/**
* Update the recalibration statistics using the information in recalInfo
*
* @param recalInfo data structure holding information about the recalibration values for a single
* read
*/
@Requires("recalInfo != null")
public void updateDataForRead(final ReadRecalibrationInfo recalInfo) {
final GATKSAMRecord read = recalInfo.getRead();
final ReadCovariates readCovariates = recalInfo.getCovariatesValues();
final RecalibrationTables tables = getUpdatableRecalibrationTables();
final NestedIntegerArray<RecalDatum> qualityScoreTable = tables.getQualityScoreTable();
for (int offset = 0; offset < read.getReadBases().length; offset++) {
if (!recalInfo.skip(offset)) {
for (final EventType eventType : EventType.values()) {
final int[] keys = readCovariates.getKeySet(offset, eventType);
final int eventIndex = eventType.ordinal();
final byte qual = recalInfo.getQual(eventType, offset);
final double isError = recalInfo.getErrorFraction(eventType, offset);
RecalUtils.incrementDatumOrPutIfNecessary(
qualityScoreTable, qual, isError, keys[0], keys[1], eventIndex);
for (int i = 2; i < covariates.length; i++) {
if (keys[i] < 0) continue;
RecalUtils.incrementDatumOrPutIfNecessary(
tables.getTable(i), qual, isError, keys[0], keys[1], keys[i], eventIndex);
}
}
}
}
}
/**
* Creates the recalibration report. Report can then be written to a stream via
* GATKReport.print(PrintStream).
*
* @return newly created recalibration report
*/
public GATKReport createGATKReport() {
return RecalUtils.createRecalibrationGATKReport(
argumentTable,
quantizationInfo,
recalibrationTables,
requestedCovariates,
RAC.SORT_BY_ALL_COLUMNS);
}
/**
* Combines two recalibration reports by adding all observations and errors
*
* <p>Note: This method DOES NOT recalculate the empirical qualities and quantized qualities. You
* have to recalculate them after combining. The reason for not calculating it is because this
* function is intended for combining a series of recalibration reports, and it only makes sense
* to calculate the empirical qualities and quantized qualities after all the recalibration
* reports have been combined. Having the user recalculate when appropriate, makes this method
* faster
*
* <p>Note2: The empirical quality reported, however, is recalculated given its simplicity.
*
* @param other the recalibration report to combine with this one
*/
public void combine(final RecalibrationReport other) {
for (int tableIndex = 0; tableIndex < recalibrationTables.numTables(); tableIndex++) {
final NestedIntegerArray<RecalDatum> myTable = recalibrationTables.getTable(tableIndex);
final NestedIntegerArray<RecalDatum> otherTable =
other.recalibrationTables.getTable(tableIndex);
RecalUtils.combineTables(myTable, otherTable);
}
}
/** Merge all of the tables from toMerge into into this set of tables */
public RecalibrationTables combine(final RecalibrationTables toMerge) {
if (numTables() != toMerge.numTables())
throw new IllegalArgumentException(
"Attempting to merge RecalibrationTables with different sizes");
for (int i = 0; i < numTables(); i++) {
final NestedIntegerArray<RecalDatum> myTable = this.allTables.get(i);
final NestedIntegerArray<RecalDatum> otherTable = toMerge.allTables.get(i);
RecalUtils.combineTables(myTable, otherTable);
}
return this;
}
public RecalibrationReport(final File recalFile, final SortedSet<String> allReadGroups) {
final GATKReport report = new GATKReport(recalFile);
argumentTable = report.getTable(RecalUtils.ARGUMENT_REPORT_TABLE_TITLE);
RAC = initializeArgumentCollectionTable(argumentTable);
GATKReportTable quantizedTable = report.getTable(RecalUtils.QUANTIZED_REPORT_TABLE_TITLE);
quantizationInfo = initializeQuantizationTable(quantizedTable);
Pair<ArrayList<Covariate>, ArrayList<Covariate>> covariates =
RecalUtils.initializeCovariates(RAC); // initialize the required and optional covariates
ArrayList<Covariate> requiredCovariates = covariates.getFirst();
ArrayList<Covariate> optionalCovariates = covariates.getSecond();
requestedCovariates = new Covariate[requiredCovariates.size() + optionalCovariates.size()];
optionalCovariateIndexes = new HashMap<String, Integer>(optionalCovariates.size());
int covariateIndex = 0;
for (final Covariate covariate : requiredCovariates)
requestedCovariates[covariateIndex++] = covariate;
for (final Covariate covariate : optionalCovariates) {
requestedCovariates[covariateIndex] = covariate;
final String covariateName =
covariate
.getClass()
.getSimpleName()
.split("Covariate")[
0]; // get the name of the covariate (without the "covariate" part of it) so we can
// match with the GATKReport
optionalCovariateIndexes.put(covariateName, covariateIndex - 2);
covariateIndex++;
}
for (Covariate cov : requestedCovariates)
cov.initialize(
RAC); // initialize any covariate member variables using the shared argument collection
recalibrationTables = new RecalibrationTables(requestedCovariates, allReadGroups.size());
initializeReadGroupCovariates(allReadGroups);
parseReadGroupTable(
report.getTable(RecalUtils.READGROUP_REPORT_TABLE_TITLE),
recalibrationTables.getReadGroupTable());
parseQualityScoreTable(
report.getTable(RecalUtils.QUALITY_SCORE_REPORT_TABLE_TITLE),
recalibrationTables.getQualityScoreTable());
parseAllCovariatesTable(
report.getTable(RecalUtils.ALL_COVARIATES_REPORT_TABLE_TITLE), recalibrationTables);
}
