Hi quick question regarding the SVPreprocess and generation of metadata. What is the quickest/most convenient way to filter out samples from the metadata contained in a single folder without having to re-run SVPreprocess?
Many of our downstream tools have ways to ignore samples, although the mechanism is not always the same.
Our CNV calling pipeline supports an -excludeReadGroup parameter, which you can use to exclude individual read groups or all read groups for a sample (which will effectively exclude the sample).
Another approach we often is to not supply all of the input .bam files to the downstream tool (i.e. leave out the samples you want to ignore). In the CNV pipeline, the samples you want to call on are defined by the -I arguments.
In fact, for the CNV pipeline, you can actually run off of a "headers only" bam file, such as metadata/headers.bam.
If you create a different headers.bam where you remove some of the samples, then only the samples you keep be used in calling.