/// <summary>
/// Creates a snpeff model for a custom gene model
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="analysisDirectory"></param>
/// <param name="genomeFastaPath"></param>
/// <param name="geneModelGtfOrGffPath"></param>
/// <returns>Name of the snpEff reference that was generated</returns>
public static string GenerateDatabase(string spritzDirectory, string analysisDirectory, string genomeFastaPath, string referenceProteinFastaPath, string geneModelGtfOrGffPath)
{
string snpEffReferenceName = Path.GetExtension(geneModelGtfOrGffPath).Substring(1).ToUpperInvariant() + geneModelGtfOrGffPath.GetHashCode().ToString();
string snpEffReferenceFolderPath = Path.Combine(spritzDirectory, "Tools", "SnpEff", "data", snpEffReferenceName);
string scriptPath = WrapperUtility.GetAnalysisScriptPath(analysisDirectory, "SnpEffDatabaseGeneration.bash");
string geneModelOption = Path.GetExtension(geneModelGtfOrGffPath).EndsWith("gtf") ? "-gtf22" : "-gff3";
// if the database is already made, don't remake it
if (File.Exists(Path.Combine(spritzDirectory, "Tools", "SnpEff", "data", snpEffReferenceName, "snpEffectPredictor.bin")))
{
return(snpEffReferenceName);
}
WrapperUtility.GenerateAndRunScript(scriptPath, new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
"cd SnpEff",
// create data folder for this reference, and copy the custom gene model (can also copy regulatory annotations)
"mkdir data/" + snpEffReferenceName,
"cp " + WrapperUtility.ConvertWindowsPath(geneModelGtfOrGffPath) + " " + WrapperUtility.ConvertWindowsPath(Path.Combine(snpEffReferenceFolderPath, "genes" + Path.GetExtension(geneModelGtfOrGffPath))),
"cp " + WrapperUtility.ConvertWindowsPath(referenceProteinFastaPath) + " " + WrapperUtility.ConvertWindowsPath(Path.Combine(snpEffReferenceFolderPath, "protein.fa")),
// copy the genome to the genomes folder
"mkdir " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "data", "genomes")),
"cp " + WrapperUtility.ConvertWindowsPath(genomeFastaPath) + " " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "data", "genomes", snpEffReferenceName + ".fa")),
// configure SnpEff for this custom reference
// note: if different organism is used in the future, this becomes pretty complex... probably would list the organisms from snpEff.config in the GUI
"echo \"\n# " + snpEffReferenceName + "\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"" + snpEffReferenceName + ".genome : Homo_sapiens\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"" + snpEffReferenceName + ".reference : ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"\t" + snpEffReferenceName + ".M.codonTable : Vertebrate_Mitochondrial\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"\t" + snpEffReferenceName + ".MT.codonTable : Vertebrate_Mitochondrial\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
// build snpEff model
"cd ..",
SnpEff(1) + " build " + geneModelOption + " -v " + snpEffReferenceName,
}).WaitForExit();
return(snpEffReferenceName);
}
public List <string> PrimaryVariantAnnotation(string spritzDirectory, string reference, string inputVcfPath, bool fromReference = false)
{
string outPrefix = Path.Combine(Path.GetDirectoryName(inputVcfPath), Path.GetFileNameWithoutExtension(inputVcfPath));
AnnotatedVcfPath = outPrefix + ".snpEffAnnotated.vcf";
HtmlReportPath = outPrefix + ".snpEffAnnotated.html";
