BRAKER User Guide

Contacts for Github Repository of BRAKER at https://github.com/Gaius-Augustus/BRAKER:

Katharina J. Hoff, University of Greifswald, Germany, [email protected], +49 3834 420 4624

Tomas Bruna, Georgia Tech, U.S.A., [email protected]

Authors of BRAKER

Katharina J. Hoff^{a, b}, Simone Lange^a, Alexandre Lomsadze^c, Tomas Bruna^c, Mark Borodovsky^{c, d, e}, Mario Stanke^{a, b}

[a] University of Greifswald, Institute for Mathematics and Computer Science, Walther-Rathenau-Str. 47, 17489 Greifswald, Germany

[b] University of Greifswald, Center for Functional Genomics of Microbes, Felix-Hausdorff-Str. 8, 17489 Greifswald, Germany

[c] Joint Georgia Tech and Emory University Wallace H Coulter Department of Biomedical Engineering, 30332 Atlanta, USA

[d] School of Computational Science and Engineering, 30332 Atlanta, USA

[e] Moscow Institute of Physics and Technology, Moscow Region 141701, Dolgoprudny, Russia

Figure 1: Current BRAKER authors, from left to right: Mario Stanke, Alexandre Lomsadze, Katharina J. Hoff, Tomas Bruna, and Mark Borodovsky.

Funding

The development of BRAKER was supported by the National Institutes of Health (NIH) [GM128145 to M.B. and M.S.].

Related Software

The Transcript Selector for BRAKER (TSEBRA) is available at https://github.com/Gaius-Augustus/TSEBRA .

Contents

• Authors
• Funding
• What is BRAKER?
• Keys to successful gene prediction
• Overview of modes for running BRAKER
• Installation
  • Supported software versions
  • BRAKER
    • Perl pipeline dependencies
    • BRAKER components
    • Bioinformatics software dependencies
      • Mandatory tools
      • Optional tools
• Running BRAKER
  • BRAKER pipeline modes
    • BRAKER with RNA-Seq data
    • BRAKER with proteins of any evolutionary distance
    • BRAKER with proteins of short evolutionary distance
    • BRAKER with RNA-Seq and protein data
  • Description of selected BRAKER command line options
    • --ab_initio
    • --augustus_args=--some_arg=bla
    • --cores=INT
    • --fungus
    • --softmasking
    • --useexisting
    • --crf
    • --lambda=int
    • --UTR=on
    • --addUTR=on
    • --stranded=+,-,.,...
    • --makehub [email protected]
• Output of BRAKER
• Example data
  • Data description
  • Testing BRAKER with RNA-Seq data
  • Testing BRAKER with proteins of any evolutionary distance
  • Testing BRAKER with proteins of any evolutionary distance and RNA-Seq
  • Testing BRAKER with proteins of close homology
  • Testing BRAKER with proteins of close homology and RNA-Seq data (RNA-Seq supported training)
  • Testing BRAKER with proteins of close homoogy and RNA-Seq data (RNA-Seq and protein supported training)
  • Testing BRAKER with pre-trained parameters
  • Testing BRAKER with genome sequence
• Starting BRAKER on the basis of previously existing BRAKER runs
• Bug reporting
  • Reporting bugs on github
  • Common problems
• Citing BRAKER and software called by BRAKER
• License

What is BRAKER?

The rapidly growing number of sequenced genomes requires fully automated methods for accurate gene structure annotation. With this goal in mind, we have developed BRAKER1^R1R0, a combination of GeneMark-ET ^R2 and AUGUSTUS ^{R3, R4}, that uses genomic and RNA-Seq data to automatically generate full gene structure annotations in novel genome.

However, the quality of RNA-Seq data that is available for annotating a novel genome is variable, and in some cases, RNA-Seq data is not available, at all.

BRAKER2 is an extension of BRAKER1 which allows for fully automated training of the gene prediction tools GeneMark-EX ^{R14, R15, R17, F1} and AUGUSTUS from RNA-Seq and/or protein homology information, and that integrates the extrinsic evidence from RNA-Seq and protein homology information into the prediction.

In contrast to other available methods that rely on protein homology information, BRAKER2 reaches high gene prediction accuracy even in the absence of the annotation of very closely related species and in the absence of RNA-Seq data.

In this user guide, we will refer to BRAKER1 and BRAKER2 simply as BRAKER because they are executed by the same script (braker.pl).

Keys to successful gene prediction

• Use a high quality genome assembly. If you have a huge number of very short scaffolds in your genome assembly, those short scaffolds will likely increase runtime dramatically but will not increase prediction accuracy.

• Use simple scaffold names in the genome file (e.g. >contig1 will work better than >contig1my custom species namesome putative function /more/information/  and lots of special characters %&!*(){}). Make the scaffold names in all your fasta files simple before running any alignment program.

• In order to predict genes accurately in a novel genome, the genome should be masked for repeats. This will avoid the prediction of false positive gene structures in repetitive and low complexitiy regions. Repeat masking is also essential for mapping RNA-Seq data to a genome with some tools (other RNA-Seq mappers, such as HISAT2, ignore masking information). In case of GeneMark-EX and AUGUSTUS, softmasking (i.e. putting repeat regions into lower case letters and all other regions into upper case letters) leads to better results than hardmasking (i.e. replacing letters in repetitive regions by the letter N for unknown nucleotide). If the genome is masked, use the --softmasking flag of braker.pl.

• Many genomes have gene structures that will be predicted accurately with standard parameters of GeneMark-EX and AUGUSTUS within BRAKER. However, some genomes have clade-specific features, i.e. special branch point model in fungi, or non-standard splice-site patterns. Please read the options section [options] in order to determine whether any of the custom options may improve gene prediction accuracy in the genome of your target species.

• Always check gene prediction results before further usage! You can e.g. use a genome browser for visual inspection of gene models in context with extrinsic evidence data. BRAKER supports the generation of track data hubs for the UCSC Genome Browser with MakeHub for this purpose.

