Picard release 1.102
4 November 2013
- Changed default read name header formatting for fastq records emitted by `IlluminaBasecallsToFastq` to include new values and, in particular, passing-filter flags. The old read name header formatting is referred to as "Illumina," and the new formatting "Casava 1.8". This change necessitated adding two new required (by default) arguments: FLOWCELL_BARCODE and MACHINE_NAME.
You may control the formatting emitted by IlluminaBasecallsToFastq with the `READ_NAME_FORMAT` argument. (For example, pass READ_NAME_FORMAT=ILLUMINA to emulate default functionality prior to this commit.)
- ValidateSamFile.java: Do not check file termination if reading from a pipe.
- Separating the VCF encoding functions from the VCFWriter class.
- Change SeekableStreamFactory to a singleton.
Allow an application to set the implementation of SeekableStreamFactory. Default implementation is the same.
- Moved `VariantContextUtil`'s `calcVCFGenotypeKeys` and a field member into `VariantContext` to avoid referencing `VariantContextUtils` in some command-line programs. Referencing `VariantContextUtils` can cause `ClassInitializationException`s because it references jexl packages which do dynamic class-loading. This is more of a band-aid than a fix.
- MakeSitesOnlyVcf.java: Bug fix: do not call `VCFHeader` constructor that copies sample names, since that forces the writer to emit genotype data for those samples (which fails the requirements of a site-only VCF).