AnnotatedGenesSummaryPath = outPrefix + ".snpEffAnnotated.genes.txt";
VariantProteinFastaPath = outPrefix + ".snpEffAnnotated.protein.fasta";
VariantProteinXmlPath = outPrefix + ".snpEffAnnotated.protein.xml";
Directory.CreateDirectory(Path.Combine(spritzDirectory, "Tools", "SnpEff", "data"));
string[] existingDatabases = Directory.GetDirectories(Path.Combine(spritzDirectory, "Tools", "SnpEff", "data"));
if (File.Exists(AnnotatedVcfPath) && new FileInfo(AnnotatedVcfPath).Length > 0)
{
return(new List <string>());
}
return(new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
SnpEff(Workers) + " -v -stats " + WrapperUtility.ConvertWindowsPath(HtmlReportPath) +
" -fastaProt " + WrapperUtility.ConvertWindowsPath(VariantProteinFastaPath) +
" -xmlProt " + WrapperUtility.ConvertWindowsPath(VariantProteinXmlPath) +
" " + Path.GetFileName(existingDatabases.FirstOrDefault(x => Path.GetFileName(x).StartsWith(reference, true, null))) +
(fromReference ? "" : $" {WrapperUtility.ConvertWindowsPath(inputVcfPath)} > {WrapperUtility.ConvertWindowsPath(AnnotatedVcfPath)}"),
// ensure that the files get closed before continuing
WrapperUtility.EnsureClosedFileCommands(WrapperUtility.ConvertWindowsPath(AnnotatedVcfPath)),
WrapperUtility.EnsureClosedFileCommands(WrapperUtility.ConvertWindowsPath(VariantProteinFastaPath)),
WrapperUtility.EnsureClosedFileCommands(WrapperUtility.ConvertWindowsPath(VariantProteinXmlPath)),
// remove the annotated VCF file if snpEff didn't work, e.g. if there was no VCF file to annotate
"if [[ ( -f " + WrapperUtility.ConvertWindowsPath(AnnotatedVcfPath) + " && ! -s " + WrapperUtility.ConvertWindowsPath(AnnotatedVcfPath) + " ) ]]; then",
" rm " + WrapperUtility.ConvertWindowsPath(AnnotatedVcfPath),
"fi",
});
}
/// <summary>
/// Aligns reads in fastq files using TopHat2.
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="bowtieIndexPrefix"></param>
/// <param name="threads"></param>
/// <param name="fastqPaths"></param>
/// <param name="geneModelGtfOrGffPath"></param>
/// <param name="strandSpecific"></param>
/// <param name="outputDirectory"></param>
public static void Align(string spritzDirectory, string analysisDirectory, string bowtieIndexPrefix, int threads, string[] fastqPaths,
bool strandSpecific, out string outputDirectory)
{
string tempDir = Path.Combine(Path.GetDirectoryName(fastqPaths[0]), "tmpDir");
outputDirectory = Path.Combine(Path.GetDirectoryName(fastqPaths[0]), Path.GetFileNameWithoutExtension(fastqPaths[0]) + "TophatOut");
Directory.CreateDirectory(tempDir);
WrapperUtility.GenerateAndRunScript(WrapperUtility.GetAnalysisScriptPath(analysisDirectory, "TophatRun.bash"), new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
"tophat-2.1.1/tophat2" +
" --num-threads " + threads.ToString() +
" --output-dir " + WrapperUtility.ConvertWindowsPath(outputDirectory) +
//" --GTF " + WrapperUtility.ConvertWindowsPath(geneModelGtfOrGffPath) + /// this triggers tophat to try building an index
" --tmp-dir " + WrapperUtility.ConvertWindowsPath(tempDir) +
(strandSpecific ? " --library-type fr-firststrand" : "") +
" " + WrapperUtility.ConvertWindowsPath(bowtieIndexPrefix) +
" " + string.Join(",", fastqPaths.Select(x => WrapperUtility.ConvertWindowsPath(x))),
"if [ -d " + WrapperUtility.ConvertWindowsPath(tempDir) + " ]; then rm -r " + WrapperUtility.ConvertWindowsPath(tempDir) + "; fi",
}).WaitForExit();
}
