RELEASE_NOTES.txt

RELEASE NOTES FOR FastQC V0.10.0
--------------------------------

The major feature of FastQC v0.10.0 is the addition of support
for fastq files generated directly by the latest version of the
illumina pipeline (Casava v1.8).  In this version the pipeline
generates gzipped fastq files by default rather than using qseq
files which can then be converted to fastq.  However the fastq
files generated by casava are unusual in two ways:

1) A single sample produces a set of fastq files with a common
   name, but an incrementing number at the end.
   
2) Casava FastQ files contain sequences from clusters which have
   failed the internal QC, and been flagged to be filtered.

FastQC v0.10.0 introduces a Casava mode which will merge together
fastq files from the same sample group and produce a single report.
It will also exclude any flagged entries from the analysis.  You
would therefore run FastQC as normal but selecting all of the
fastq files from Casava and using casava mode for your analysis.

Casava mode is activated from the command line by adding the 
--casava option to the launch command. From the interactive 
application you need to select 'Casava FastQ Files' from the drop 
down file selector filter options.

If you want to analyse casava fastq files without these extra
options then you can use treat them as normal fastq files with 
no problems.

In addition to this change there have also been changes to allow 
the wrapper script to work properly under windows, and a bug was
fixed which missed of the last possible Kmer from every sequence
in a library.


RELEASE NOTES FOR FastQC V0.9.6
-------------------------------

FastQC v0.9.6 fixes a couple of bugs which aren't likely to have
affected the majority of fastqc users:

- Fixed a crash in the Kmer module when analysing a sequence where
  every sequence in a library contained a poly-N stretch at its
  3' end.
  
- Fixed the wrapper script so that OSX users launching fastqc
  through the script rather than the Mac application bundle get
  their classpath set correctly, and can therefore analyse bam/sam
  files.


RELEASE NOTES FOR FastQC V0.9.5
-------------------------------

FastQC v0.9.5 fixes some bugs in the programs text output and
improves a few things in the graphical interface.  Main changes
are:

- Progress calculations are now exact and not approximate

- The UI now has a welcome screen so you're not just presented
  with a blank screen when the program starts

- The wrapper script now sets the classpath correctly in windows
  as well as linux.

- The text report for per-base sequence content now reports
  correct values for grouped bases

- The HTML report uses a custom stylesheet for print output so
  graphs aren't cut off when reports are printed.

- Fixed a bug in testing for a warning in the per-base sequence
  content module.
  
- Alt text in HTML reports now matches the graphic it describes


RELEASE NOTES FOR FastQC V0.9.4
-------------------------------

FastQC v0.9.4 is a minor bugfix release which changes the offline
version of the program so that if a file fails to be processed a
full backtrace of the error is produced, rather than just a
simple generic error message.

RELEASE NOTES FOR FastQC V0.9.3
-------------------------------

FastQC v0.9.3 adds support for fastq files compressed with
bzip2 in addition to its existing support for gzip compressed
files.  It's worth noting that although bzip2 offers a reduction
in the file size of the compressed files (about a 5-fold size
reduction compared to raw fastq.  Gzip is a 4-fold decrease),
there is a significant penalty in terms of the speed of
decompression of these files.  In our tests gzipped files were
actually processed slightly faster than raw FastQ files, presumably
due to the lower amount of data transfer from disk required, 
however bzip2 compressed files took around 6X as long to 
process as gzipped files.

The other big change in this release is an update to the default
CSS layout such that viewing the HTML reports doesn't require
lots of scrolling up and down the page. As before the CSS can
be edited and customised by editing the templates shipped with
the program.  Many thanks to Phil Ewels who did much of the 
work on the new layout.



RELEASE NOTES FOR FastQC V0.9.2
-------------------------------

FastQC v0.9.2 fixes two bugs which were identified in the
previous release.

1) In the text output for the per-base quality module the 
correct base numbers weren't being included for files which
used grouped base ranges.

