/// <summary>
/// Internal code to determine the type of the genotype from the alleles vector </summary>
/// <returns> the type </returns>
protected internal virtual GenotypeType determineType() // we should never call if already calculated
{
// TODO -- this code is slow and could be optimized for the diploid case
IList <Allele> alleles = Alleles;
if (alleles.Count == 0)
{
return(GenotypeType.UNAVAILABLE);
}
bool sawNoCall = false, sawMultipleAlleles = false;
Allele observedAllele = null;
foreach (Allele allele in alleles)
{
if (allele.NoCall)
{
sawNoCall = true;
}
else if (observedAllele == null)
{
observedAllele = allele;
}
else if (!allele.Equals(observedAllele))
{
sawMultipleAlleles = true;
}
}
if (sawNoCall)
{
if (observedAllele == null)
{
return(GenotypeType.NO_CALL);
}
return(GenotypeType.MIXED);
}
if (observedAllele == null)
{
throw new Exception("BUG: there are no alleles present in this genotype but the alleles list is not null");
}
return(sawMultipleAlleles ? GenotypeType.HET : observedAllele.Reference?GenotypeType.HOM_REF : GenotypeType.HOM_VAR);
}
private static VariantType typeOfBiallelicVariant(Allele reference, Allele allele)
{
if (reference.Symbolic)
{
throw new Exception("Unexpected error: encountered a record with a symbolic reference allele");
}
if (allele.Symbolic)
{
return(VariantType.SYMBOLIC);
}
if (reference.Length == allele.Length)
{
if (allele.Length == 1)
{
return(VariantType.SNP);
}
else
{
return(VariantType.MNP);
}
}
// Important note: previously we were checking that one allele is the prefix of the other. However, that's not an
// appropriate check as can be seen from the following example:
// REF = CTTA and ALT = C,CT,CA
// This should be assigned the INDEL type but was being marked as a MIXED type because of the prefix check.
// In truth, it should be absolutely impossible to return a MIXED type from this method because it simply
// performs a pairwise comparison of a single alternate allele against the reference allele (whereas the MIXED type
// is reserved for cases of multiple alternate alleles of different types). Therefore, if we've reached this point
// in the code (so we're not a SNP, MNP, or symbolic allele), we absolutely must be an INDEL.
return(VariantType.INDEL);
// old incorrect logic:
// if (oneIsPrefixOfOther(ref, allele))
// return Type.INDEL;
// else
// return Type.MIXED;
}
/// <summary>
/// Compute the end position for this VariantContext from the alleles themselves
///
/// In the trivial case this is a single BP event and end = start (open intervals)
/// In general the end is start + ref length - 1, handling the case where ref length == 0
/// However, if alleles contains a symbolic allele then we use endForSymbolicAllele in all cases
/// </summary>
/// <param name="alleles"> the list of alleles to consider. The reference allele must be the first one </param>
/// <param name="start"> the known start position of this event </param>
/// <param name="endForSymbolicAlleles"> the end position to use if any of the alleles is symbolic. Can be -1
/// if no is expected but will throw an error if one is found </param>
/// <returns> this builder </returns>
public static int ComputeEndFromAlleles(IList <Allele> alleles, int start, int endForSymbolicAlleles)
{
Allele reference = alleles [0];
if (reference.NonReference)
{
throw new Exception("computeEndFromAlleles requires first allele to be reference");
}
if (VariantContext.HasSymbolicAlleles(alleles))
{
if (endForSymbolicAlleles == -1)
{
throw new Exception("computeEndFromAlleles found a symbolic allele but endForSymbolicAlleles was provided");
}
return(endForSymbolicAlleles);
}
else
{
return(start + Math.Max(reference.Length - 1, 0));
}
}
/// <summary>
/// check to make sure the allele is an acceptable allele </summary>
/// <param name="allele"> the allele to check </param>
/// <param name="isRef"> are we the reference allele? </param>
/// <param name="lineNo"> the line number for this record </param>
private static void checkAllele(string allele, bool isRef, int lineNo)
{
if (allele == null || allele.Length == 0)
{
generateException("Empty alleles are not permitted in VCF records", lineNo);
}
if (MAX_ALLELE_SIZE_BEFORE_WARNING != -1 && allele.Length > MAX_ALLELE_SIZE_BEFORE_WARNING)
{
throw new VCFParsingError(string.Format("Allele detected with length {0:D} exceeding max size {1:D} at approximately line {2:D}, likely resulting in degraded VCF processing performance", allele.Length, MAX_ALLELE_SIZE_BEFORE_WARNING, lineNo));
}
if (isSymbolicAllele(allele))
{
if (isRef)
{
generateException("Symbolic alleles not allowed as reference allele: " + allele, lineNo);
}
}