Overview of modes for running BRAKER

BRAKER mainly features semi-unsupervised, extrinsic evidence data (RNA-Seq and/or protein spliced alignment information) supported training of GeneMark-EX^[F1] and subsequent training of AUGUSTUS with integration of extrinsic evidence in the final gene prediction step. However, there are now a number of additional pipelines included in BRAKER. In the following, we give an overview of possible input files and pipelines:

• Genome file, only. In this mode, GeneMark-ES is trained on the genome sequence, alone. Long genes predicted by GeneMark-ES are selected for training AUGUSTUS. Final predictions by AUGUSTUS are ab initio. This approach will likely yield lower prediction accuracy than all other here described pipelines. (see Figure 2),

Figure 2: BRAKER pipeline A: training GeneMark-ES on genome data, only; ab initio gene prediction withAUGUSTUS

• Genome and RNA-Seq file from the same species (see figure 3); this approach is suitable for short read RNA-Seq libraries with a good coverage of the transcriptome, important: this approach requires that each intron is covered by many alignments, i.e. it does not work with assembled transcriptome mappings. In principle, also alignments of long read RNA-Seq data may lead to sufficient data for running BRAKER, but only if each transcript that will go into training was sequenced and aligned to the genome multiple times. Please be aware that at the current point in time, BRAKER does not officially support the integration of long read RNA-Seq data, yet.

Figure 3: BRAKER pipeline B: training GeneMark-ET supported by RNA-Seq spliced alignment information, prediction with AUGUSTUS with that same spliced alignment information.

• Genome file and database of proteins that may be of unknown evolutionary distance to the target species (see Figure 4); this approach is particularly suitable if no RNA-Seq data is available. This method will work better with proteins from species that are rather close to the target species, but accuracy will drop only very little if the reference proteins are more distant from the target species. Important: This approach requires a database of protein families, i.e. many representatives of each protein family must be present in the database. BRAKER has been tested with OrthoDB ^R19, successfully. The ProtHint ^R18 protein mapping pipeline for generating required hints for BRAKER is available for download at https://github.com/gatech-genemark/ProtHint, the instructions on how to prepare the OrthoDB input proteins are documented at https://github.com/gatech-genemark/ProtHint#protein-database-preparation. You may add proteins of a closely related species to the OrthoDB fasta file in order to incorporate additional evidence into gene prediction.

Figure 4: BRAKER pipeline C: training GeneMark-EP+ on protein spliced alignment, start and stop information, prediction with AUGUSTUS with that same information, in addition chained CDSpart hints. Proteins used here can be of any evolutionary distance to the target organism.

• Genome and RNA-Seq file from the same species, and proteins that may be of unknown evolutionary distance to the target species (see figure 5); important: this approach requires a database of protein families, i.e. many representatives of each protein family must be present in the database, e.g. OrthoDB is suitable. (You may add proteins of a closely related species to the OrthoDB fasta file in order to incorporate additional evidence into gene prediction.)

Figure 5: BRAKER pipeline D: training GeneMark-ETP+ supported by RNA-Seq alignment information and information from proteins (proteins can be of any evolutionary distance). Please be aware that GeneMark-ETP+ is still under development, BRAKER can currently execute a precursor of the mature version. Introns supported by both RNA-Seq and protein alignment information are treated as “true positive introns”, their prediction in gene structures by GeneMark-ETP+ and AUGUSTUS is enforced. Important: It is not always best to use all evidence! So far, we found this approach to work well for large genomes, but accuracy on small and medium sized genomes is unstable. Please have a look at the poster from PAG 2020 before running this pipeline.

• Genome file and file with proteins of short evolutionary distance (see Figure 6); this approach is suitable if RNA-Seq data is not available and if the reference species is very closely related. Note: This pipeline is deprecated since pipeline C can also use proteins of closely related species in addition to OrthoDB.

Figure 6: Additional pipeline B: training AUGUSTUS on the basis of spliced alignment information from proteins of a very closely related species against the target genome.

• Genome and RNA-Seq file and proteins of short evolutionary distance (see Figures 6 and 7). In both cases, GeneMark-ET is trained supported by RNA-Seq data, and the resulting gene predictions are used for training AUGUSTUS. In approach A), protein alignment information is used in the gene prediction step with AUGUSTUS, only. In approach C), protein spliced alignment data is used to complement the training set for AUGUSTUS. The latter approach is in particular suitable if RNA-Seq data does not produce a sufficiently high number of training gene structures for AUGUSTUS, and if a very closely related and already annotated species is available. Note: This pipeline is deprecated since pipeline D can also use proteins of closely related species in addition to OrthoDB.

Figure 7: Additional pipeline A: training GeneMark-ET supported by RNA-Seq spliced alignment information, prediction with AUGUSTUS with spliced alignment information from RNA-Seq data and with gene features determined by alignments from proteins of a very closely related species against the target genome. Note: This pipeline is deprecated since pipeline C can also use proteins of closely related species in addition to OrthoDB.

Figure 8: Additional pipeline C: training GeneMark-ET on the basis of RNA-Seq spliced alignment information, training AUGUSTUS on a set of training gene structures compiled from RNA-Seq supported gene structures predicted by GeneMark-ET and spliced alignment of proteins of a very closely related species. Note: This pipeline is deprecated since pipeline D can also use proteins of closely related species in addition to OrthoDB.

Installation

Supported software versions

At the time of release, this BRAKER version was tested with:

• AUGUSTUS 3.4.0 ^F2

• GeneMark-ES/ET/EP 4.64_lic

• BAMTOOLS 2.5.1^R5

• SAMTOOLS 1.7-4-g93586ed^R6

• ProtHint 2.6.0

• GenomeThreader 1.7.0^R7

• Spaln 2.3.3d ^{R8, R9, R10, F3}

• (Exonerate 2.2.0 ^R11)^[F3]

• NCBI BLAST+ 2.2.31+ ^{R12, R13}

• DIAMOND 0.9.24

• cdbfasta 0.99

• cdbyank 0.981

• GUSHR 1.0.0

BRAKER

Perl pipeline dependencies

Running BRAKER requires a Linux-system with bash and Perl. Furthermore, BRAKER requires the following CPAN-Perl modules to be installed:

• File::Spec::Functions

• Hash::Merge

• List::Util

• MCE::Mutex

• Module::Load::Conditional

• Parallel::ForkManager

• POSIX

• Scalar::Util::Numeric

• YAML

• Math::Utils

For ProtHint, used when protein input is supplied, also install:

• threads

On Ubuntu, for example, install the modules with CPANminus^F4: sudo cpanm Module::Name, e.g. sudo cpanm Hash::Merge.

BRAKER also uses a Perl module helpMod.pm that is not available on CPAN. This module is part of the BRAKER release and does not require separate installation.