2) The Kmer analysis module could crash when analysing very
small files, such that no position in the file had more than
1000 instances of an enriched Kmer.

Both of these issues should now be resolved.


RELEASE NOTES FOR FastQC V0.9.1
-------------------------------

FastQC v0.9.1 adds some new command line options and fixes a
couple of usability issues.

The new command line options are:

--quiet Will suppress all progress messages and ensure that only
warnings or errors are written to stderr.  This might be useful
for people running fastqc as part of a pipeline.

--nogroup Will turn off the dynamic grouping of bases in the 
various per-base plots.  This would allow you to see a result
for each base of a 100bp run for example.  This option should
not be used for really long reads (454, PacBio etc) since it
has the potential to crash the program or generate very wide
output plots

In addition to these the following changes have been made:

The basic stats module now includes a line to say which type
of quality encoding was found so this information isn't just
present in the header of the per-base quality plot, and will
appear in the text based output.

We now distinguish between Illumina <1.3, 1.3 and 1.5 encodings
rather than just <1.3 and >=1.3.  The Sanger encoding is now
labelled as Sanger / Illumina 1.9+ to allow for the change in
encoding in the latest illumina pipeline.

A bug was fixed which caused the program to crash when encountering
a zero length colorspace file (which shouldn't happen anyway, but
crashing wasn't the correct response).

The interactive over-represented sequence table now allows you
to copy out each cell individually which makes it easy to copy
any unknown sequences into other programs.


RELEASE NOTES FOR FastQC V0.9.0
-------------------------------

FastQC v0.9.0 makes a number of changes primarily targetted at
datasets containing longer reads.  It allows all of the analyses
within FastQC to be completed sensibly and without excessive
resource usage on runs containing reads up to tens of kilobases
in length.

For long reads the program now converts many of its plots to
variably sized bins so that, for example you see every base for
the first 10 bases, then every 5 bases for the next 40 bases,
then every 50, then 100 then 1kb per bin, until the end of the
sequence is reached.  For sequences below 75bp the reports will
look exactly the same as before.  For groups running slightly
longer Illumina or ABI runs you'll see some compression at the
end of your reads, and people using 454 or PacBio will get 
sensible results for all analyses for the first time.

One additional change which will impact anyone using reads over
75bp is that the duplicate sequence and overrepresented sequence
plots now only use the first 50bp of each read if the total read
length is over 75bp.  This is because these plots work on the 
basis of an exact sequence match, and longer reads tend to show
more errors at the end which makes them look like different
sequences when they're actually the same.  50bp should be enough
that you won't see exact matches by chance.



RELEASE NOTES FOR FastQC V0.8.0
-------------------------------

FastQC v0.8.0 adds some new options for parsing BAM/SAM files and
makes the graphs in the report easier to interpret.

All graphs in v0.8.0 now have marker lines going across them to make
it easier to relate the data in the graph to the y-axis.  The per
base quality boxplots have shading behind the graph to indicate ranges
of good, medium and bad quality sequence.

For BAM/SAM files you can now specify that you wish to analyse only
the mapped sequences in the file.  This is particularly of use to
people working on colorspace data where the mapped data should
produce reliable sequence, whereas the current raw conversion to
base space may overrepresent any errors which are present.  The 
option to use only mapped data is set by using the

-f bam_mapped or -f sam_mapped

option on the command line, or by specifying Mapped BAM/SAM files
from the drop down file filter in the file chooser in the interactive
version of the program.

For FastQ files the parser has been updated to not treat blank lines
between entries or at the end of the file as a format violation since
many sequences in the public repositories can have this problem.

From the command line we now offer the option to process multiple
files in parallel by setting the -t/--threads option.  Please note that
using more than 1 thread will increase the amount of allocated memory
(250MB per thread), so you need to be sure you have enough memory (and
disk bandwidth) to be able to process more than one file.  The
interactive application still defaults to a single processing thread
but you could theoretically change this by passing the correct java
properties in the startup command.