else
{
// check for VCF3 insertions or deletions
if ((allele [0] == VCFConstants.DELETION_ALLELE_v3) || (allele [0] == VCFConstants.INSERTION_ALLELE_v3))
{
generateException("Insertions/Deletions are not supported when reading 3.x VCF's. Please" + " convert your file to VCF4 using VCFTools, available at http://vcftools.sourceforge.net/index.html", lineNo);
}
if (!Allele.AcceptableAlleleBases(allele))
{
generateException("Unparsable vcf record with allele " + allele, lineNo);
}
if (isRef && allele.Equals(VCFConstants.EMPTY_ALLELE))
{
generateException("The reference allele cannot be missing", lineNo);
}
}
}
/// <summary>
/// parse out the alleles </summary>
/// <param name="reference"> the reference base </param>
/// <param name="alts"> a string of alternates to break into alleles </param>
/// <param name="lineNo"> the line number for this record </param>
/// <returns> a list of alleles, and a pair of the shortest and longest sequence </returns>
protected internal static IList <Allele> parseAlleles(string reference, string alts, int lineNo)
{
IList <Allele> alleles = new List <Allele> (2); // we are almost always biallelic
// ref
checkAllele(reference, true, lineNo);
Allele refAllele = Allele.Create(reference, true);
alleles.Add(refAllele);
if (alts.IndexOf(",") == -1) // only 1 alternatives, don't call string split
{
parseSingleAltAllele(alleles, alts, lineNo);
}
else
{
foreach (string alt in alts.Split(VCFConstants.COMMA_AS_CHAR_ARRAY, StringSplitOptions.RemoveEmptyEntries))
{
parseSingleAltAllele(alleles, alt, lineNo);
}
}
return(alleles);
}
/// <summary>
/// Add a record to the file
/// </summary>
/// <param name="vc">The Variant Context object </param>
protected string getVariantLinetoWrite(VariantContext vc)
{
if (doNotWriteGenotypes)
{
vc = (new VariantContextBuilder(vc)).noGenotypes().make();
}
try
{
//Convert alleles to 1,2,3,etc. numbering
IDictionary <Allele, string> alleleMap = buildAlleleMap(vc);
// CHROM
StringBuilder lineToWrite = new StringBuilder();
//Add chr, pos, id, ref
lineToWrite.Append(String.Join(VCFConstants.FIELD_SEPARATOR, vc.Chr, vc.Start.ToString(), vc.ID, vc.Reference.DisplayString));
// ALT
if (vc.Variant)
{
Allele altAllele = vc.GetAlternateAllele(0);
string alt = altAllele.DisplayString;
lineToWrite.Append(alt);
for (int i = 1; i < vc.AlternateAlleles.Count; i++)
{
altAllele = vc.GetAlternateAllele(i);
alt = altAllele.DisplayString;
lineToWrite.Append(",");
lineToWrite.Append(alt);
}
}
else
{
lineToWrite.Append(VCFConstants.EMPTY_ALTERNATE_ALLELE_FIELD);
}
lineToWrite.Append(VCFConstants.FIELD_SEPARATOR);
// QUAL
if (!vc.HasLog10PError)
{
lineToWrite.Append(VCFConstants.MISSING_VALUE_v4);
}
else
{
lineToWrite.Append(formatQualValue(vc.PhredScaledQual));
}
lineToWrite.Append(VCFConstants.FIELD_SEPARATOR);
// FILTER
string filters = getFilterString(vc);
lineToWrite.Append(filters);
lineToWrite.Append(VCFConstants.FIELD_SEPARATOR);
// INFO
IDictionary <string, string> infoFields = new SortedDictionary <string, string>();
foreach (KeyValuePair <string, object> field in vc.Attributes)
{
string key = field.Key;
if (!mHeader.hasInfoLine(key))
{
fieldIsMissingFromHeaderError(vc, key, "INFO");
}
string outputValue = formatVCFField(field.Value);
if (outputValue != null)
{
infoFields[key] = outputValue;
}
}
lineToWrite.Append(getInfoString(infoFields));;
// FORMAT
GenotypesContext gc = vc.Genotypes;
if (gc.LazyWithData && ((LazyGenotypesContext)gc).UnparsedGenotypeData is string)
{
lineToWrite.Append(VCFConstants.FIELD_SEPARATOR);
lineToWrite.Append(((LazyGenotypesContext)gc).UnparsedGenotypeData.ToString());
}
else
{
IList <string> genotypeAttributeKeys = calcVCFGenotypeKeys(vc);
if (genotypeAttributeKeys.Count > 0)
{
foreach (String format in genotypeAttributeKeys)
{
if (!mHeader.hasFormatLine(format))
{
fieldIsMissingFromHeaderError(vc, format, "FORMAT");
}
}
string genotypeFormatString = String.Join(VCFConstants.GENOTYPE_FIELD_SEPARATOR, genotypeAttributeKeys);
lineToWrite.Append(VCFConstants.FIELD_SEPARATOR);
lineToWrite.Append(genotypeFormatString);
lineToWrite.Append(getGenotypeDataText(vc, alleleMap, genotypeAttributeKeys));
}
}
lineToWrite.Append("\n");
return(lineToWrite.ToString());
}
catch (IOException e)
{
throw new Exception("Unable to write the VCF object:\n " + vc.ToString() + "\n", e);
}
}
/// <summary>
/// Returns how many times allele appears in this genotype object?
/// </summary>
/// <param name="allele"> </param>
/// <returns> a value >= 0 indicating how many times the allele occurred in this sample's genotype </returns>
public virtual int CountAlleles(Allele allele)
{
return(Alleles.Count(x => x.Equals(allele)));
}