ubic.gemma.core.analysis.preprocess

Interface PreprocessorService

All Known Implementing Classes:

PreprocessorServiceImpl

public interface PreprocessorService

Encapsulates steps that are done to expression data sets after they are loaded and experimental design curated.

This can also be used to 'refresh' everything.

The following steps are performed:

1. Deleting old analysis files and results, as these are invalidated by the subsequent steps.S
2. Computing missing values (two-channel)
3. Creation of "processed" vectors
4. Batch-correction
5. PCA
6. Computing sample-wise correlation matrices for diagnostic plot
7. Computing mean-variance data for diagnostic plots
8. GEEQ scoring
9. Redoing any DEA (if this is a 'refresh')

Note that since each step can be replayed and the whole process is lengthy and likely to lock parts of not whole tables, it is marked as Propagation.NEVER to prevent execution within another transaction.

Author:

paul

Method Summary

Modifier and Type	Method and Description
default void	process(ExpressionExperiment ee) Preprocess a dataset.
default void	process(ExpressionExperiment ee, boolean ignoreQuantitationMismatch) Preprocess a dataset.
void	process(ExpressionExperiment ee, boolean ignoreQuantitationMismatch, boolean ignoreDiagnosticFailure) Preprocess a dataset.
void	processDiagnostics(ExpressionExperiment ee) Create or update the sample correlation, PCA and M-V data.

Method Detail

process

default void process(ExpressionExperiment ee)
              throws PreprocessingException

Preprocess a dataset.

Mismatched quantitation types are ignored by default, diagnostics failure will in a PreprocessingException.

Throws:

PreprocessingException

See Also:

process(ExpressionExperiment, boolean, boolean)

process

default void process(ExpressionExperiment ee,
                     boolean ignoreQuantitationMismatch)
              throws PreprocessingException

Preprocess a dataset.

Diagnostic failure will result in a PreprocessingException.

Throws:

PreprocessingException

See Also:

process(ExpressionExperiment, boolean, boolean)

process

void process(ExpressionExperiment ee,
             boolean ignoreQuantitationMismatch,
             boolean ignoreDiagnosticFailure)
      throws PreprocessingException

Preprocess a dataset.

Parameters:

ee - the expression experiment to process

ignoreQuantitationMismatch - ignore quantitation mismatch when generating processed EVs

ignoreDiagnosticFailure - simply warn if a diagnostic fails instead of interrupting the pre-processing and raising an exception

Throws:

PreprocessingException - if there was a problem during the processing

processDiagnostics

void processDiagnostics(ExpressionExperiment ee)
                 throws PreprocessingException

Create or update the sample correlation, PCA and M-V data.

This is also done as part of process(ExpressionExperiment, boolean, boolean) so should only be called if only a refresh is needed.

Throws:

PreprocessingException