RELEASE NOTES FOR FastQC V0.7.2
-------------------------------

FastQC v0.7.2 fixes a bug in libraries where no unique sequences
were observed.  It also improves the collection of duplicate 
statistics on libraries with very low diversity.

A new command line option has been added to allow the user to
manually specify a contaminant's file rather than using the sitewide
default.  This would be useful if you have different sets of
contaminants to screen against for different libraries, or if you
wanted to make a custom set of contaminants, but didn't have 
sufficient privileges to modify the sitewide contaminants file.


RELEASE NOTES FOR FastQC V0.7.1
-------------------------------

FastQC v0.7.1 makes some significant enhancements to the fastqc
wrapper script which make it easier to use as part of a pipeline.

You can now use normal unix options to create your fastqc
command rather than having to pass java system properties.  The
old options will continue to work though, so the updated 
wrapper is still compatible with any previous commands you may
have in place.  Full details of the new options can be found
by running 

fastqc --help

One new option is the --format option which allows you to manually
specify the input format of a file, rather than having FastQC
guess the format from the file name.  This would allow you to have
a BAM file called test.dat and process it using:

fastqc --format bam test.dat




RELEASE NOTES FOR FastQC V0.7.0
-------------------------------

FastQC v0.7.0 introduces a new analysis module which looks at the 
enrichment of short sequences within the library.  It is possible
to get enrichment of unaligned subsequences to quite a high degree
without this being apparent in any of the existing modules.  This
new module should find problems such as read through into the 
adapters at the other end of libraries, which other analyses would
miss.

Other changes in this release include:

* Altering the fastqc wrapper script so it can identify when it's
  being used on the source distribution of the software so it can
  issue an appropriate warning.

* Tidied up all of the y-axes on the graphs so that scaling should
  now always be perfect in all graphs.


RELEASE NOTES FOR FastQC V0.6.1
-------------------------------

FastQC v0.6.1 is a bugfix release which fixes a problem with sequences
from BAM/SAM files which map to the reverse strand of the reference.

In these cases the sequence contained in the BAM/SAM file is reverse
complemented and the qualities are reversed relative to the original
sequence which came off the sequencer.  In the previous release this
meant that the plots were incorrectly showing a mix of forward and 
reversed sequences.

In v0.6.1 any sequences mapping to the reverse strand of the reference
are converted back to their original state before being analysed which
should give a clearer view of the overall qualities and sequence biases
within the run.


RELEASE NOTES FOR FastQC V0.6.0
-------------------------------

FastQC v0.6 adds support for reading SAM/BAM files as well as still
supporting fastq files.  File type detection is based on the filename
so SAM/BAM files need to be named .sam or .bam.  Any other form of
filename is assumed to be a fastq file.

SAM/BAM reading was added through the use of the picard libraries, which
means that the launcher for FastQC has had to be modified to include
the picard libraries into the classpath.  If you use the bat file launcher,
the wrapper script or the Mac application bundle then you won't need to
do anything to get the new version to work, but if you've created your
own launcher then you will need to modify your classpath statement to
include the sam-1.32.jar file into the classpath.

The only other change in this release is that the line graphs have been
improved to use smoother lines for the graphs.



RELEASE NOTES FOR FastQC V0.5.1
-------------------------------

Release 0.5.1 fixes a couple of bugs and makes some improvements to 
existing functions.

* A bug was fixed which caused the headers of the overrepresented
  sequences results to not be separated in the text output.
  
* A bug was fixed which caused spikes to appear in the %GC profile
  when using read lengths >100bp

* The fitting of the theoretical distribution to the %GC profile
  was improved
  
* Some new entries were added to the contaminants file to cover
  Illumina oligos for multiplexing, tag expression and small RNA
  protocols (thanks to Aaron Statham for providing these)


RELEASE NOTES FOR FastQC V0.5.0
-------------------------------

FastQC v0.4.4 makes a number of changes to the previous
release which hopefully improve the relevance of the output.

