FastQC Files

This is a major release of the program with multiple changes. Changes to the build environment and the options in the launch script mean that this may not be a drop-in replacement for existing installations. The main change which affects switching is that the command line launch script moved from perl to python3 so a python interpreter is now required to launch the program. No additional libraries outside the python standard library are required.

Because java 6 is now not supported by the latest compilers, FastQC now requires java >= 8 to work. This is still ancient so shouldn't pose any problems but if you have something somewhere still running java6 then you'll need to update that.

The main functional changes in this release are:

• @ewels and @pditommaso have contributed significant improvements in both the threading model and general efficiency of processing. The threading model now allows for up to 4 threads to be used to process a single fastq file, and this scales nicely with more threads allowing multiple files to process in parallel. The combined effect of this is that processing files is significantly faster than in previous versions.
• The HTML reports now generate graphs in SVG format by default, making it much easier to adapt them for publication. PNG images can still be generated by using the --png command line flag. *bzip decompression now uses apache commons as its back end which fixed issues with concatenated bzip2 files being truncated in the report.
• Allow the interactive application to detect and process bzip compressed files.
• Fixes in reporting metrics which overflowed the 2^32 boundary for integers.
• Basic stats now reports mean and median sequence lengths
• Autodetection of Phred encoding is now disabled and all analysis defaults to Phred32. Analysis of legacy Phred64 files requires the use of --phred64 command line option. We still check the data to ensure that the encoding selected is actually technically feasible and will give warnings or errors if it isn't.
• The old, very broken --min_length option has been replaced with new --min_length and --max_length options which pre-filter reads to fall between the specified lengths. Reads outside the range set are discarded before any analysis is done.