testFunctional: Functions dedicated to the functional validation of Pasha...
In Pasha: Preprocessing of Aligned Sequences from HTS Analyses

Description Usage Arguments Details Value Author(s) See Also

The goal of these two functions is to generate a reference dataset with the currently installed Pasha package in the user's environment and/or to compare it to the precomputed results in the developpers environment.

 
.testFunctional(folderOutput=tempdir(), testType="regular",
verbose=FALSE) 

.testFunctional_generateResults(testFileBAM_folderName,
                                testFileBAM_fileName, 
                                testFileMulti_fileName,
                                testType="complete")

`folderOutput`	An existing folder (this is tested and can eventually raise an error) in which the generated results will be written and the precomputed results will be inflated.
`testFileBAM_folderName`	The folder in which the BAM file for results generation stands (.testFunctional passes the path to the package-embedded test BAM file)
`testFileBAM_fileName`	The filename of the BAM file used to generate the results (.testFunctional passes the name of the package-embedded test BAM file)
`testFileMulti_fileName`	The filename of the file that is used as test to generate ouptut results for comparison. It simulates reads that were treated with the multiread processing functions before injecting it in the pipeline.
`testType`	A value in c("regular", "complete"). A regular test will generate ~15 WIG and GFF files, a complete test however generates ~750 result files from the test BAM file and aims at passing in all possible loops and blocks of the pipeline.
`verbose`	Logical. If verbose is FALSE, the pipeline stdout will be redirected to a log file to avoid overflow of test messages.

Once installed, Pasha gives access to a functional test with the function '.testFunctional'.

The package includes a SAM/BAM file designed to cover most of the cases for paired-ends or single-ends datasets (see publication for details). This test consists in using the provided pipeline and the embedded BAM file to generate results and compare them to a previously generated (precomputed) dataset that has been functionally validated by developpers. The precomputed dataset is embedded in the package in a zip file and contains all the WIG and GFF files that could be necessary for validation. This dataset is generated and compiled by package developpers using .testFunctional_generateResults. Two kind of tests can be computed: A regular test will go through the most common options for single-end and paired-ends, it will generate 15 to 20 results file and compare them to the references, the generation and comparisons should take less than a minute to complete. A complete test will launch the generation of results by using as much as possible different options and parameters to the pipeline function. It will typically generate ~1000 wig and gff files and compare them to the validated reference, this test can be long (15 to 30 minutes) but guarantees to obtain the same results as the developpers for all kind of analysis.

.testFunctional prints to stdout a summary of file generation and comparison with reference dataset.

Romain Fenouil

AlignedData-class processPipeline

Pasha documentation built on Jan. 15, 2017, 6:21 p.m.