private static MutableGene GetFlattenedGene(MutableGene seedGene, List <MutableGene> genesWithSameGeneId,
int overlapStart, int overlapEnd)
{
var flattenedGene = MutableGene.Clone(seedGene);
bool useOverlap = overlapStart != -1 && overlapEnd != -1;
foreach (var gene in genesWithSameGeneId)
{
if (gene.Invalid || flattenedGene.OnReverseStrand != gene.OnReverseStrand ||
flattenedGene.ReferenceIndex != gene.ReferenceIndex)
{
continue;
}
if (useOverlap && !Overlap.Partial(overlapStart, overlapEnd, gene.Start, gene.End))
{
continue;
}
if (!useOverlap && !Overlap.Partial(flattenedGene.Start, flattenedGene.End, gene.Start, gene.End))
{
continue;
}
UpdateCoordinates(gene, flattenedGene);
gene.Invalid = true;
}
return(flattenedGene);
}
private List <MutableGene> GetValidGenes(MutableGene seedGene, List <MutableGene> genes, out int start,
out int end)
{
var validGenes = new List <MutableGene>();
start = seedGene.Start;
end = seedGene.End;
foreach (var gene in genes)
{
if (gene.Invalid || seedGene.OnReverseStrand != gene.OnReverseStrand ||
seedGene.ReferenceIndex != gene.ReferenceIndex || !Overlap.Partial(start, end, gene.Start, gene.End))
{
continue;
}
validGenes.Add(gene);
if (gene.Start < start)
{
start = gene.Start;
}
if (gene.End > end)
{
end = gene.End;
}
}
return(validGenes);
}
private void MergesGenesWithSameSymbol(MutableGene seedGene, List <MutableGene> genesWithSameSymbol)
{
int overlapStart, overlapEnd;
var validGenes = GetValidGenes(seedGene, genesWithSameSymbol, out overlapStart, out overlapEnd);
var ensemblGenes = GeneUtilities.GetGenesByDataSource(validGenes, TranscriptDataSource.Ensembl);
var refSeqGenes = GeneUtilities.GetGenesByDataSource(validGenes, TranscriptDataSource.RefSeq);
var ensemblFlattener = new GeneFlattener(ensemblGenes, "Ensembl", false);
var flatEnsemblGenes = ensemblFlattener.Flatten(overlapStart, overlapEnd);
var refSeqFlattener = new GeneFlattener(refSeqGenes, "RefSeq", false);
var flatRefSeqGenes = refSeqFlattener.Flatten(overlapStart, overlapEnd);
foreach (var ensemblGene in flatEnsemblGenes)
{
// add the unused Ensembl genes
string linkedEntrezId;
if (!_linkedEnsemblIds.TryGetValue(ensemblGene.EnsemblId.ToString(), out linkedEntrezId))
{
AddEnsemblOrphan(ensemblGene);
continue;
}
var refSeqGene = GeneUtilities.GetRefSeqGeneById(flatRefSeqGenes, linkedEntrezId);
if (refSeqGene == null)
{
AddEnsemblOrphan(ensemblGene);
continue;
}
// merge the Ensembl and RefSeq gene
var mergedGene = MutableGene.Clone(ensemblGene);
mergedGene.TranscriptDataSource = TranscriptDataSource.BothRefSeqAndEnsembl;
UpdateCoordinates(refSeqGene, mergedGene);
if (mergedGene.HgncId == -1 && refSeqGene.HgncId != -1)
{
mergedGene.HgncId = refSeqGene.HgncId;
}
mergedGene.EntrezGeneId = refSeqGene.EntrezGeneId;
_mergedGenes.Add(mergedGene);
refSeqGene.Invalid = true;
ensemblGene.Invalid = true;
_numMergedGenes++;
}
// add the unused RefSeq genes
foreach (var refSeqGene in flatRefSeqGenes)
{
if (refSeqGene.Invalid)
{
continue;
}
AddRefSeqOrphan(refSeqGene);
}
}
private static void UpdateCoordinates(MutableGene source, MutableGene dest)
{
if (source.Start < dest.Start)
{
dest.Start = source.Start;
}
if (source.End > dest.End)
{
dest.End = source.End;
}
}
private static bool GeneEquals(MutableGene x, MutableGene y)
{
return(x.Chromosome.Index == y.Chromosome.Index &&
x.Start == y.Start &&
x.End == y.End &&
x.OnReverseStrand == y.OnReverseStrand &&
x.GeneId == y.GeneId &&
x.Symbol == y.Symbol &&
x.HgncId == y.HgncId &&