If you do not have root permissions on the Linux machine, try setting up an Anaconda (https://www.anaconda.com/distribution/) environment as follows:

wget https://repo.anaconda.com/archive/Anaconda3-2018.12-Linux-x86_64.sh
bash bin/Anaconda3-2018.12-Linux-x86_64.sh # do not install VS (needs root privileges)
conda install -c anaconda perl
conda install -c bioconda perl-app-cpanminus
conda install -c bioconda perl-hash-merge
conda install -c bioconda perl-parallel-forkmanager
conda install -c bioconda perl-scalar-util-numeric
conda install -c bioconda perl-yaml
conda install -c bioconda perl-class-data-inheritable
conda install -c bioconda perl-exception-class
conda install -c bioconda perl-test-pod
conda install -c anaconda biopython
conda install -c bioconda perl-file-homedir
conda install -c bioconda perl-file-which # skip if you are not comparing to reference annotation
conda install -c bioconda perl-mce
conda install -c bioconda perl-threaded
conda install -c bioconda perl-list-util
conda install -c bioconda perl-list-moreutils
conda install -c bioconda perl-math-utils
conda install -c bioconda cdbtools

Subsequently install BRAKER and other software "as usual" while being in your conda environment. Note: There is a bioconda braker package, and a bioconda augustus package. They work. But they are usually lagging behind the development code of both tools on github. We therefore recommend manual installation and usage of lastest sources.

BRAKER components

BRAKER is a collection of Perl and Python scripts and a Perl module. The main script that will be called in order to run BRAKER is braker.pl. Additional Perl and Python components are:

• align2hints.pl

• filterGenemark.pl

• filterIntronsFindStrand.pl

• startAlign.pl

• helpMod.pm

• findGenesInIntrons.pl

• downsample_traingenes.pl

• ensure_n_training_genes.py

All scripts (files ending with *.pl and *.py) that are part of BRAKER must be executable in order to run BRAKER. This should already be the case if you download BRAKER from GitHub. Executability may be overwritten if you e.g. transfer BRAKER on a USB-stick to another computer. In order to check whether required files are executable, run the following command in the directory that contains BRAKER Perl scripts:

ls -l *.pl *.py

The output should be similar to this:

    -rwxr-xr-x 1 katharina katharina  18191 Mai  7 10:25 align2hints.pl
    -rwxr-xr-x 1 katharina katharina   6090 Feb 19 09:35 braker_cleanup.pl
    -rwxr-xr-x 1 katharina katharina 408782 Aug 17 18:24 braker.pl
    -rwxr-xr-x 1 katharina katharina   5024 Mai  7 10:25 downsample_traingenes.pl
    -rwxr-xr-x 1 katharina katharina   5024 Mai  7 10:23 ensure_n_training_genes.py
    -rwxr-xr-x 1 katharina katharina   4542 Apr  3  2019 filter_augustus_gff.pl
    -rwxr-xr-x 1 katharina katharina  30453 Mai  7 10:25 filterGenemark.pl
    -rwxr-xr-x 1 katharina katharina   5754 Mai  7 10:25 filterIntronsFindStrand.pl
    -rwxr-xr-x 1 katharina katharina   7765 Mai  7 10:25 findGenesInIntrons.pl
    -rwxr-xr-x 1 katharina katharina   1664 Feb 12  2019 gatech_pmp2hints.pl
    -rwxr-xr-x 1 katharina katharina   2250 Jan  9 13:55 log_reg_prothints.pl
    -rwxr-xr-x 1 katharina katharina   4679 Jan  9 13:55 merge_transcript_sets.pl
    -rwxr-xr-x 1 katharina katharina  41674 Mai  7 10:25 startAlign.pl

It is important that the x in -rwxr-xr-x is present for each script. If that is not the case, run

`chmod a+x *.pl *.py`

in order to change file attributes.

You may find it helpful to add the directory in which BRAKER perl scripts reside to your $PATH environment variable. For a single bash session, enter:

    PATH=/your_path_to_braker/:$PATH
    export PATH

To make this $PATH modification available to all bash sessions, add the above lines to a startup script (e.g.~/.bashrc).

Bioinformatics software dependencies

BRAKER calls upon various bioinformatics software tools that are not part of BRAKER. Some tools are obligatory, i.e. BRAKER will not run at all if these tools are not present on your system. Other tools are optional. Please install all tools that are required for running BRAKER in the mode of your choice.

Mandatory tools

GeneMark-EX

Download GeneMark-EX^F1 from http://exon.gatech.edu/GeneMark/license_download.cgi (the GeneMark-ES/ET/EP) option. Unpack and install GeneMark-EX as described in GeneMark-EX’s README file.

If already contained in your $PATH variable, BRAKER will guess the location of gmes_petap.pl, automatically. Otherwise, BRAKER can find GeneMark-EX executables either by locating them in an environment variable GENEMARK_PATH, or by taking a command line argument (--GENEMARK_PATH=/your_path_to_GeneMark-EX/).

In order to set the environment variable for your current Bash session, type:

export GENEMARK_PATH=/your_path_to_GeneMark-EX/

Add the above lines to a startup script (e.g. ~/.bashrc) in order to make it available to all bash sessions.^F5

Important: GeneMark-EX will only run if a valid key file resides in your home directory. The key file will expire after 200 days, which means that you have to download a new GeneMark-EX release and a new key file after 200 days. The key file is downloaded as gm_key.gz. Unpack the key file and move it to a hidden file in your home directory as follows:

cd # change to your home directory
gunzip gm_key_64.gz
mv gm_key_64 .gm_key

Perl scripts within GeneMark-EX are configured with default Perl location at /usr/bin/perl.

If you are running GeneMark-EX in an Anaconda environment (or want to use Perl from the $PATH variable for any other reason), modify the shebang of all GeneMark-EX scripts with the following command located inside GeneMark-EX folder:

perl change_path_in_perl_scripts.pl "/usr/bin/env perl"

You can check whether GeneMark-EX is installed properly by running the check_install.bash and/or executing examples in GeneMark-E-tests directory.

AUGUSTUS

Download AUGUSTUS from its master branch at https://github.com/Gaius-Augustus/Augustus. Unpack AUGUSTUS and install AUGUSTUS according to AUGUSTUS README.TXT. Do not use outdated AUGUSTUS versions from other sources, e.g. Debian package or Bioconda package! BRAKER highly depends in particular on an up-to-date Augustus/scripts directory, and other sources are often lagging behind.

You should compile AUGUSTUS on your own system in order to avoid problems with versions of libraries used by AUGUSTUS. Compilation instructions are provided in the AUGUSTUS README.TXT file (Augustus/README.txt).

