
private void CheckProperties(string bamPath, string geneModelPath, Genome genome, double minFractionStrandSpecific)
{
GeneModel gm = new GeneModel(genome, geneModelPath);
using (var reader = File.OpenRead(bamPath))
{
Console.WriteLine("Reading BAM file.");
// read bam, and filter out reads that are QC failures, unmapped, duplicates, or secondary
BAMParser bam = new BAMParser();
var unfilteredReads = bam.Parse(reader).ToList();
var reads = unfilteredReads.Where(read =>
!read.Flag.HasFlag(SAMFlags.QualityCheckFailure) && !read.Flag.HasFlag(SAMFlags.UnmappedQuery) &&
!read.Flag.HasFlag(SAMFlags.Duplicate) && !read.Flag.HasFlag(SAMFlags.NonPrimeAlignment)).ToList();
Console.WriteLine("Evaluating reads.");
Parallel.ForEach(reads, read =>
{
// set the interval contained by this read, and get the gene regions nearby
bool isReversed = read.Flag.HasFlag(SAMFlags.QueryOnReverseStrand);
Interval readInterval = new Interval(null, read.RName, "source", isReversed ? "-" : "+", read.Pos, read.RefEndPos, null);
bool hasNearbyRegion = gm.GenomeForest.Forest.TryGetValue(readInterval.ChromosomeID, out IntervalTree nearbyGeneTree);
if (hasNearbyRegion)
{
List <Interval> nearbyGeneRegions = nearbyGeneTree.Query(readInterval);
if (nearbyGeneRegions.Count > 0)
{
// count up paired-end or single-end read properties
string mapStrand = isReversed ? "-" : "+";
bool isPaired = read.Flag.HasFlag(SAMFlags.PairedRead);
bool isRead1 = read.Flag.HasFlag(SAMFlags.FirstReadInPair);
bool isRead2 = read.Flag.HasFlag(SAMFlags.SecondReadInPair);
string readId = isRead1 ? "1" : isRead2 ? "2" : null;
HashSet <string> strandFromGene = new HashSet <string>(nearbyGeneRegions.Select(x => x.Strand));
foreach (string strand in strandFromGene)
{
Dictionary <string, int> dict = isPaired ? PairedStrandedness : SingleStrandedness;
string key = isPaired ?
readId + mapStrand + strand :
mapStrand + strand;
lock (dict)
{
if (dict.TryGetValue(key, out int count))
{
count++;
}
else
{
dict[key] = 1;
}
}
}
}
}
});
// From RSeQC:
// Not strand specific:
// This is PairEnd Data
// Fraction of reads failed to determine: 0.0172
// Fraction of reads explained by "1++,1--,2+-,2-+": 0.4903
// Fraction of reads explained by "1+-,1-+,2++,2--": 0.4925
// Strand specific:
// This is PairEnd Data
// Fraction of reads failed to determine: 0.0072
// Fraction of reads explained by "1++,1--,2+-,2-+": 0.9441
// Fraction of reads explained by "1+-,1-+,2++,2--": 0.0487
SingleStrandedness.TryGetValue("++", out int sForward1);
SingleStrandedness.TryGetValue("--", out int sForward2);
SingleStrandedness.TryGetValue("+-", out int sReverse1);
SingleStrandedness.TryGetValue("-+", out int sReverse2);
PairedStrandedness.TryGetValue("1++", out int pForward1);
PairedStrandedness.TryGetValue("1--", out int pForward2);
PairedStrandedness.TryGetValue("2+-", out int pForward3);
PairedStrandedness.TryGetValue("2-+", out int pForward4);
PairedStrandedness.TryGetValue("1+-", out int pReverse1);
PairedStrandedness.TryGetValue("1-+", out int pReverse2);
PairedStrandedness.TryGetValue("2++", out int pReverse3);
PairedStrandedness.TryGetValue("2--", out int pReverse4);
if (PairedStrandedness.Count > 0 && SingleStrandedness.Count == 0)
{
Protocol = RnaSeqProtocol.PairedEnd;
FractionForwardStranded = (double)(pForward1 + pForward2 + pForward3 + pForward4) / (double)PairedStrandedness.Values.Sum();
FractionReverseStranded = (double)(pReverse1 + pReverse2 + pReverse3 + pReverse4) / (double)PairedStrandedness.Values.Sum();
FractionUndetermined = 1 - FractionForwardStranded - FractionReverseStranded;
if (FractionUndetermined > 0.5)
{
throw new ArgumentException("A large number of reads failed to determine the standedness of the protocol within " + bamPath);
}
Strandedness = FractionForwardStranded >= minFractionStrandSpecific ? Strandedness.Forward :
FractionReverseStranded >= minFractionStrandSpecific ? Strandedness.Reverse :
Strandedness.None;
}
else if (SingleStrandedness.Count > 0 && PairedStrandedness.Count == 0)
{
Protocol = RnaSeqProtocol.SingleEnd;
FractionForwardStranded = (double)(sForward1 + sForward2) / (double)SingleStrandedness.Values.Sum();
FractionReverseStranded = (double)(sReverse1 + sReverse2) / (double)SingleStrandedness.Values.Sum();
FractionUndetermined = 1 - FractionForwardStranded - FractionReverseStranded;
if (FractionUndetermined > 0.5)
{
throw new ArgumentException("A large number of reads failed to determine the standedness of the protocol within " + bamPath);
}
Strandedness = FractionForwardStranded >= minFractionStrandSpecific ? Strandedness.Forward :
FractionReverseStranded >= minFractionStrandSpecific ? Strandedness.Reverse :
Strandedness.None;
}
else
{
Protocol = RnaSeqProtocol.Mixture;
Strandedness = Strandedness.None;
FractionForwardStranded = (double)(sForward1 + sForward2 + pForward1 + pForward2 + pForward3 + pForward4) / (double)PairedStrandedness.Values.Sum();
FractionReverseStranded = (double)(sReverse1 + sReverse2 + pReverse1 + pReverse2 + pReverse3 + pReverse4) / (double)PairedStrandedness.Values.Sum();
FractionUndetermined = 1 - FractionForwardStranded - FractionReverseStranded;
if (FractionUndetermined > 0.5)
{
throw new ArgumentException("A large number of reads failed to determine the standedness of the protocol within " + bamPath);
}
Strandedness = FractionForwardStranded >= minFractionStrandSpecific ? Strandedness.Forward :
FractionReverseStranded >= minFractionStrandSpecific ? Strandedness.Reverse :
Strandedness.None;
}
}
}