New stable release

• Migrated build system and testing framework to a server at my new institute (NCI)
• Upgraded builder to sbt 1.1.4
• Copied over some utility code from vArmyKnife
• Added the "--prefilterImproperPairs"
• Fixed error message typo (thanks @fbearoff !)

1.3.0

New stable release.

Minor changes and updates.

• Updated the scala libraries to 1.22.4 in order to solve a discovered bug in the AnyRefMap collection class (see here). This may have caused random crashes when attempting to calculate reference mismatch rates and processing certain reads. Crashes would be expected to occur roughly once every 4 billion read-pairs.
• Added an option to allow downsampling of bam files when generating wiggle files. The downsampling is fairly memory intensive, and there are probably better ways to do this. But it's fast.
• Removed methods that are deprecated in scala v1.22. (specifically scala.collection.JavaConversions)
• Minor documentation fixes.

v1.2.42

A few relatively minor changes.

• Added a "build.plotter.advanced" function that gives much greater control over highlighting and coloring. It allows the highlighting and coloring of arbitrarily defined groups and sub-groups of replicates. Should provide a solution to issue 37.
• Fixed problem with the (still under beta-testing) reference-mismatch engine in which it would store a larger buffer of the genome fasta file than actually needed. This caused extreme memory-use issues when the genome fasta file and the bam file don't have the same chromosome ordering.

1.2.37

Note that this version is still under heavy development and is not yet comprehensively tested.

New to 1.2.37:

• Fixed major bug with plotting (thanks to Roy Francis!)
• Fixed various typos and dead links in vignette, help docs, and web documentation.

v1.2.26

Minor bugfixes and performance improvements.

• Genome fasta parser no longer requires that the chromosomes be in the same order as found in the BAM file. In addition, the --genomeFA option will now accept a list of individual-chromosome fasta files as well as a single multi-chromosome fasta.
• Added more internal testing to the build pipeline.

1.2.25

Fixed major bug with the FASTA parser, which sometimes causes the reference-mismatch calculator to vastly overestimate the mismatch rate.

1.2.19

Minor bugfixes.

EDIT: NOTE: This version of QoRTs has a major known bug! Under certain (common) circumstances, the reference-mismatch calculator vastly overestimates the number of reference mismatches. This bug is patched in v1.2.25.

v1.2.11

EDIT: NOTE: This version of QoRTs has a major known bug! Under certain (common) circumstances, the reference-mismatch calculator vastly overestimates the number of reference mismatches. This bug is patched in v1.2.25.

Major update. Numerous new features have been added and are still undergoing beta testing. These new features may be subject to change in the next stable release.

Among the changes:

• Added support for whole-exome or whole-genome datasets in addition to RNA-Seq. (Maybe rename the tool?)
• Added numerous metrics which may be relevant to variant calling.
• Intermediate file documentation: The raw QC metric files produced by QoRTs are now better documented. You can cause QoRTs QC to generate a documentation file using the parameter "--addFunctions writeDocs"

Added an array of new metrics:

• "Overlap Mismatch": various metrics relating to the rate at which overlapped paired-end reads are found to mismatch one another. This can be used as a proxy for the sequencing error, since the two paired-end reads sequence the same physical cDNA fragment. Mismatch rates are calculated by base-swap type, by quality score, and by position in the reads.
• "Reference Mismatch": various metrics relating to the rate at which reads have point-mismatches with the reference genome. Requires that a genome fasta file (via the --genomeFA parameter for the QoRTs QC step). Mismatch rates are calculated by base-swap type, by quality score, and by position in the reads.
• "On-Target Rate": For Exome data only. Uses a target bed file to calculate rate of on-target reads. Can also be used to filter reads to only on-target reads. Requires a target BED file (set via the --targetRegionBed parameter)
• "Read Length Rates": Rates of observed read lengths. Useful if data is hard-trimmed prior to alignment.
• Performance Plot: Plot shows the runtime performance of the QoRTs QC run.
• Raw FASTQ QC: Added modules to QoRTs QC that allow the specification of a fastq file. QoRTs will run some basic QC on the FASTQ file (NVC, missingness rate, GC rate, read-length distribution, and quality score metrics).

New Plotting Functionality:

• Added more flexible multiplotting. Can now arbitrarily set the number of rows or columns of multiplots, and QoRTs will automatically fit the requested plots in the requested plots.
• Can now change which plots to include in multiplots. Plotters will automatically resize and reorganize sub-plots.
• By default, makemultiplot will automatically remove plots that cannot be created due to missing data.
• Can now create more flexible plotters with manually-set coloration and highlighting.

Internal changes:

• Improved performance of several internal utilities (in particular, NVC calculator is now almost twice as fast). Most of the improvement is offset by the addition of new metrics, so with all new modules active runtime is roughly the same as in 1.1.
• New paired iterator that sorts pairs by the lowest genomic position. This is necessary for the fast and efficient calculation of reference mismatches, but somewhat reduces performance and increases memory usage. Therefore it is only used when reference mismatch rates are calculated.

v1.1.8

Added two new options to makePlot.biotype.rates. See the associated issue..

Minor internal changes.

v1.1.6

Minor additions and tweaks:

• Added new optional tag to "flattened" GFF file, containing transcript coding region information. This is intended for use by a future update to JunctionSeq.
• Fixed x-axis label autofitting to not produce oversized x-axis labels when there is only one column to label.
• Legend plot now autofits and autoselects the number of columns. Added cex.legend and ncol parameters to these functions to allow users to override this behavior.
• Fixed bug with maxObsReadLength field.