// see here for how to generate them from scratch: http://lab.loman.net/2012/11/16/how-to-get-snpeff-working-with-bacterial-genomes-from-ncbi/
public void DownloadSnpEffDatabase(string spritzDirectory, string analysisDirectory, string reference)
{
DatabaseListPath = Path.Combine(spritzDirectory, "snpEffDatabases.txt");
// check for existing list and database
bool databaseListExists = File.Exists(DatabaseListPath);
string databaseDirectory = Path.Combine(spritzDirectory, "Tools", "SnpEff", "data");
string[] existingDatabases = Directory.Exists(databaseDirectory) ? Directory.GetDirectories(databaseDirectory) : new string[0];
bool databaseExists = existingDatabases.Any(d => Path.GetFileName(d).StartsWith(reference, true, null));
if (databaseListExists && databaseExists)
{
return;
}
// download database list
string scriptPath = WrapperUtility.GetAnalysisScriptPath(analysisDirectory, "SnpEffDatabaseDownloadList.bash");
WrapperUtility.GenerateAndRunScript(scriptPath, new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
"echo \"Downloading list of SnpEff references\"",
SnpEff(Workers) + " databases > " + WrapperUtility.ConvertWindowsPath(DatabaseListPath),
WrapperUtility.EnsureClosedFileCommands(DatabaseListPath)
}).WaitForExit();
List <string> databases = new List <string>();
using (StreamReader reader = new StreamReader(DatabaseListPath))
{
while (true)
{
string line = reader.ReadLine();
if (line == null)
{
break;
}
databases.Add(line.Split('\t')[0].TrimEnd());
}
}
string snpeffReference = databases.FirstOrDefault(d => d.StartsWith(reference, true, CultureInfo.InvariantCulture));
// download database (it downloads automatically now, with more feedback), but still need the mitochondrial references
scriptPath = WrapperUtility.GetAnalysisScriptPath(analysisDirectory, "SnpEffDatabaseDownload.bash");
WrapperUtility.GenerateAndRunScript(scriptPath, new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
"echo \"\n# " + snpeffReference + "\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"" + snpeffReference + ".genome : Human genome " + snpeffReference.Split('.')[0] + " using RefSeq transcripts\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"" + snpeffReference + ".reference : ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"\t" + snpeffReference + ".M.codonTable : Vertebrate_Mitochondrial\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
"echo \"\t" + snpeffReference + ".MT.codonTable : Vertebrate_Mitochondrial\" >> " + WrapperUtility.ConvertWindowsPath(Path.Combine(spritzDirectory, "Tools", "SnpEff", "snpEff.config")),
}).WaitForExit();
}
/// <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),
});
}
public static List <string> Bowtie2Align(string spritzDirectory, string analysisDirectory, string bowtieIndexPrefix, int threads, string[] fastqPaths,
bool strandSpecific, out string sortedBamFilePath)
{
sortedBamFilePath = Path.Combine(Path.GetDirectoryName(fastqPaths[0]), Path.GetFileNameWithoutExtension(fastqPaths[0])) + ".sorted.bam";
return(new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
"if [[ ( ! -f " + WrapperUtility.ConvertWindowsPath(sortedBamFilePath) + " || ! -s " + WrapperUtility.ConvertWindowsPath(sortedBamFilePath) + " ) ]]; then",
" bowtie2-2.3.4/bowtie2 " +
" -x " + WrapperUtility.ConvertWindowsPath(bowtieIndexPrefix) +
" -p " + threads.ToString() +
(fastqPaths.Length == 1 ?