* The fitting of the normal curve to the %GC distribution has
  been improved for shorter sequences
  
* Each section of the HTML report output now has a pass/warn/fail
  icon next to it, rather than just having them at the top

* The duplicate level analysis now estimates the total percentage
  of sequences which are not unique, reports this on the graph and
  uses this as the basis for the pass/warn/fail filtering
  
* The structure of the HTML output folder has been changed so that
  icons and images are put into subfolders so the only files at
  the top level are ones which the user is intended to open directly.


RELEASE NOTES FOR FastQC V0.4.3
-------------------------------

FastQC v0.4.3 is a bugfix release which fixes a bug and adds
an extra check to the interactive program.

In versions of the program since v0.2 the total sequence count
in the Basic Stats module may have been incorrect by a few percent
(either high or low).  This is because instead of using the real
sequence count the module was using an estimated count which should
only be used for updating the progress indicators.  The module
has now been fixed to report the actual sequence count.

In the interactive application there was no warning given if you 
chose to save a report and opted to overwrite an existing report.  You
will now be warned in this case an offered the chance to use a
different filename.  This change won't affect the non-interactive
mode of the program where report files will be overwritten if the
program is run more than once on the same set of data.

RELEASE NOTES FOR FastQC V0.4.2
-------------------------------

FastQC v0.4.2 is a minor release which fixes some bugs and improves
the warnings on some of the analysis modules.

Bugs which were fixed were:

* The per-base quality plot showed an incorrect scale on the
  y-axis.  This was usually off by about 2, but could be more
  depending on the data.  The relative relationships shown
  were correct, it was just the scale bar which was wrong.
  
* The sequence parser incorrectly identified some base called
  files as colorspace files where they contained dots in place
  of N base calls.  This has now been improved but will still
  fail for the base where a base called file starts with an 
  entry which is a single called base followed by all dots
  since this is indistinguishable from a colorspace fastq file.
  
* Some JREs (notably OpenJDK) can't cope with a headless environment
  being set from within the program.  I've therefore added code
  to the fastqc wrapper script to externally set the headless 
  environment up to bypass this problem.
  
Other improvements which have been made are:

* When writing out graphs for the HTML report we now scale some
  graphs by the length of the sequences analysed so we don't get
  really squashed graphs.  The interactive application will still
  scale the graphs to the size of the window
  
* More checks have been added to the parsing of FastQ files to
  ensure we're looking at the correct file format.  A better reporting
  mechanism has been added to allow the program to cope better with
  problems during parsing.
  
* The per-sequence GC plot has been modified to add in a modelled
  normal distribtion over the observed distribution so you can see
  how well the two fit.
  
* All modules (apart from the general stats) now have valid checks
  in place, and all can now produce warnings and failures.  Some of
  the checks in the existing modules have been changed to better cope
  with libraries with very low or high GC content.


RELEASE NOTES FOR FastQC V0.4.1
-------------------------------

FastQC v0.4.1 is a bugfix release which changes the operation
of the duplicate sequence and overrepresented sequence modules.

Both of these modules now only track sequences which appear in
the first 200,000 sequences of a file.  This change was made
because people using longer read lengths found that tracking
the first million sequences led to the program exhausting its
allocated memory.

The other change is that the duplicate sequence module now
tracks sequence counts to the end of the file rather than
stopping after a million sequences.  This means the final
duplication counts seen are much more realistic and should
match with what you would see in a genome browser.  Because
of this change the cutoffs to flag a file with a warning or
an error have been increased.  You now get a warning when 
the sum of duplicated sequences is more than 20% of the
unique sequences.  You get an error when there are more
duplicated sequences than unique sequences.


RELEASE NOTES FOR FastQC V0.4
-----------------------------

FastQC v0.4 introduces a new analysis module, and easier way
to launch the program from the command line and a new output 
file, as well as fixing a few minor bugs.