AUGUSTUS consists of augustus, the gene prediction tool, additional C++ tools located in Augustus/auxprogs and Perl scripts located in Augustus/scripts. Perl scripts must be executable (see instructions in section BRAKER components.

The C++ tool bam2hints is an essential component of BRAKER when run with RNA-Seq. Sources are located in Augustus/auxprogs/bam2hints. Make sure that you compile bam2hints on your system (it should be automatically compiled when AUGUSTUS is compiled, but in case of problems with bam2hints, please read troubleshooting instructions in Augustus/auxprogs/bam2hints/README).

Since BRAKER is a pipeline that trains AUGUSTUS, i.e. writes species specific parameter files, BRAKER needs writing access to the configuration directory of AUGUSTUS that contains such files (Augustus/config/). If you install AUGUSTUS globally on your system, the config folder will typically not be writable by all users. Either make the directory where config resides recursively writable to users of AUGUSTUS, or copy the config/ folder (recursively) to a location where users have writing permission.

AUGUSTUS will locate the config folder by looking for an environment variable $AUGUSTUS_CONFIG_PATH. If the $AUGUSTUS_CONFIG_PATH environment variable is not set, then BRAKER will look in the path ../config relative to the directory in which it finds an AUGUSTUS executable. Alternatively, you can supply the variable as a command line argument to BRAKER (--AUGUSTUS_CONFIG_PATH=/your_path_to_AUGUSTUS/Augustus/config/). We recommend that you export the variable e.g. for your current bash session:

    export AUGUSTUS_CONFIG_PATH=/your_path_to_AUGUSTUS/Augustus/config/

In order to make the variable available to all Bash sessions, add the above line to a startup script, e.g. ~/.bashrc.

Important:

BRAKER expects the entire config directory of AUGUSTUS at $AUGUSTUS_CONFIG_PATH, i.e. the subfolders species with its contents (at least generic) and extrinsic! Providing a writable but empty folder at $AUGUSTUS_CONFIG_PATH will not work for BRAKER. If you need to separate augustus binary and $AUGUSTUS_CONFIG_PATH, we recommend that you recursively copy the un-writable config contents to a writable location.

If you have a system-wide installation of AUGUSTUS at /usr/bin/augustus, an unwritable copy of config sits at /usr/bin/augustus_config/. The folder /home/yours/ is writable to you. Copy with the following command (and additionally set the then required variables):

cp -r /usr/bin/Augustus/config/ /home/yours/
export AUGUSTUS_CONFIG_PATH=/home/yours/augustus_config
export AUGUSTUS_BIN_PATH=/usr/bin
export AUGUSTUS_SCRIPTS_PATH=/usr/bin/augustus_scripts

Modification of $PATH

Adding directories of AUGUSTUS binaries and scripts to your $PATH variable enables your system to locate these tools, automatically. It is not a requirement for running BRAKER to do this, because BRAKER will try to guess them from the location of another environment variable ($AUGUSTUS_CONFIG_PATH), or both directories can be supplied as command line arguments to braker.pl, but we recommend to add them to your $PATH variable. For your current bash session, type:

    PATH=:/your_path_to_augustus/bin/:/your_path_to_augustus/scripts/:$PATH
    export PATH

For all your BASH sessions, add the above lines to a startup script (e.g.~/.bashrc).

Python3

On Ubuntu, Python3 is usually installed by default, python3 will be in your $PATH variable, by default, and BRAKER will automatically locate it. However, you have the option to specify the python3 binary location in two other ways:

1. Export an environment variable $PYTHON3_PATH, e.g. in your ~/.bashrc file:

   export PYTHON3_PATH=/path/to/python3/
   

2. Specify the command line option --PYTHON3_PATH=/path/to/python3/ to braker.pl.

Bamtools

Download BAMTOOLS (e.g. git clone https://github.com/pezmaster31/bamtools.git). Install BAMTOOLS by typing the following in your shell:

    cd your-bamtools-directory mkdir build cd build cmake .. make

If already in your $PATH variable, BRAKER will find bamtools, automatically. Otherwise, BRAKER can locate the bamtools binary either by using an environment variable $BAMTOOLS_PATH, or by taking a command line argument (--BAMTOOLS_PATH=/your_path_to_bamtools/bin/^F6). In order to set the environment variable e.g. for your current bash session, type:

    export BAMTOOLS_PATH=/your_path_to_bamtools/bin/

Add the above line to a startup script (e.g. ~/.bashrc) in order to set the environment variable for all bash sessions.

NCBI BLAST+ or DIAMOND

You can use either NCBI BLAST+ or DIAMOND for removal of redundant training genes. You do not need both tools. If DIAMOND is present, it will be preferred because it is much faster.

Obtain and unpack DIAMOND as follows:

    wget http://github.com/bbuchfink/diamond/releases/download/v0.9.24/diamond-linux64.tar.gz
    tar xzf diamond-linux64.tar.gz

If already in your $PATH variable, BRAKER will find diamond, automatically. Otherwise, BRAKER can locate the diamond binary either by using an environment variable $DIAMOND_PATH, or by taking a command line argument (--DIAMOND_PATH=/your_path_to_diamond). In order to set the environment variable e.g. for your current bash session, type:

    export DIAMOND_PATH=/your_path_to_diamond/

Add the above line to a startup script (e.g. ~/.bashrc) in order to set the environment variable for all bash sessions.

If you decide for BLAST+, install NCBI BLAST+ with sudo apt-get install ncbi-blast+.

If already in your $PATH variable, BRAKER will find blastp, automatically. Otherwise, BRAKER can locate the blastp binary either by using an environment variable $BLAST_PATH, or by taking a command line argument (--BLAST_PATH=/your_path_to_blast/). In order to set the environment variable e.g. for your current bash session, type:

    export BLAST_PATH=/your_path_to_blast/

Add the above line to a startup script (e.g. ~/.bashrc) in order to set the environment variable for all bash sessions.

ProtHint

ProtHint is a pipeline for generating hints for GeneMark-EX and AUGUSTUS from proteins of any evolutionary distance. If protein sequences are given on input, BRAKER runs ProtHint automatically. Alternatively, ProtHint can be executed as a separate step during data preparation. ProtHint is available from https://github.com/gatech-genemark/ProtHint. Download as follows:

git clone https://github.com/gatech-genemark/ProtHint.git

or by getting the latest release from https://github.com/gatech-genemark/ProtHint/releases.

