Ryuto is a tool for exact and fast transcript assembly and quantification, using network flows and a novel extension of splice-graphs.
Cite: Gatter, Thomas, and Peter F. Stadler. "Ryūtō: network-flow based transcriptome reconstruction." BMC bioinformatics 20.1 (2019): 190.
Ryuto is available via bioconda
conda install ryuto
| Operation System | Binary |
|---|---|
| Generic Linux, dynamic linked (recommended) | binary |
| Generic Linux, static libc | binary |
All users have to install g++ in order for the pre-compiled dynamic binaries to work. e.g, for Fedora use:
sudo dnf install gcc-c++
The static version provides libstdc++ and libgcc static linked for linux distributions that cannot provide libraries supporting >= C11.
In its basic usecase, Ryuto needs to be provided only an output directory and the library type used by RNA-Seq: You must provide at least one mapping file in sam or bam format.
ryuto -o [output dir] -l [fr-unstranded | fr-firststrand | fr-secondstrand] <input1.bam> <input2.bam> ...
Please always make sure that your bam files are sorted and index files are available.
We recommnd to use multiple threads via option -t [CPU COUNT].
If an annotation is available for your organism, you may provide it via option -g [GTF].
See
ryuto --help
for a full list of options.
For best results in combination with Ryuto, especially for multi-sample assembly, we strongly recommend mapping reads with STAR with option --outSAMstrandField intronMotif.
3 files will be produced:
- transcripts.gtf: The main result of the assembly. Transcripts are given in GTF format.
- transcripts.count: Counting table produced for e.g. differential transcript expression analysis. Formatted as a table as follows:
Gene-Name, Transcript-Name, Length of the Transcript, Read Count Sample 1 [, Read Count Sample 2 [, Read Count Sample 3 ...]] - transcripts.errcount: If multiple samples are assembled into a consensus, individual inputs may provide signals on how strongly they disagree with it. Higher numbers indicate a higher level of disagreement, e.g. because of undetected fold-changes. Formatted as a table as follows:
Chromosome, Strand, Start, End, Feature (Exon or Splice-Junction), Disagreement Sample 1 [, Disagreement Sample 2 [, Disagreement Sample 3 ...]]
Download the newest source code from: https://github.com/studla/RYUTO
The following additional libraries need to installed in order to run Ryuto: zlib openmp boost htslib
Ryuto provides the following packages with simplified installers. clp lemon
Compile Ryuto with:
./build_all.sh --prefix=[Install Path] [options]
The compiled binary is installed to the specified prefix.
htslib and thus Ryuto relies on zlib. If zlib is not installed on your system, you have to install it first. Download and install zlib from (https://zlib.net/).
Use the following commands to install zlib:
./configure
make
make install
Fedora users may alternatively run:
sudo dnf install zlib-devel
Ubuntu users may alternatively run:
sudo apt-get install zlib1g-dev
You may specify a non-standard zlib installation with --with-zlib=/path/to/your/zlib to the Ryuto ./build_all.sh call.
OpenMP is used for parallelization and needs to be installed in order for this feature to work.
Download and install boost from (http://www.boost.org).
On Fedora you may run:
sudo dnf install boost-devel
On Ubuntu you may run:
sudo apt-get install libboost-all-dev
If installed to a non-standard path, add with --with-boost=/path/to/your/boost to the Ryuto ./build_all.sh call.
Download and install htslib from (https://github.com/samtools/htslib). Install with:
./configure
make
make install
You may alternatively use the following call if bz2 is not installed.
./configure --disable-bz2 --disable-lzma
If installed to a non-standard path, add with --with-htslib=/path/to/your/htslib to the Ryuto ./build_all.sh call.
Clp and Lemon are automatically built and during the ./build_all.sh call. You can find both in the "extern" subfolder.
Lemon needs to be linked to CLP if you want to provide those libraries yourself. If installed to a non-standard path, add with --with-clp=/path/to/your/clp --with-lemon=/path/to/your/lemon and build the project manually.
You can find our in-house scripts to create simulated samples and to use Ryuto counts for DESEQ2 in a seperate repository here.
You may download the alignments of real data used for the first publication here (Size: 82GB).