Continuous integration on master branch - build status:
| Linux | gcc 4.7 | gcc 4.8 | gcc 4.9 | clang 3.6 | clang 3.9 | Valgrind |
|---|---|---|---|---|---|---|
| Debian 8 | - | |||||
| Debian 7 | - | - | - | - | ||
| Fedora 20 | - | - | - | - | - |
| Mac OSX | clang-600 | gcc 4.2.1 |
|---|---|---|
| 10.9 |
GATB is made of two master projects:
-
The GATB-CORE project provides a set of highly efficient algorithms to analyse NGS data sets. These methods enable the analysis of data sets of any size on multi-core desktop computers, including very huge amount of reads data coming from any kind of organisms such as bacteria, plants, animals and even complex samples (e.g. metagenomes). Read more about GATB at https://gatb.inria.fr/.
-
The GATB-TOOLS project contains a set of ready-to-use softwares relying on GATB-CORE algorithms. You can have a look at available tools at https://gatb.inria.fr/software/.
GATB-CORE is a high-performance and low memory footprint C++ library.
GATB-Core natively provides the following operations:
-
Reads handling:
-
FASTA/FASTQ parsing
-
Parallel iteration of sequences
-
K-mer:
-
K-mer counting
-
Minimizer computation of k-mers, partitioning of datasets by minimizers
-
Bloom data structure of k-mers
-
Hash table of k-mers
-
Minimal perfect hash function of k-mers
-
Arbitrarily large k-mers representations
-
de Bruijn graph:
-
graph construction
-
graph traversal operations (contigs, unitigs)
-
graph simplifications for assembly (tip removal, bulge removal)
By itself GATB-CORE is not an NGS data analysis tool. However, it can be used to create such tools; see section [Quickly create a new GATB-TOOL software](#Quickly create a new GATB-TOOL software), below.
They already exist a set of ready-to-use tools relying on GATB-CORE library: see https://gatb.inria.fr/software/
All the needed material of GATB-CORE is contained in the current directory in order to generate the wanted artifacts:
-
dynamic and static libraries holding the services component
-
120+ unit tests of the entire library
The following third parties have to be already installed to compile GATB-Core:
- a C++/11 capable compiler (e.g. gcc 4.7+, clang 3.5+, Apple/clang 6.0+)
- CMake 3.10+
In addition, you could install these optional tools:
- doxygen: to compile a local copy of the GATB-Core documentation
- cppunit: to compile and run Unit tests
Type:
cd <some_directory>
git clone https://github.com/GATB/gatb-core.git
cd gatb-core/gatb-core
mkdir build ; cd build ; cmake .. ; make -j8
Type same as above, except for the CMake command:
cmake -D CMAKE_BUILD_TYPE=Debug ..
make -j8
- cppunit is required
- compile using the command above (Release or Debug mode)
Then type:
# enter gatb-core build directory
cd gatb-core/gatb-core/build
# set verbose mode to on so that we have name of failing tests (if any)
export CPPUNIT_VERBOSE=1
# Copy database for unit tests
cp -r ../test/db ./test/
# Launch the full test suite
cd bin
./gatb-core-cppunit
The gatb-core-cppunit command may also take as argument the categories of tests that show up in the verbose output, e.g. './gatb-core-cppunit TestBank'.
More about GATB-CORE code compiling instruction is available here.
Read this documentation.
Read this documentation.
You can follow this link to start the GATB-Core Online Tutorial trail.
The project also contains many code examples that can be easily compiled and executed to review how to use GATB-Core APIs.
The complete GATB-Core documentation is available here. It contains: API, code snippets, compile instructions, etc.
Nevertheless, you can create a local copy of the documentation as follows (we suppose you already compiled the c++ code, see above; requires 'doxygen'):
cd gatb-core/gatb-core/build
make doc
Documentation is then available in build/html/index.html
By default, the library is compiled for supporting 4 ranges of kmers :
- k1 : for kmerSize < k1 (default value 32)
- k2 : for k1 <= kmerSize < k2 (default value 64)
- k3 : for k2 <= kmerSize < k3 (default value 96)
- k4 : for k3 <= kmerSize < k4 (default value 128)
You can customize these values through cmake, provided they rebuild the project from scratch. For instance:
rm -Rf build; mkdir build ; cd build ; cmake -DKSIZE_LIST="64 96 128 162" ..
Tools may set a default kmer lists in their CMakeFiles.txt, as such (see for instance Minia):
list (APPEND KSIZE_DEFAULT_LIST 32 64 96 128 160 192 224 256)
-
README: this file
-
CMakeList.txt: global cmake file
-
doc:
- design design documentation for the component
- doxygen pages for doxygen
-
examples: snippets showing how to use the library
-
src: source code for the component
-
test: tests directory
- src source code for unit tests
- db FASTA databases for unit tests
-
thirdparty: third parties
It contains several sub directories, each one corresponding to one software package.
A package may be composed of sub packages; the directory hierarchy should represent this packages tree structure.
For one atomic package (or sub package), we should have:
-
directory 'api' API of the package
-
directory 'impl' several implementations of the API
A GATB-TOOL is a new software relying upon GATB-CORE.
You use GATB-CORE to create a new tool project, with the following script:
sh scripts/NewProject/NewProject.sh -d directory -n toolName
where:
'directory' is the directory where the project will be created
'toolName' is the name of the project.
The script will automatically creates the full path 'directory/toolName' to deploy a self-contained tool.
By default, the following part will be included in the project:
- a CMakeLists.txt file used for building the project
- a 'tools' directory holding a default source code using GATB-Core
- a 'scripts' directory holding a script to automatically package the tool
- an optional 'thirdparty' directory holding the gatb-core resources
The 'thirdparty' directory is only available for tool created outside the GATB-Tools repository.
Tools located within GATB-Tools rely on a common GATB-CORE sub-module already available in this repository.
The directory where the project is created has no link to any external resources. You can therefore move it anywhere you want.
Such a project can be a start for building applications based on GATB-CORE.
More on creating a new GATB-Core based project: http://gatb-core.gforge.inria.fr/doc/api/new_project.html
#Contact
To contact a developer, request help, etc, use:
https://gatb.inria.fr/contact/
GATB is free software; you can redistribute it and/or modify it under the Affero GPL v3 license.