" -U " + WrapperUtility.ConvertWindowsPath(fastqPaths[0]) :
" -1 " + WrapperUtility.ConvertWindowsPath(fastqPaths[0]) + " -2 " + WrapperUtility.ConvertWindowsPath(fastqPaths[1])) +
" | samtools view -b - " +
" | " + SamtoolsWrapper.SortBamFromStdin(sortedBamFilePath, threads),
"fi",
});
}
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>
/// 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);
}
/// <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);
}
public string DownloadEnsemblKnownVariantSites(string spritzDirectory, bool commonOnly, string reference, bool dryRun)
{
DownloadUCSCKnownVariantSites(spritzDirectory, commonOnly, reference, dryRun);
EnsemblKnownSitesPath = ConvertVCFChromosomesUCSC2Ensembl(spritzDirectory, UcscKnownSitesPath, reference, dryRun);
if (!dryRun)
{
// indexing is used for most GATK tools
WrapperUtility.GenerateAndRunScript(WrapperUtility.GetAnalysisScriptPath(spritzDirectory, "IndexKnownVariantSites.bash"), new List <string>
{
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(UcscKnownSitesPath) + ".idx ]; then " + Gatk(Workers) + " IndexFeatureFile -F " + WrapperUtility.ConvertWindowsPath(UcscKnownSitesPath) + "; fi",
"if [ ! -f " + WrapperUtility.ConvertWindowsPath(EnsemblKnownSitesPath) + ".idx ]; then " + Gatk(Workers) + " IndexFeatureFile -F " + WrapperUtility.ConvertWindowsPath(EnsemblKnownSitesPath) + "; fi",
}).WaitForExit();
}
return(EnsemblKnownSitesPath);
}
/// <summary>
/// Downloads dbSNP reference VCF file if it doesn't exist
/// </summary>
/// <param name="spritzDirectory"></param>
/// <param name="commonOnly"></param>
/// <param name="reference"></param>
/// <returns></returns>
public string DownloadUCSCKnownVariantSites(string spritzDirectory, bool commonOnly, string reference, bool dryRun)
{
bool knownSitesExists = KnownVariantSitesFileExists(spritzDirectory, commonOnly, reference);
if (!knownSitesExists && !dryRun)
{
WrapperUtility.GenerateAndRunScript(WrapperUtility.GetAnalysisScriptPath(spritzDirectory, "DownloadUcscVariants.bash"), new List <string>
{
"cd " + WrapperUtility.ConvertWindowsPath(spritzDirectory),
"wget " + TargetFileLocation,
"gunzip " + WrapperUtility.ConvertWindowsPath(UcscKnownSitesDownloadPath) + ".gz",
"rm " + WrapperUtility.ConvertWindowsPath(UcscKnownSitesDownloadPath) + ".gz",
"mv " + WrapperUtility.ConvertWindowsPath(UcscKnownSitesDownloadPath) + " " + WrapperUtility.ConvertWindowsPath(UcscKnownSitesPath)
}).WaitForExit();
}
return(UcscKnownSitesPath);
}
public void Fetch(string spritzDirectory, int threads, string analysisDirectory, string sraAccession)
{
LogPath = Path.Combine(analysisDirectory, sraAccession + "download.log");
FastqPaths = new[] { sraAccession + "_1.fastq", sraAccession + "_2.fastq" }.Select(f => Path.Combine(analysisDirectory, f)).Where(f => File.Exists(f)).ToArray();
if (FastqPaths.Length > 0) // already downloaded
{
FastqPaths = FastqPaths.Where(x => x != null && !x.Contains("trimmed") && x.EndsWith(".fastq")).ToArray();
return;
}
;
string scriptPath = WrapperUtility.GetAnalysisScriptPath(analysisDirectory, "Download" + sraAccession + ".bash");
WrapperUtility.GenerateAndRunScript(scriptPath, new List <string>
{
$"echo \"Downloading {sraAccession}\"",
WrapperUtility.ChangeToToolsDirectoryCommand(spritzDirectory),
$"sratoolkit*/bin/fasterq-dump --progress --threads {threads.ToString()} --split-files --outdir \"{WrapperUtility.ConvertWindowsPath(analysisDirectory)}\" {sraAccession} 2> {WrapperUtility.ConvertWindowsPath(LogPath)}",
}).WaitForExit();
FastqPaths = Directory.GetFiles(analysisDirectory, sraAccession + "*.fastq").ToArray();
}