ProtHint has software requirements of its own. In addition to the Perl modules required by BRAKER, it needs

threads

You can easily verify ProtHint's installation by running the test in https://github.com/gatech-genemark/ProtHint/tree/master/example.

ProtHint requires DIAMOND and Spaln, both of which come with ProtHint's installation. ProtHint's requirement of GeneMark-ES will already be fulfilled if you installed BRAKER dependencies above. For further installation instructions, please check https://github.com/gatech-genemark/ProtHint.

If already in your $PATH variable, BRAKER will find prothint.py, automatically. Otherwise, BRAKER will try to locate the prothint.py executable by using an environment variable $PROTHINT_PATH. Alternatively, this can be supplied as a command line argument --PROTHINT_PATH=/your/path/to/ProtHint/bin.

Optional tools

Samtools

Samtools is not required for running BRAKER if all your files are formatted, correctly (i.e. all sequences should have short and unique fasta names). If you are not sure whether all your files are fomatted correctly, it might be helpful to have Samtools installed because BRAKER can automatically fix certain format issues by using Samtools.

As a prerequisite for Samtools, download and install htslib (e.g. git clone https://github.com/samtools/htslib.git, follow the htslib documentation for installation).

Download and install Samtools (e.g. git clone git://github.com/samtools/samtools.git), subsequently follow Samtools documentation for installation).

If already in your $PATH variable, BRAKER will find samtools, automatically. Otherwise, BRAKER can find Samtools either by taking a command line argument (--SAMTOOLS_PATH=/your_path_to_samtools/), or by using an environment variable $SAMTOOLS_PATH. For exporting the variable, e.g. for your current bash session, type:

    export SAMTOOLS_PATH=/your_path_to_samtools/

Add the above line to a startup script (e.g. ~/.bashrc) in order to set the environment variable for all bash sessions.

Biopython

If Biopython is installed, BRAKER can generate FASTA-files with coding sequences and protein sequences predicted by AUGUSTUS and generate track data hubs for visualization of a BRAKER run with MakeHub ^R16. These are optional steps. The first can be disabled with the command-line flag --skipGetAnnoFromFasta, the second can be activated by using the command-line options --makehub [email protected], Biopython is not required if neither of these optional steps shall be performed.

On Ubuntu, install Python3 package manager with:

`sudo apt-get install python3-pip`

Then, install Biopython with:

`sudo pip3 install biopython`

cdbfasta

cdbfasta and cdbyank are required by BRAKER for correcting AUGUSTUS genes with in frame stop codons (spliced stop codons) using the AUGUSTUS script fix_in_frame_stop_codon_genes.py. This can be skipped with --skip_fixing_broken_genes.

On Ubuntu, install cdbfasta with:

`sudo apt-get install cdbfasta`

For other systems, you can for example obtain cdbfasta from https://github.com/gpertea/cdbfasta, e.g.:

        git clone https://github.com/gpertea/cdbfasta.git`
        cd cdbfasta
        make all

On Ubuntu, cdbfasta and cdbyank will be in your $PATH variable after installation, and BRAKER will automatically locate them. However, you have the option to specify the cdbfasta and cdbyank binary location in two other ways:

1. Export an environment variable $CDBTOOLS_PATH, e.g. in your ~/.bashrc file:

        export CDBTOOLS_PATH=/path/to/cdbtools/

2. Specify the command line option --CDBTOOLS_PATH=/path/to/cdbtools/ to braker.pl.

GenomeThreader

Note: Support of GenomeThreader within BRAKER is deprecated.

This tool is required, only, if you would like to run protein to genome alignments with BRAKER using GenomeThreader. This is a suitable approach only if an annotated species of short evolutionary distance to your target genome is available. Download GenomeThreader from http://genomethreader.org/. Unpack and install according to gth/README.

BRAKER will try to locate the GenomeThreader executable by using an environment variable $ALIGNMENT_TOOL_PATH. Alternatively, this can be supplied as command line argument (--ALIGNMENT_TOOL_PATH=/your/path/to/gth).

Spaln

Note: Support of stand-alone Spaln (ouside of ProtHint) within BRAKER is deprecated.

This tool is required if you run ProtHint or if you would like to run protein to genome alignments with BRAKER using Spaln outside of ProtHint. Using Spaln outside of ProtHint is a suitable approach only if an annotated species of short evolutionary distance to your target genome is available. We recommend running Spaln through ProtHint for BRAKER. ProtHint brings along a Spaln binary. If that does not work on your system, download Spaln from https://github.com/ogotoh/spaln. Unpack and install according to spaln/doc/SpalnReadMe22.pdf.

BRAKER will try to locate the Spaln executable by using an environment variable $ALIGNMENT_TOOL_PATH. Alternatively, this can be supplied as command line argument (--ALIGNMENT_TOOL_PATH=/your/path/to/spaln).

Exonerate

Note: Support of Exonerate within BRAKER is deprecated.

This tool is required, only, if you would like to run protein to genome alignments with BRAKER using Exonerate. This is a suitable approach only if an annotated species of short evolutionary distance to your target genome is available. (We recommend the usage of GenomeThreader instad of Exonerate because Exonerate is comparably slower and has lower specificity than GenomeThreader.) Download Exonerate from https://github.com/nathanweeks/exonerate. Unpack and install according to exonerate/README. (On Ubuntu, download and install by typing sudo apt-get install exonerate.)

BRAKER will try to locate the Exonerate executable by using an environment variable $ALIGNMENT_TOOL_PATH. Alternatively, this can be supplied as command line argument (--ALIGNMENT_TOOL_PATH=/your/path/to/exonerate).

GUSHR

This tool is only required if you want either add UTRs (from RNA-Seq data) to predicted genes or if you want to train UTR parameters for AUGUSTUS and predict genes with UTRs. In any case, GUSHR requires the input of RNA-Seq data.

GUSHR is available for download at https://github.com/Gaius-Augustus/GUSHR. Obtain it by typing:

    git clone https://github.com/Gaius-Augustus/GUSHR.git

GUSHR executes a GeMoMa jar file ^{R19, R20, R21}, and this jar file requires Java 1.8. On Ubuntu, you can install Java 1.8 with the following command:

sudo apt-get install openjdk-8-jdk

If you have several java versions installed on your system, make sure that you enable 1.8 prior running BRAKER with java by running

sudo update-alternatives --config java 

and selecting the correct version.

