public VCFHeader(ISet <VCFHeaderLine> metaData, IList <string> genotypeSampleNames) : this(metaData)
{
if (genotypeSampleNames.Count != (new HashSet <string>(genotypeSampleNames)).Count)
{
throw new VCFParsingError("BUG: VCF header has duplicate sample names");
}
GenotypeSampleNames.AddRange(genotypeSampleNames);
SamplesWereAlreadySorted = ParsingUtils.IsSorted(genotypeSampleNames);
buildVCFReaderMaps(genotypeSampleNames);
}
/// <summary>
/// Return a VCF-like string representation for the alleles of this genotype.
///
/// If ignoreRefState is true, will not append the reference * marker on the alleles.
/// </summary>
/// <returns> a string representing the genotypes, or null if the type is unavailable. </returns>
public virtual string getGenotypeString(bool ignoreRefState)
{
if (Ploidy == 0)
{
return("NA");
}
// Notes:
// 1. Make sure to use the appropriate separator depending on whether the genotype is phased
// 2. If ignoreRefState is true, then we want just the bases of the Alleles (ignoring the '*' indicating a ref Allele)
// 3. So that everything is deterministic with regards to integration tests, we sort Alleles (when the genotype isn't phased, of course)
return(String.Join(Phased ? PHASED_ALLELE_SEPARATOR : UNPHASED_ALLELE_SEPARATOR, ignoreRefState ? AlleleStrings :
(Phased ? Alleles.Select(x => x.ToString()) : ParsingUtils.SortList(Alleles).Select(x => x.ToString()))));
}
// --------------------------------------------------------------------------------
//
// implementation functions
//
// --------------------------------------------------------------------------------
private string getFilterString(VariantContext vc)
{
if (vc.Filtered)
{
foreach (String filter in vc.Filters)
{
if (!mHeader.hasFilterLine(filter))
{
fieldIsMissingFromHeaderError(vc, filter, "FILTER");
}
}
return(String.Join(";", ParsingUtils.SortList(vc.Filters.ToList()).ToArray()));
}
else if (vc.FiltersWereApplied)
{
return(VCFConstants.PASSES_FILTERS_v4);
}
else
{
return(VCFConstants.UNFILTERED);
}
}
/// <summary>
/// Determine which genotype fields are in use in the genotypes in VC </summary>
/// <param name="vc"> </param>
/// <returns> an ordered list of genotype fields in use in VC. If vc has genotypes this will always include GT first </returns>
private static IList <string> calcVCFGenotypeKeys(VariantContext vc)
{
//TODO: not sure who wrote this, these boolean flags should be removed though
HashSet <string> keys = new HashSet <string>();
bool sawGoodGT = false;
bool sawGoodQual = false;
bool sawGenotypeFilter = false;
bool sawDP = false;
bool sawAD = false;
bool sawPL = false;
foreach (Genotype g in vc.Genotypes)
{
//todo, make this a string later
foreach (string s in g.ExtendedAttributes.Keys.Select(x => x.ToString()))
{
keys.Add(s);
}
if (g.Available)
{
sawGoodGT = true;
}
if (g.HasGQ)
{
sawGoodQual = true;
}
if (g.HasDP)
{
sawDP = true;
}
if (g.HasAD)
{
sawAD = true;
}
if (g.HasPL)
{
sawPL = true;
}
if (g.Filtered)
{
sawGenotypeFilter = true;
}
}
if (sawGoodQual)
{
}
if (sawDP)
{
keys.Add(VCFConstants.DEPTH_KEY);
}
if (sawAD)
{
keys.Add(VCFConstants.GENOTYPE_ALLELE_DEPTHS);
}
if (sawPL)
{
keys.Add(VCFConstants.GENOTYPE_PL_KEY);
}
if (sawGenotypeFilter)
{
keys.Add(VCFConstants.GENOTYPE_FILTER_KEY);
}
IList <string> sortedList = ParsingUtils.SortList(new List <string>(keys));
// make sure the GT is first
if (sawGoodGT)
{
IList <string> newList = new List <string>(sortedList.Count + 1);
newList.Add(VCFConstants.GENOTYPE_KEY);
foreach (string s in sortedList)
{
newList.Add(s);
}
sortedList = newList;
}
if (sortedList.Count == 0)
{
// this needs to be done in case all samples are no-calls
return(new List <string>()
{
VCFConstants.GENOTYPE_KEY
});
}
else
{
return(sortedList);
}
}