/// <summary>
/// Creates recalibration table and recalibrates reads.
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="genomeFasta"></param>
/// <param name="bam"></param>
/// <param name="recalibrationTablePath"></param>
/// <param name="knownSitesVcf"></param>
public List <string> BaseRecalibration(string spritzDirectory, string analysisDirectory, string genomeFasta, string bam, string knownSitesVcf)
{
RecalibrationTablePath = Path.Combine(Path.GetDirectoryName(bam), Path.GetFileNameWithoutExtension(bam) + ".recaltable");
RecalibratedBamPath = Path.Combine(Path.GetDirectoryName(bam), Path.GetFileNameWithoutExtension(bam) + ".recal.bam");
return(new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
SamtoolsWrapper.GenomeFastaIndexCommand(genomeFasta),
GenomeDictionaryIndexCommand(genomeFasta),
// check that reference VCF is indexed
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(knownSitesVcf) + ".idx ]; then " + Gatk(Workers) + " IndexFeatureFile -F " + WrapperUtility.ConvertWindowsPath(knownSitesVcf) + "; fi",
"if [[ ( ! -f " + WrapperUtility.ConvertWindowsPath(RecalibrationTablePath) + " || ! -s " + WrapperUtility.ConvertWindowsPath(RecalibrationTablePath) + " ) ]]; then " +
Gatk(Workers) +
" BaseRecalibrator" +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -I " + WrapperUtility.ConvertWindowsPath(bam) +
(knownSitesVcf != "" ? " --known-sites " + WrapperUtility.ConvertWindowsPath(knownSitesVcf) : "") +
" -O " + WrapperUtility.ConvertWindowsPath(RecalibrationTablePath) +
"; fi",
"if [[ ( ! -f " + WrapperUtility.ConvertWindowsPath(RecalibratedBamPath) + " || ! -s " + WrapperUtility.ConvertWindowsPath(RecalibratedBamPath) + " ) ]]; then " +
Gatk(Workers) +
" ApplyBQSR" +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -I " + WrapperUtility.ConvertWindowsPath(bam) +
" --bqsr-recal-file " + WrapperUtility.ConvertWindowsPath(RecalibrationTablePath) +
" -O " + WrapperUtility.ConvertWindowsPath(RecalibratedBamPath) +
"; fi",
SamtoolsWrapper.IndexBamCommand(RecalibratedBamPath),
});
}
/// <summary>
/// Groups (I'm just using one group, so it's more a formality) and sorts reads. Marks duplicates.
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="threads"></param>
/// <param name="bam"></param>
/// <param name="genomeFasta"></param>
/// <param name="reference"></param>
/// <param name="newBam"></param>
/// <param name="convertToUCSC"></param>
public List <string> PrepareBamAndFasta(string spritzDirectory, string analysisDirectory, int threads, string bam, string genomeFasta, string reference)
{
string sortedCheckPath = Path.Combine(Path.GetDirectoryName(bam), Path.GetFileNameWithoutExtension(bam) + ".headerSorted");
string readGroupedCheckfile = Path.Combine(Path.GetDirectoryName(bam), Path.GetFileNameWithoutExtension(bam) + ".headerReadGrouped");
string sortedBam = Path.Combine(Path.GetDirectoryName(bam), Path.GetFileNameWithoutExtension(bam) + ".sorted.bam");
string groupedBamPath = Path.Combine(Path.GetDirectoryName(sortedBam), Path.GetFileNameWithoutExtension(sortedBam) + ".grouped.bam");
string markedDuplicatesBamPath = Path.Combine(Path.GetDirectoryName(groupedBamPath), Path.GetFileNameWithoutExtension(groupedBamPath) + ".marked.bam");
string markedDuplicateMetrics = Path.Combine(Path.GetDirectoryName(groupedBamPath), Path.GetFileNameWithoutExtension(groupedBamPath) + ".marked.metrics");
string tmpDir = Path.Combine(spritzDirectory, "tmp");
Directory.CreateDirectory(tmpDir);
List <string> commands = new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
SamtoolsWrapper.GenomeFastaIndexCommand(genomeFasta),
GenomeDictionaryIndexCommand(genomeFasta),
"samtools view -H " + WrapperUtility.ConvertWindowsPath(bam) + " | grep SO:coordinate > " + WrapperUtility.ConvertWindowsPath(sortedCheckPath),
"samtools view -H " + WrapperUtility.ConvertWindowsPath(bam) + " | grep '^@RG' > " + WrapperUtility.ConvertWindowsPath(readGroupedCheckfile),