x.SymbolSource == y.SymbolSource);
}
private static UgaGene GetMergedGene(MutableGene geneA, MutableGene geneB, bool isGrch37)
{
(MutableGene ensemblGene, MutableGene refSeqGene) = geneA.GeneId.StartsWith("ENSG") ? (geneA, geneB) : (geneB, geneA);
if (ensemblGene.Chromosome.Index != refSeqGene.Chromosome.Index)
{
throw new InvalidDataException($"The two genes are on different chromosomes: {geneA.GeneId} & {geneB.GeneId}");
}
if (ensemblGene.OnReverseStrand != refSeqGene.OnReverseStrand)
{
throw new InvalidDataException($"Both genes do not have the same orientation: {geneA.GeneId} & {geneB.GeneId}");
}
IInterval interval = GetMergedInterval(ensemblGene, refSeqGene);
(IInterval grch37, IInterval grch38) = isGrch37 ? (interval, null as IInterval) : (null as IInterval, interval);
return(new UgaGene(ensemblGene.Chromosome, grch37, grch38, ensemblGene.OnReverseStrand, refSeqGene.GeneId,
ensemblGene.GeneId, ensemblGene.Symbol, ensemblGene.HgncId));
}
private static string GetGeneKey(MutableGene gene) => gene.GeneId + '|' + gene.Chromosome.UcscName + '|' + gene.Start + '|' + gene.End + '|' + (gene.OnReverseStrand ? 'R' : 'F');
private static void WriteGene(TextWriter writer, MutableGene gene)
{
char strand = gene.OnReverseStrand ? 'R' : 'F';
writer.WriteLine($"Gene\t{gene.GeneId}\t{gene.Chromosome.UcscName}\t{gene.Chromosome.Index}\t{gene.Start}\t{gene.End}\t{strand}\t{gene.Symbol}\t{(int)gene.SymbolSource}\t{gene.HgncId}");
}
private void AddRefSeqOrphan(MutableGene gene)
{
_mergedGenes.Add(gene);
gene.Invalid = true;
_numOrphanRefSeqGenes++;
}
private void AddEnsemblOrphan(MutableGene gene)
{
_mergedGenes.Add(gene);
gene.Invalid = true;
_numOrphanEnsemblGenes++;
}
private static IInterval GetMergedInterval(MutableGene geneA, MutableGene geneB) => new Interval(Math.Min(geneA.Start, geneB.Start), Math.Max(geneA.End, geneB.End));
/// <summary>
/// parses the relevant data from each transcript
/// </summary>
public static MutableTranscript Parse(ObjectValueNode objectValue, IChromosome chromosome, Source source)
{
// IDs
string transcriptId = null;
byte transcriptVersion = 1;
string proteinId = null;
byte proteinVersion = 0;
string ccdsId = null;
string refSeqId = null;
string geneId = null;
int hgncId = -1;
// gene
int geneStart = -1;
int geneEnd = -1;
var geneOnReverseStrand = false;
string geneSymbol = null;
var geneSymbolSource = GeneSymbolSource.Unknown;
// translation
int translationStart = -1;
int translationEnd = -1;
MutableExon translationStartExon = null;
MutableExon translationEndExon = null;
// predictions
string siftData = null;
string polyphenData = null;
var bioType = BioType.other;
IInterval[] microRnas = null;
MutableTranscriptRegion[] cdnaMaps = null;
IInterval[] introns = null;
string peptideSequence = null;
string translateableSequence = null;
var isCanonical = false;
int compDnaCodingStart = -1;
int compDnaCodingEnd = -1;
int start = -1;
int end = -1;
MutableExon[] exons = null;
var cdsStartNotFound = false;
var cdsEndNotFound = false;
int[] selenocysteinePositions = null;
IRnaEdit[] rnaEdits = null;
string bamEditStatus = null;
foreach (var node in objectValue.Values)
{
// sanity check: make sure we know about the keys are used for
if (!KnownKeys.Contains(node.Key))
{
throw new InvalidDataException($"Encountered an unknown key in the dumper transcript object: {node.Key}");
}
// handle each key
switch (node.Key)
{
case ImportKeys.CodingRegionEnd:
case ImportKeys.CodingRegionStart:
case ImportKeys.CreatedDate:
