Finish Release-2.3.2

KAT now allows users to pipe in gzipped files via process substitution.

Updated distribution analysis script to make it more robust and added calculations for estimated genome size and level of heterozygos content.

Fixed a bug in spectra hist plot title.

Fixed a bug in the help message for '-n' option in kat comp.

Author: Daniel Mapleson

.gitignore

@@ -40,3 +40,4 @@ Makefile.in
 *.pc
 *.la
 *.out
+/tags

configure.ac

@@ -4,7 +4,7 @@
 
 # Autoconf initialistion.  Sets package name version and contact details
 AC_PREREQ([2.68])
-AC_INIT([kat],[2.3.1],[https://github.com/TGAC/KAT/issues],[kat],[https://github.com/TGAC/KAT])
+AC_INIT([kat],[2.3.2],[https://github.com/TGAC/KAT/issues],[kat],[https://github.com/TGAC/KAT])
 AC_CONFIG_SRCDIR([src/kat.cc])
 AC_CONFIG_AUX_DIR([build-aux])
 AC_CONFIG_MACRO_DIR([m4])
@@ -176,6 +176,7 @@ else
     if [[ -z "${BOOST_TIMER_STATIC_LIB}" ]] || [[ -z "${BOOST_CHRONO_STATIC_LIB}" ]] || [[ -z "${BOOST_FILESYSTEM_STATIC_LIB}" ]] || [[ -z "${BOOST_PROGRAM_OPTIONS_STATIC_LIB}" ]] || [[ -z "${BOOST_SYSTEM_STATIC_LIB}" ]]; then
         AC_MSG_WARN([Not all static boost libraries could be found.  Will use dynamic libraries instead.])
         BOOST_LIBS="${BOOST_DYN_LIBS}"
+        dynboost="yes"
     else
         BOOST_LIBS="${BOOST_STATIC_LIBS}"
     fi

doc/source/conf.py

@@ -52,9 +52,9 @@
 # built documents.
 #
 # The short X.Y version.
-version = '2.3.1'
+version = '2.3.2'
 # The full version, including alpha/beta/rc tags.
-release = '2.3.1'
+release = '2.3.2'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

doc/source/faq.rst

@@ -4,10 +4,11 @@
 Frequently Asked Questions
 ==========================
 
-Can KAT handle gzipped sequence files?
---------------------------------------
+Can KAT handle compressed sequence files?
+-----------------------------------------
 
-Yes, but only via named pipes.  For example, say we wanted to run ``kat hist`` using
+Yes, via named pipes.  Anonymous named pipes (process substitution) is also supported.
+For example, say we wanted to run ``kat hist`` using
 gzipped paired end dataset, we can use a named pipe to do this as follows::
 
     mkfifo pe_dataset.fq && gunzip -c pe_dataset_?.fq.gz > pe_dataset.fq &
@@ -21,7 +22,17 @@ consuming from the named pipe will take data that has been gunzipped first.  To
 clear this means you do not have to decompress the gzipped files to disk, this happens
 on the fly as consumed by KAT.
 
-Thanks to John Davey for suggesting this.
+Alternatively, using process substitution we could write the previous example more 
+concisely in a single line like this::
+
+    kat hist -o oe_dataset.hist <(gunzip -c pe_dataset_?.fq.gz)
+
+As a more complex example, the KAT comp tool can be driven in spectra-cn mode using
+both compressed paired end reads and a compressed assembly as follows::
+
+    kat comp -o oe_spectra_cn <(gunzip -c pe_dataset_?.fq.gz) <(gunzip -c asm.fa.gz)
+
+Thanks to John Davey and Torsten Seeman for suggesting this.
 
 
 Why is jellyfish bundled with KAT?

doc/source/images/distanalysis_console.png

@@ -308,8 +308,9 @@ Genome assembly analysis using k-mer spectra
 --------------------------------------------
 
 One of the most frequently used tools in KAT are the so called "assembly spectra
-copy number plots" or spectra-cn. We use these as a fast first analysis for assembly coherence
-to the data in the reads they are representing. Basically we represent how many elements
+copy number plots" or spectra-cn. We use these as a fast first analysis to check
+assembly coherence against
+the content within reads that were used to produce the assembly. Basically we represent how many elements
 of each frequency on the read’s spectrum ended up not included in the assembly, included
 once, included twice etc.
 
@@ -374,6 +375,36 @@ duplications, inclusion of extra variation, etc:
     :scale: 33%
 
 
+Distribution decomposition analysis
+-----------------------------------
+
+It's useful to be able to fit distributions to each peak in a k-mer histogram or spectra-cn matrix
+in order to work out how many distinct k-mers can be associated with those distributions.  By counting
+k-mers in this way we can make predictions around genome size, heterozygous rates (if diploid) and 
+assembly completeness.  To this end we bundle a script with kat called kat_distanalysis.py.  It takes
+in either a spectra-cn matrix file, or kat histogram file as input, then proceeds to identify peaks
+and fit distributions to each one.  In the case of spectra-cn matrix files it also identifies peaks
+for each copy number for an assembly.
+
+The user can help to get correct predictions out of the tool by providing an approximate frequency for
+the homozygous part of the distribution.  By default, this is assumed to be the last peak.  For example,
+this command::
+
+        kat_distanalysis.py --plot spectra-cn.mx
+
+might produce the following output for this tetraploid genome:
+
+.. image:: images/distanalysis_console.png
+    :scale: 100%
+
+.. image:: images/distanalysis_plot.png
+    :scale: 100%
+
+
+
+
+
+
 Finding repetitive regions in assemblies
 ----------------------------------------
 

doc/source/images/distanalysis_plot.png

@@ -4,3 +4,4 @@
 *.lo
 *.o
 *.kat-2.1.pc.swp
+/tags

doc/source/walkthrough.rst

@@ -0,0 +1 @@
+/tags

lib/.gitignore

@@ -0,0 +1 @@
+/tags