// group and sort (note, using picard-tools works, but picard.jar somehow is trucating the BAM files)
"if [[ ( ! -f " + WrapperUtility.ConvertWindowsPath(groupedBamPath) + " || ! -s " + WrapperUtility.ConvertWindowsPath(groupedBamPath) + " ) && " +
" ( ! -f " + WrapperUtility.ConvertWindowsPath(markedDuplicatesBamPath) + " || ! -s " + WrapperUtility.ConvertWindowsPath(markedDuplicatesBamPath) + " ) ]]; then " +
Gatk(Workers, 2) +
" AddOrReplaceReadGroups" +
" -PU platform -PL illumina -SM sample -LB library" +
" -I " + WrapperUtility.ConvertWindowsPath(bam) +
" -O " + WrapperUtility.ConvertWindowsPath(groupedBamPath) +
" -SO coordinate" +
" --TMP_DIR " + WrapperUtility.ConvertWindowsPath(tmpDir) +
"; fi",
SamtoolsWrapper.IndexBamCommand(groupedBamPath),
// mark duplicates (AS means assume sorted; note, using picard-tools works, but picard.jar somehow is trucating the BAM files)
"if [[ ( ! -f " + WrapperUtility.ConvertWindowsPath(markedDuplicatesBamPath) + " || ! -s " + WrapperUtility.ConvertWindowsPath(markedDuplicatesBamPath) + " ) ]]; then " +
Gatk(Workers) +
" MarkDuplicates" + // formerly picard
" -I " + WrapperUtility.ConvertWindowsPath(groupedBamPath) +
" -O " + WrapperUtility.ConvertWindowsPath(markedDuplicatesBamPath) +
" -M " + WrapperUtility.ConvertWindowsPath(markedDuplicateMetrics) +
" --TMP_DIR " + WrapperUtility.ConvertWindowsPath(tmpDir) +
" -AS true" +
"; fi",
SamtoolsWrapper.IndexBamCommand(markedDuplicatesBamPath),
// clean up
"if [[ ( ! -f " + WrapperUtility.ConvertWindowsPath(markedDuplicatesBamPath) + " || ! -s " + WrapperUtility.ConvertWindowsPath(markedDuplicatesBamPath) + " ) ]]; then " +
"rm " + WrapperUtility.ConvertWindowsPath(groupedBamPath) + "; fi",
};
PreparedBamPath = markedDuplicatesBamPath;
return(commands);
}
/// <summary>
/// HaplotypeCaller for calling variants on each RNA-Seq BAM file individually.
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="threads"></param>
/// <param name="genomeFasta"></param>
/// <param name="splitTrimBam"></param>
/// <param name="dbsnpReferenceVcfPath"></param>
/// <param name="newVcf"></param>
public List <string> VariantCalling(string spritzDirectory, ExperimentType experimentType, int threads, string genomeFasta, string splitTrimBam, string dbsnpReferenceVcfPath)
{
HaplotypeCallerGvcfPath = Path.Combine(Path.GetDirectoryName(splitTrimBam), Path.GetFileNameWithoutExtension(splitTrimBam) + ".g.vcf.gz");
HaplotypeCallerVcfPath = Path.Combine(Path.GetDirectoryName(splitTrimBam), Path.GetFileNameWithoutExtension(splitTrimBam) + ".g.gt.vcf");
FilteredHaplotypeCallerVcfPath = Path.Combine(Path.GetDirectoryName(splitTrimBam), Path.GetFileNameWithoutExtension(splitTrimBam) + ".g.gt.NoIndels.vcf");
var vcftools = new VcfToolsWrapper();
List <string> commands = new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
SamtoolsWrapper.GenomeFastaIndexCommand(genomeFasta),
GenomeDictionaryIndexCommand(genomeFasta),
// check that reference VCF is indexed
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(dbsnpReferenceVcfPath) + ".idx ]; then " + Gatk(Workers) + " IndexFeatureFile -F " + WrapperUtility.ConvertWindowsPath(dbsnpReferenceVcfPath) + "; fi",
// call variants
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) + " ] || [ " + " ! -s " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) + " ]; then " +
Gatk(Workers, 2) +
" HaplotypeCaller" +
" --native-pair-hmm-threads " + threads.ToString() +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -I " + WrapperUtility.ConvertWindowsPath(splitTrimBam) +
" --min-base-quality-score 20" +
(experimentType == ExperimentType.RNASequencing ? " --dont-use-soft-clipped-bases true" : "") +
" --dbsnp " + WrapperUtility.ConvertWindowsPath(dbsnpReferenceVcfPath) +
" -O " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) +
" -ERC GVCF" + // this prompts phasing!
