protected override void ProgramExecution()
{
Logger.WriteToLog("Generating counts file");
string countsFile = Counts.WriteCountsFile(_options.InputVcf, _options.OutputDirectory, _options.LociCount);
Logger.WriteToLog("Starting Recalibration");
try
{
QualityRecalibration.Recalibrate(_options.InputVcf, countsFile, _options.OutputDirectory, _options.BaseQNoise,
_options.ZFactor, _options.MaxQScore, _options.FilterQScore, _options.QuotedCommandLineArgumentsString);
}
catch (Exception e)
{
Logger.WriteToLog("*** Error encountered: {0}", e);
throw e;
}
Logger.WriteToLog("Work complete.");
}
static int Main(string[] arguments)
{
if (arguments.Length == 0)
{
ApplicationOptions.PrintUsageInfo();
return(1);
}
var options = ApplicationOptions.ParseCommandLine(arguments);
Init(options);
if (options == null)
{
ApplicationOptions.PrintUsageInfo();
return(1);
}
try
{
if (!(options.InputVcf).ToLower().EndsWith(".genome.vcf"))
{
Logger.WriteToLog(">>> **Warning**: VCF supplied to Recalibration algorithms should be genome VCF. Was this the intent?");
Logger.WriteToLog(">>> continuing...");
}
Logger.WriteToLog(">>> generating counts file");
string countsFile = Counts.WriteCountsFile(options.InputVcf, options.OutputDirectory);
Logger.WriteToLog(">>> starting Recalibration");
QualityRecalibration.Recalibrate(options.InputVcf, countsFile, options.BaseQNoise,
options.ZFactor, options.MaxQScore, options.FilterQScore);
}
catch (Exception e)
{
Logger.WriteToLog("*** Error encountered: {0}", e);
}
Logger.WriteToLog(">>> Work complete.");
Logger.TryCloseLog();
return(0);
}
Dictionary <MutationCategory, int> GetPhredScaledCalibratedRates(int baselineQNoise, double zFactor, string noiseFile)
{
double baseNoiseRate = MathOperations.QtoP(baselineQNoise);
var counts = new Counts();
counts.LoadCountsFile(noiseFile);
var countsByCategory = counts.CountsByCategory;
var PhredScaledRatesByCategory = new Dictionary <MutationCategory, int>();
countsByCategory.Remove(MutationCategory.Deletion);
countsByCategory.Remove(MutationCategory.Insertion);
countsByCategory.Remove(MutationCategory.Other);
countsByCategory.Remove(MutationCategory.Reference);
double[] sortedFinalCounts = countsByCategory.Values.OrderBy(d => d).ToArray();
if (countsByCategory.Keys.Count != 12)
{
return(null);
}
//take the average value, throwing out the top two and bottom two outlyiers.
double numDataPoints = 8; //12 - 4;
double avg = 0;
for (int i = 2; i < 10; i++)
{
avg += (sortedFinalCounts[i] / numDataPoints);
}
//get the variance
double variance = 0;
for (int i = 2; i < 10; i++)
{
variance += ((avg - sortedFinalCounts[i]) * (avg - sortedFinalCounts[i]) / numDataPoints);
}
//threshold = avg + z * sigma
double threshold = avg + zFactor * Math.Sqrt(variance);
foreach (var cat in countsByCategory.Keys)
{
double mutationCount = countsByCategory[cat];
if (mutationCount > threshold)
{
//baseline noise level is 'b' .
//the observed transition-rate is how frequently we observe a vf >b.
//so our expected freq due to noise =
// (prob of observation f_i <= b) + (prob of observation f_i > b)
double observedNoiseRate = 0;
if (counts.NumPossibleVariants > 0)
{
observedNoiseRate = mutationCount / counts.NumPossibleVariants;
}
//deliberately taking floor instead of rounding.
PhredScaledRatesByCategory.Add(cat, (int)(MathOperations.PtoQ(observedNoiseRate + baseNoiseRate)));
}
}
return(PhredScaledRatesByCategory);
}