Tools from UCSC

If you switch --UTR=on, bamToWig.py will require the following tools that can be downloaded from http://hgdownload.soe.ucsc.edu/admin/exe:

• twoBitInfo

• faToTwoBit

It is optional to install these tools into your $PATH. If you don't, and you switch --UTR=on, bamToWig.py will automatically download them into the working directory.

MakeHub

If you wish to automaticaly generate a track data hub of your BRAKER run, the MakeHub software, available at https://github.com/Gaius-Augustus/MakeHub is required. Download the software (either by running git clone https://github.com/Gaius-Augustus/MakeHub.git, or by picking a release from https://github.com/Gaius-Augustus/MakeHub/releases. Extract the release package if you downloaded a release (e.g. unzip MakeHub.zip or tar -zxvf MakeHub.tar.gz.

BRAKER will try to locate the make_hub.py script by using an environment variable $MAKEHUB_PATH. Alternatively, this can be supplied as command line argument (--MAKEHUB_PATH=/your/path/to/MakeHub/). BRAKER can also try to guess the location of MakeHub on your system.

Running BRAKER

Different BRAKER pipeline modes

In the following, we describe “typical” BRAKER calls for different input data types. In general, we recommend that you run BRAKER on genomic sequences that have been softmasked for Repeats. If your genome has been softmasked, include the --softmasking flag in your BRAKER call!

BRAKER with RNA-Seq data

This approach is suitable for genomes of species for which RNA-Seq libraries with a good coverage of the transcriptome are available. The pipeline is illustrated in Figure 2.

BRAKER can either extract RNA-Seq spliced alignment information from bam files, or it can use such extracted information, directly.

In order to run BRAKER with RNA-Seq data supplied as bam file(s) (in case of multiple files, separate them by comma), run:

    braker.pl --species=yourSpecies --genome=genome.fasta \
       --bam=file1.bam,file2.bam

In order to run BRAKER with RNA-Seq spliced alignment information that has already been extracted, run:

    braker.pl --species=yourSpecies --genome=genome.fasta \
       --hints=hints1.gff,hints2.gff

The format of such a hints file must be as follows (tabulator separated file):

    chrName b2h intron  6591    8003    1   +   .   pri=4;src=E
    chrName b2h intron  6136    9084    11  +   .   mult=11;pri=4;src=E
    ...

The source b2h in the second column and the source tag src=E in the last column are essential for BRAKER to determine whether a hint has been generated from RNA-Seq data.

BRAKER with proteins of any evolutionary distance

This approach is suitable for genomes of species for which no RNA-Seq libraries are available. A large database of proteins (with possibly longer evolutionary distance to the target species) should be used in this case. This mode is illustrated in figure 9.

Figure 9: BRAKER with proteins of any evolutionary distance. ProtHint protein mapping pipelines is used to generate protein hints. ProtHint automatically determines which alignments are from close relatives, and which are from rather distant relatives.

For running BRAKER in this mode, type:

braker.pl --genome=genome.fa --prot_seq=proteins.fa --softmasking

We recommend using OrthoDB as basis for proteins.fa. The instructions on how to prepare the input OrthoDB proteins are documented here: https://github.com/gatech-genemark/ProtHint#protein-database-preparation.

You can of course add additional protein sequences to that file, or try with a completely different database. Any database will need several representatives for each protein, though.

Instead of having BRAKER run ProtHint, you can also start BRAKER with hints already produced by ProtHint, by providing ProtHint's prothint_augustus.gff output:

braker.pl --genome=genome.fa --hints=prothint_augustus.gff --softmasking

The format of prothint_augustus.gff in this mode looks like this:

2R ProtHint intron 11506230 11506648 4 + . src=M;mult=4;pri=4
2R ProtHint intron 9563406  9563473  1 + . grp=69004_0:001de1_702_g;src=C;pri=4;
2R ProtHint intron 8446312  8446371  1 + . grp=43151_0:001cae_473_g;src=C;pri=4;
2R ProtHint intron 8011796  8011865  2 - . src=P;mult=1;pri=4;al_score=0.12;
2R ProtHint start  234524   234526   1 + . src=P;mult=1;pri=4;al_score=0.08;

The prediction of all hints with src=M will be enforced. Hints with src=C are 'chained evidence', i.e. they will only be incorporated if all members of the group (grp=...) can be incorporated in a single transcript. All other hints have src=P in the last column. Supported features in column 3 are intron, start, stop and CDSpart.

Training and prediction of UTRs, integration of coverage information

If RNA-Seq (and only RNA-Seq) data is provided to BRAKER as a bam-file, and if the genome is softmasked for repeats, BRAKER can automatically train UTR parameters for AUGUSTUS. After successful training of UTR parameters, BRAKER will automatically predict genes including coverage information form RNA-Seq data. Example call:

braker.pl --species=yourSpecies --genome=genome.fasta \
   --bam=file.bam --softmasking --UTR=on

Warnings:

1. This feature is experimental!

2. --UTR=on is currently not compatible with bamToWig.py as released in AUGUSTUS 3.3.3; it requires the current development code version from the github repository (git clone https://github.com/Gaius-Augustus/Augustus.git).

3. --UTR=on increases memory consumption of AUGUSTUS. Carefully monitor jobs if your machine was close to maxing RAM without --UTR=on! Reducing the number of cores will also reduce RAM consumption.

4. UTR prediction sometimes improves coding sequence prediction accuracy, but not always. If you try this feature, carefully compare results with and without UTR parameters, afterwards (e.g. in UCSC Genome Browser).

Stranded RNA-Seq alignments

For running BRAKER without UTR parameters, it is not very important whether RNA-Seq data was generated by a stranded protocol (because spliced alignments are ’artificially stranded’ by checking the splice site pattern). However, for UTR training and prediction, stranded libraries may provide information that is valuable for BRAKER.

After alignment of the stranded RNA-Seq libraries, separate the resulting bam file entries into two files: one for plus strand mappings, one for minus strand mappings. Call BRAKER as follows:

braker.pl --species=yourSpecies --genome=genome.fasta \
   --softmasking --bam=plus.bam,minus.bam --stranded=+,- \
    --UTR=on

You may additionally include bam files from unstranded libraries. Those files will not used for generating UTR training examples, but they will be included in the final gene prediction step as unstranded coverage information, example call:

braker.pl --species=yourSpecies --genome=genome.fasta \
   --softmasking --bam=plus.bam,minus.bam,unstranded.bam \
   --stranded=+,-,. --UTR=on

Warning: This feature is experimental and currently has low priority on our maintenance list!