case ImportKeys.DbId:
case ImportKeys.Description:
case ImportKeys.DisplayXref:
case ImportKeys.ExternalDb:
case ImportKeys.ExternalDisplayName:
case ImportKeys.ExternalName:
case ImportKeys.ExternalStatus:
case ImportKeys.GenePhenotype:
case ImportKeys.GeneStableId:
case ImportKeys.ModifiedDate:
case ImportKeys.Protein:
case ImportKeys.Slice:
case ImportKeys.Source:
case ImportKeys.Strand:
case ImportKeys.SwissProt:
case ImportKeys.Trembl:
case ImportKeys.UniParc:
case ImportKeys.VepLazyLoaded:
// not used
break;
case ImportKeys.BamEditStatus:
bamEditStatus = node.GetString();
break;
case ImportKeys.Attributes:
(microRnas, rnaEdits, cdsStartNotFound, cdsEndNotFound) = Attribute.ParseList(node);
break;
case ImportKeys.Biotype:
bioType = TranscriptUtilities.GetBiotype(node);
break;
case ImportKeys.Ccds:
ccdsId = node.GetString();
break;
case ImportKeys.CdnaCodingEnd:
compDnaCodingEnd = node.GetInt32();
break;
case ImportKeys.CdnaCodingStart:
compDnaCodingStart = node.GetInt32();
break;
case ImportKeys.End:
end = node.GetInt32();
break;
case ImportKeys.GeneHgncId:
hgncId = node.GetHgncId();
break;
case ImportKeys.GeneSymbol:
case ImportKeys.GeneHgnc: // older key
geneSymbol = node.GetString();
break;
case ImportKeys.GeneSymbolSource:
geneSymbolSource = GeneSymbolSourceHelper.GetGeneSymbolSource(node.GetString());
break;
case ImportKeys.Gene:
(geneStart, geneEnd, geneId, geneOnReverseStrand) = ImportGene.Parse(node);
break;
case ImportKeys.IsCanonical:
isCanonical = node.GetBool();
break;
case ImportKeys.Refseq:
refSeqId = node.GetString();
break;
case ImportKeys.StableId:
transcriptId = node.GetString();
break;
case ImportKeys.Start:
start = node.GetInt32();
break;
case ImportKeys.TransExonArray:
exons = ImportExon.ParseList(node, chromosome);
break;
case ImportKeys.Translation:
(translationStart, translationEnd, proteinId, proteinVersion, translationStartExon, translationEndExon) = ImportTranslation.Parse(node, chromosome);
break;
case ImportKeys.VariationEffectFeatureCache:
(cdnaMaps, introns, peptideSequence, translateableSequence, siftData, polyphenData, selenocysteinePositions) = ImportVariantEffectFeatureCache.Parse(node);
break;
case ImportKeys.Version:
transcriptVersion = (byte)node.GetInt32();
break;
default:
throw new InvalidDataException($"Unknown key found: {node.Key}");
}
}
var fixedTranscript = AccessionUtilities.GetMaxVersion(transcriptId, transcriptVersion);
var fixedProtein = AccessionUtilities.GetMaxVersion(proteinId, proteinVersion);
var gene = new MutableGene(chromosome, geneStart, geneEnd, geneOnReverseStrand, geneSymbol,
geneSymbolSource, geneId, hgncId);
var codingRegion = new CodingRegion(GetCodingRegionStart(geneOnReverseStrand, translationStartExon, translationEndExon, translationStart, translationEnd),
GetCodingRegionEnd(geneOnReverseStrand, translationStartExon, translationEndExon, translationStart, translationEnd),
compDnaCodingStart, compDnaCodingEnd, 0);
int totalExonLength = GetTotalExonLength(exons);
int startExonPhase = translationStartExon?.Phase ?? int.MinValue;
return(new MutableTranscript(chromosome, start, end, fixedTranscript.Id, fixedTranscript.Version, ccdsId,
refSeqId, bioType, isCanonical, codingRegion, fixedProtein.Id, fixedProtein.Version,
peptideSequence, source, gene, exons, startExonPhase, totalExonLength, introns, cdnaMaps,
siftData, polyphenData, translateableSequence, microRnas, cdsStartNotFound, cdsEndNotFound,
selenocysteinePositions, rnaEdits, bamEditStatus));
}