" --max-mnp-distance 3" + // note: this can't be used for joint genotyping here, but this setting is available in mutect2 for doing tumor vs normal calls
"; fi",
// index compressed gvcf file
$"if [ ! -f {WrapperUtility.ConvertWindowsPath($"{HaplotypeCallerGvcfPath}.tbi")} ]; then {Gatk(Workers)} IndexFeatureFile -F {WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath)}; fi",
// genotype the gvcf file into a traditional vcf file
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) + " ] || [ " + " ! -s " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) + " ]; then " +
Gatk(Workers, 2) +
" GenotypeGVCFs" +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -V " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) +
" -O " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) +
"; fi",
$"if [ ! -f {WrapperUtility.ConvertWindowsPath($"{HaplotypeCallerVcfPath}.idx")} ]; then {Gatk(Workers)} IndexFeatureFile -F {WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath)}; fi",
// filter out indels
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(FilteredHaplotypeCallerVcfPath) + " ] || [ " + " ! -s " + WrapperUtility.ConvertWindowsPath(FilteredHaplotypeCallerVcfPath) + " ]; then " +
Gatk(Workers, 2) +
" SelectVariants" +
" --select-type-to-exclude INDEL" +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -V " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) +
" -O " + WrapperUtility.ConvertWindowsPath(FilteredHaplotypeCallerVcfPath) +
"; fi",
$"if [ ! -f {WrapperUtility.ConvertWindowsPath($"{FilteredHaplotypeCallerVcfPath}.idx")} ]; then {Gatk(Workers)} IndexFeatureFile -F {WrapperUtility.ConvertWindowsPath(FilteredHaplotypeCallerVcfPath)}; fi",
// filter variants (RNA-Seq specific params... need to check out recommendations before using DNA-Seq)
//"if [ ! -f " + WrapperUtility.ConvertWindowsPath(newVcf) + " ] || [ " + " ! -s " + WrapperUtility.ConvertWindowsPath(newVcf) + " ]; then " +
// Gatk() +
// " -T VariantFiltration" +
// " -nct " + threads.ToString() +
// " -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
// " -V " + WrapperUtility.ConvertWindowsPath(unfliteredVcf) +
// " -window 35 -cluster 3" + // filter out clusters of 3 snps within 35 bases (https://software.broadinstitute.org/gatk/documentation/topic?name=methods)
// " -filterName FS -filter \"FS > 30.0\"" +
// " -filterName QD -filter \"QD < 2.0\"" +
// " -o " + WrapperUtility.ConvertWindowsPath(newVcf) +
// "; fi",
};
return(commands);
}
public List <string> CombineAndGenotypeGvcfs(string spritzDirectory, string genomeFasta, List <string> gvcfPaths)
{
if (gvcfPaths == null || gvcfPaths.Count <= 1)
{
throw new ArgumentException("CombineAndGenotypeGvcfs exception: no gvcfs were specified to combine");
}
int uniqueSuffix = 1;
foreach (string f in gvcfPaths)
{
uniqueSuffix = uniqueSuffix ^ f.GetHashCode();
}
HaplotypeCallerGvcfPath = Path.Combine(Path.GetDirectoryName(gvcfPaths.First()), $"combined{uniqueSuffix}.g.vcf.gz");
HaplotypeCallerVcfPath = Path.Combine(Path.GetDirectoryName(HaplotypeCallerGvcfPath), $"{Path.GetFileNameWithoutExtension(Path.GetFileNameWithoutExtension(HaplotypeCallerGvcfPath))}.gt.vcf");
FilteredHaplotypeCallerVcfPath = Path.Combine(Path.GetDirectoryName(HaplotypeCallerGvcfPath), $"{Path.GetFileNameWithoutExtension(HaplotypeCallerGvcfPath)}.NoIndels.vcf");
List <string> commands = new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
SamtoolsWrapper.GenomeFastaIndexCommand(genomeFasta),
GenomeDictionaryIndexCommand(genomeFasta)
};
foreach (string gvcf in gvcfPaths)
{
// double check that the compressed gvcf file is indexed
commands.Add($"if [ ! -f {WrapperUtility.ConvertWindowsPath(gvcf)}.idx ]; then {Gatk(Workers)} IndexFeatureFile -F {WrapperUtility.ConvertWindowsPath(gvcf)}; fi");
}
// combine GVCFs
string combineCommand =
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) + " ] || [ " + " ! -s " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) + " ]; then " +
Gatk(Workers) +
" CombineGVCFs" +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -V " + string.Join(" -V ", gvcfPaths.Select(gvcf => WrapperUtility.ConvertWindowsPath(gvcf))) +
" -O " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) +
"; fi";
commands.Add(combineCommand);
commands.Add($"if [ ! -f {WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath)}.idx ]; then {Gatk(Workers)} IndexFeatureFile -F {WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath)}; fi");
// genotype the gvcf file into a traditional vcf file
string genotypeCommand =
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) + " ] || [ " + " ! -s " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) + " ]; then " +
Gatk(Workers) +
" GenotypeGVCFs" +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -V " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerGvcfPath) +
" -O " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) +
"; fi";
commands.Add(genotypeCommand);
// filter out indels
string filterIndelsCommand =
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) + " ] || [ " + " ! -s " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) + " ]; then " +
Gatk(Workers) +
" SelectVariants" +
" --select-type-to-exclude INDEL" +
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -V " + WrapperUtility.ConvertWindowsPath(HaplotypeCallerVcfPath) +
" -O " + WrapperUtility.ConvertWindowsPath(FilteredHaplotypeCallerVcfPath) +
"; fi";
commands.Add(filterIndelsCommand);
return(commands);
}
/// <summary>
/// Splits and trims reads splice junction reads with SplitNCigarReads.