The new analysis module is the sequence duplication level
module.  This is a complement to the existing overrepresented
sequences module in that it looks at sequences which occur
more than once in your data.  The new module takes a more
global view and says what proportion of all of your sequences
occur once, twice, three times etc. In a diverse library most
sequences should occur only once.  A highly enriched library
may have some duplication, but higher levels of duplication
may indicate a problem, such as a PCR overamplification.

In response to several requests we've also now introduced a
new output file into the report.  This is a text based, tab
delimited file which includes all of the data show in the
graphs in the graphical report.  This would allow people
running pipelines to store the data generated by fastQC and
analyse it systematically rather than just taking the 
pass/fail/warn summary, or reviewing the reports manually.

Finally, if you're running fastqc from the command line we've
now included a 'fastqc' wrapper script which you can launch
directly rather than having to construct a java launch
command.  You can still pass -Dxxx options through to the 
program, but for simple analyses you can now simply run:

fastqc [some files]

..once you have included the FastQC install directory into your
path.  More details are in the install document.

Other minor changes:

- The over-represented sequences module now scans the first
million sequences to decide which sequences to track to the
end of the file.

- The labeling on the per-base N content graph is now correct



RELEASE NOTES FOR FastQC V0.3.1
-------------------------------

V0.3.1 is a bugfix release which fixes a few minor problems.

1) The template system now correctly checks that it only imports
graphics files from the templates directory into the report
files and doesn't break when it encounters unexpected files.

2) The offline mode now correctly reports the progress of all
processed files rather than just the first one.

3) The documentation has been updated to include information on
the blue mean line in the per base quality plot.


RELEASE NOTES FOR FastQC V0.3
-----------------------------

Major new additions to v0.3 are listed below:

1) Support for gzip compressed fastq files.  You can now
load gzip compressed fastq files into FastQC in the same
way as uncompressed files.  The files will be decompressed
interactively and can be viewed in the same way as for
uncompressed files.

2) Contaminant identification.  If your library is found to
have overrepresented sequences in it these are now scanned 
against a database of common contaminants (primers, adapters etc)
to see if the source of the contamination can be identified.
The database is stored in a new Contaminants folder in the
main installation directory, and can be updated with your
own protocol specific sequences.

3) When in non-interactive mode you can now pass a 
preference -Dfastqc.output_dir to provide an alternative location
to save reports to, rather than having them in the same
directory as the source fastq files.

Some improvements have also been made to the colorspace support
so this version should support a wider range of colorspace
fastq files.


RELEASE NOTES FOR FastQC V0.2
-----------------------------

There are a few new additions in v0.2 of FastQC

1) Colorspace support: We now have rudimentary support for the
analysis of fastq files in colorspace.  The analysis is done by
converting the colorspace calls to basecalls, which isn't ideal
but is hopefully more useful than nothing.

2) Option to unzip reports.  The default action in non-interactive
mode is now to both create a zip file containing the FastQC report
and to unzip this into a folder of the same name.  If you just want
to generate the zip files you can add -Dfastqc.unzip=false to the
command line to suppress this new behaviour

3) It is now possible to customise the HTML report for your site.
There is an HTML template which can be edited to add your own 
branding to the reports you generate.

4) In addition to the human readable HTML report we now also
generate a tab delimited summary file which should make it 
easier to integrate FastQC into an automated reporting system
which spots any potential problems with the data.


RELEASE NOTES FOR FastQC V0.1.1
-------------------------------

This point release fixes a problem when running FastQC in a
non-interactive mode on a system without a graphical display.
The program should now operate correctly on headless systems
as long as the filename(s) to process are specified on the
command line.


RELEASE NOTES FOR FastQC V0.1
-------------------------------

FastQC v0.1 is a beta release it should work in its present state but
we are keen to get feedback on the program.  In particular we are
interested to hear if anyone has:

1) Suggestions for other checks we could be performing.

2) Comments about the criteria we set for issuing warnings or errors and
suggestions for how these could be improved.

You can report feedback either though our bug reporting tool at:

www.bioinformatics.bbsrc.ac.uk/bugzilla/

...or directly to simon.andrews@bbsrc.ac.uk