BRAKER with proteins of short evolutionary distance

This is a deprecated pipeline that was before the only suitable approach if RNA-Seq data for the species of the target genome is not available and if a well annotated and very closely related reference species is available and you don't want to use the approach for proteins of any evolutionary distance (where we back then weren't sure how it'd perform on proteins of short evolutionary distance).

For running BRAKER in this mode, type:

    braker.pl --species=yourSpecies --genome=genome.fasta \
       --prot_seq=proteins.fa --prg=gth \
       --ALIGNMENT_TOOL_PATH=/path/to/gth/binary \
       --trainFromGth

It is possible to generate protein alignments externally, prior running BRAKER, itself. The compatible command for running GenomeThreader prior running BRAKER, is:

    gth -genomic genome.fa  -protein protein.fa -gff3out \
       -skipalignmentout -o gth.aln

In order to use such externally created alignment files, run:

    braker.pl --species=yourSpecies --genome=genome.fasta \
       --prot_aln=proteins.aln --prg=gth --trainFromGth

It is also possible to run BRAKER in this mode using an already prepared hints file. In this case, run:

    braker.pl --species=yourSpecies --genome=genome.fasta \
       --hints=hints.gff --prg=gth --trainFromGth

Format of the hints file should look like this:

    chrName   gth2h   CDSpart 105984  106633  .     -    .    src=P;grp=FBpp0285205;pri=4
    chrName   gth2h   start   106646  106648  .     -    .    src=P;grp=FBpp0285205;pri=4

Supported features in column 3 are intron, CDSpart, start, stop.

BRAKER with RNA-Seq and protein data

Even though BRAKER supports the combination of RNA-Seq and protein data within the BRAKER pipeline, we strongly recommend to run BRAKER twice (1x with RNA-Seq only, 1x with protein data only) and subsequently combine the results of both runs with TSEBRA, the Transcript Selector for BRAKER (https://github.com/Gaius-Augustus/TSEBRA). You find more information on TSEBRA at https://www.biorxiv.org/content/10.1101/2021.06.07.447316v1

In the following, we describe that it is possible to combine both data sources with BRAKER, alone, for the sake of completeness.

The native mode for running BRAKER with RNA-Seq and protein data is --etpmode. This will call GeneMark-ETP (which is currently only available as a premature version by using current state GeneMark-ES/ET/EP/EP+, improvements are to be expected, soon), which will use RNA-Seq and protein hints for training GeneMark-ETP. Hints that are supported by both sources and proteins hints of particularly high quality are enforced in gene prediction with GeneMark-ETP. Subsequently, AUGUSTUS is trained on GeneMark-ETP predictions and genes with hints are predicted by AUGUSTUS. To call the pipeline in this mode, run:

    braker.pl --genome=genome.fa --prot_seq=orthodb.fa \
       --hints=rnaseq.hints --etpmode --softmasking

You can or course replace the rnaseq.gff hints file by a BAM-file, e.g. --bam=ranseq.bam.

In addition, the following pipelines can be executed by BRAKER (deprecated pipelines):

• Adding protein data of short evolutionary distance to gene prediction step

• Extending training gene set with proteins of short evolutionary distance

Adding protein data of short evolutionary distance to gene prediction step

This pipeline is illustrated in Figure 7.

In general, add the options

       --prot_seq=proteins.fa --prg=(gth|exonerate|spaln)

to the BRAKER call that is described in section BRAKER with RNA-Seq data. Select one protein alignment tool from GenomeThreader (gth, recommended), Spaln (spaln) or Exonerate (exonerate). Of course, you may also specify the protein information as protein alignment files or hints files as described in section BRAKER with proteins of short evolutionary distance). This may result in a call similar to:

    braker.pl --species=yourSpecies --genome=genome.fasta \
       --bam=file1.bam,file2.bam --prot_seq=proteins.fa \
       --prg=(gth|exonerate|spaln)

Extending training gene set with proteins of short evolutionary distance

If the number of training gene structures identified by RNA-Seq data, only, seems to be too small, you may add training gene structures generated by protein alignments with GenomeThreader to the training gene set. This pipeline is illustrated in Figure 8.

In general, add the options

       --prot_seq=proteins.fa --prg=gth --gth2traingenes

to the BRAKER call that is described in section BRAKER with RNA-Seq data. This may result in a call similar to:

    braker.pl --species=yourSpecies --genome=genome.fasta \
       --bam=file1.bam,file2.bam --prot_seq=proteins.fa \
       --prg=gth --gth2traingenes

Description of selected BRAKER command line options

Please run braker.pl --help to obtain a full list of options.

--epmode

Run BRAKER in EP-mode, i.e. with proteins of any evolutionary distance as processed by ProtHint within BRAKER. This mode is turned on by default when only protein input is detected. Should be provided with either --prot_seq=orthodb.fa or protein hints --hints=prothint_augustus.gff.

--etpmode

Run BRAKER in ETP-mode, i.e. with proteins of any evolutionary distance processed by ProtHint, and with RNA-Seq data. Should to be provided with prot_seq=orthodb.fa and --bam=rnaseq.bam. Alternatively, the RNA-Seq and protein hints can be provided as processed hints with the --hints opiton. Please consider using TSEBRA (https://github.com/Gaius-Augustus/TSEBRA) instead of BRAKER in ETP-mode.

--ab_initio

Compute AUGUSTUS ab initio predictions in addition to AUGUSTUS predictions with hints (additional output files: augustus.ab_initio.*. This may be useful for estimating the quality of training gene parameters when inspecting predictions in a Browser.

--augustus_args="--some_arg=bla"

One or several command line arguments to be passed to AUGUSTUS, if several arguments are given, separate them by whitespace, i.e. "--first_arg=sth --second_arg=sth". This may be be useful if you know that gene prediction in your particular species benefits from a particular AUGUSTUS argument during the prediction step.

--cores=INT

Specifies the maximum number of cores that can be used during computation. BRAKER has to run some steps on a single core, others can take advantage of multiple cores. If you use more than 8 cores, this will not speed up all parallelized steps, in particular, the time consuming optimize_augustus.pl will not use more than 8 cores. However, if you don’t mind some cores being idle, using more than 8 cores will speed up other steps.

--fungus

GeneMark-EX option: run algorithm with branch point model. Use this option if you genome is a fungus.

--softmasking

Softmasking option for soft masked genome files. (Disabled by default.)

--useexisting

Use the present config and parameter files if they exist for 'species'; will overwrite original parameters if BRAKER performs an AUGUSTUS training.