/// Apparently cigars are genomic intervals, and splice junctions are represented by a bunch of N's (unkonwn nucleotide), HaplotypeCaller requires splitting them in the BAM file.
///
/// It's tempting to want to run a few of these at the same time because it's not well parallelized. It's just not worth it. It uses quite a bit of RAM and racks the I/O at the beginning when reading the BAM files.
/// Could possibly do 4 at a time on 128 GB RAM and 28 processors.
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="genomeFasta"></param>
/// <param name="dedupedBam"></param>
/// <param name="splitTrimBam"></param>
/// <returns></returns>
public List <string> SplitNCigarReads(string spritzDirectory, string genomeFasta, string dedupedBam)
{
string fixedQualsBam = Path.Combine(Path.GetDirectoryName(dedupedBam), Path.GetFileNameWithoutExtension(dedupedBam) + ".fixedQuals.bam");
SplitTrimBamPath = Path.Combine(Path.GetDirectoryName(fixedQualsBam), Path.GetFileNameWithoutExtension(fixedQualsBam) + ".split.bam");
// This also filters malformed reads
string fixMisencodedQualsCmd =
Gatk(Workers) +
" FixMisencodedBaseQualityReads" +
" -I " + WrapperUtility.ConvertWindowsPath(dedupedBam) +
" -O " + WrapperUtility.ConvertWindowsPath(fixedQualsBam);
string splitNCigarReadsCmd1 =
Gatk(Workers) +
" SplitNCigarReads" +
//" --num_threads " + threads.ToString() + // not supported
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -I " + WrapperUtility.ConvertWindowsPath(fixedQualsBam) +
" -O " + WrapperUtility.ConvertWindowsPath(SplitTrimBamPath)
//" -rf ReassignOneMappingQuality" + // doing this with STAR
//" -RMQF 255" +
//" -RMQT 60" + // default mapping quality is 60; required for RNA-Seq aligners
//" -U ALLOW_N_CIGAR_READS"
;
string splitNCigarReadsCmd2 =
Gatk(Workers) +
" SplitNCigarReads" +
//" --num_threads " + threads.ToString() + // not supported
" -R " + WrapperUtility.ConvertWindowsPath(genomeFasta) +
" -I " + WrapperUtility.ConvertWindowsPath(dedupedBam) +
" -O " + WrapperUtility.ConvertWindowsPath(SplitTrimBamPath)
//" -rf ReassignOneMappingQuality" + // doing this with STAR
//" -RMQF 255" +
//" -RMQT 60" + // default mapping quality is 60; required for RNA-Seq aligners
//" -U ALLOW_N_CIGAR_READS"
;
List <string> commands = new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
SamtoolsWrapper.GenomeFastaIndexCommand(genomeFasta),
GenomeDictionaryIndexCommand(genomeFasta),
// split and trim reads (some datasets are probably going to have misencoded quality scores; -fixMisencodedQuals just subtracts 31 from all quality scores if possible...)
// exit code of 2 means that the FixMisencodedQualityBaseReads errored out because there were correctly encode base quality scores
SamtoolsWrapper.IndexBamCommand(dedupedBam),
"if [[ ( ! -f " + WrapperUtility.ConvertWindowsPath(fixedQualsBam) + " || ! -s " + WrapperUtility.ConvertWindowsPath(fixedQualsBam) + " ) ]]; then",
" " + fixMisencodedQualsCmd,
" if [ $? -ne 2 ]; then",
" " + splitNCigarReadsCmd1,
" else",
" " + splitNCigarReadsCmd2,
" fi",
"fi",
SamtoolsWrapper.IndexBamCommand(SplitTrimBamPath),
};
return(commands);
}