--crf

Execute CRF training for AUGUSTUS; resulting parameters are only kept for final predictions if they show higher accuracy than HMM parameters. This increases runtime!

--lambda=int

Change the parameter $\lambda$ of the Poisson distribution that is used for downsampling training genes according to their number of introns (only genes with up to 5 introns are downsampled). The default value is $\lambda=2$. You might want to set it to 0 for organisms that mainly have single-exon genes. (Generally, single-exon genes contribute less value to increasing AUGUSTUS parameters compared to genes with many exons.)

--UTR=on

Generate UTR training examples for AUGUSTUS from RNA-Seq coverage information, train AUGUSTUS UTR parameters and predict genes with AUGUSTUS and UTRs, including coverage information for RNA-Seq as evidence. This flag only works if --softmasking is also enabled. This is an experimental feature!

If you performed a BRAKER run without --UTR=on, you can add UTR parameter training and gene prediction with UTR parameters (and only RNA-Seq hints) with the following command:

braker.pl --genome=../genome.fa --addUTR=on --softmasking \
    --bam=../RNAseq.bam --workingdir=$wd \
    --AUGUSTUS_hints_preds=augustus.hints.gtf \
    --cores=8 --skipAllTraining --species=somespecies

Modify augustus.hints.gtf to point to the AUGUSTUS predictions with hints from previous BRAKER run; modify flaning_DNA value to the flanking region from the log file of your previous BRAKER run; modify some_new_working_directory to the location where BRAKER should store results of the additional BRAKER run; modify somespecies to the species name used in your previous BRAKER run.

--addUTR=on

Add UTRs from RNA-Seq converage information to AUGUSTUS gene predictions using GUSHR. No training of UTR parameters and no gene prediction with UTR parameters is performed.

If you performed a BRAKER run without --addUTR=on, you can add UTRs results of a previous BRAKER run with the following command:

braker.pl --genome=../genome.fa --addUTR=on --softmasking \
    --bam=../RNAseq.bam --workingdir=$wd \
    --AUGUSTUS_hints_preds=augustus.hints.gtf --cores=8 \
    --skipAllTraining --species=somespecies

Modify augustus.hints.gtf to point to the AUGUSTUS predictions with hints from previous BRAKER run; modify some_new_working_directory to the location where BRAKER should store results of the additional BRAKER run; this run will not modify AUGUSTUS parameters. We recommend that you specify the original species of the original run with --species=somespecies. Otherwise, BRAKER will create an unneeded species parameters directory Sp_*.

--stranded=+,-,.,...

If --UTR=on is enabled, strand-separated bam-files can be provided with --bam=plus.bam,minus.bam. In that case, --stranded=... should hold the strands of the bam files (+ for plus strand, - for minus strand, . for unstranded). Note that unstranded data will be used in the gene prediction step, only, if the parameter --stranded=... is set. This is an experimental feature! GUSHR currently does not take advantage of stranded data.

--makehub --email=[email protected]

If --makehub and [email protected] (with your valid e-mail adress) are provided, a track data hub for visualizing results with the UCSC Genome Browser will be generated using MakeHub (https://github.com/Gaius-Augustus/MakeHub).

--gc_probability=DECIMAL

By default, GeneMark-EX uses a probability of 0.001 for predicting the donor splice site pattern GC (instead of GT). It may make sense to increase this value for species where this donor splice site is more common. For example, in the species Emiliania huxleyi, about 50% of donor splice sites have the pattern GC (https://media.nature.com/original/nature-assets/nature/journal/v499/n7457/extref/nature12221-s2.pdf, page 5).

Output of BRAKER

BRAKER produces several important output files in the working directory.

• augustus.hints.gtf: Genes predicted by AUGUSTUS with hints from given extrinsic evidence. This file will be missing if BRAKER was run with the option --esmode.

• augustus.hints_utr.gtf: This file may contain different contents depending on how you called BRAKER:

  • If you ran BRAKER with --UTR=on, then this file will contain genes predicted by AUGUSTUS with UTR parameters and coverage information from RNA-Seq data in GTF format.

  • If you ran BRAKER with --addUTR=on, then this file will contain genes predicted by AUGUSTUS without UTR parameters and without coverage information from RNA-Seq data. Instead, AUGUSTUS gene predictions with hints will only be extended by UTRs if RNA-Seq coverage allows it (i.e. no separate AUGUSTUS training or run was performed, UTRs are only added from running GUSHR). Genes in are in GTF format.

This file will only be present if BRAKER was executed with the options --UTR=on or --addUTR=on and a RNA-Seq BAM file.

• augustus.ab_initio.gtf: Genes predicted by AUGUSTUS in ab initio mode in GTF-format. The file will always be present if AUGUSTUS has been run with the option --esmode. Otherwise, it will only be present if BRAKER was run with the option --AUGUSTUS_ab_initio.

• augustus.ab_initio_utr.gtf: This file may contain gene predictions with UTRs if you ran BRAKER with --UTR=on.

This file will only be present if BRAKER was executed with the options --UTR=on or --addUTR=on and a RNA-Seq BAM-file, and with the option --AUGUSTUS_ab_initio.

• GeneMark-E*/genemark.gtf: Genes predicted by GeneMark-ES/ET/EP/EP+ in GTF-format. This file will be missing if BRAKER was executed with proteins of close homology and the option --trainFromGth.

• braker.gtf: Union of augustus.hints.gtf and reliable GeneMark-EX predictions (genes fully supported by external evidence). In --esmode, this is the union of augustus.ab_initio.gtf and all GeneMark-ES genes. Thus, this set is generally more sensitive (more genes correctly predicted) and can be less specific (more false-positive predictions can be present).

• hintsfile.gff: The extrinsic evidence data extracted from RNAseq.bam and/or protein data.

AUGUSTUS output files may be present with the following name endings and formats:

• GTF-format is always produced.

• GFF3-format is produced if the flag --gff3 was specified to BRAKER.

• Coding sequences in FASTA-format are produced if the flag --skipGetAnnoFromFasta was not set.

• Protein sequence files in FASTA-format are produced if the flag --skipGetAnnoFromFasta was not set.

For details about gtf format, see http://www.sanger.ac.uk/Software/formats/GFF/. A GTF-format file contains one line per predicted exon. Example:

    HS04636 AUGUSTUS initial   966 1017 . + 0 transcript_id "g1.1"; gene_id "g1";
    HS04636 AUGUSTUS internal 1818 1934 . + 2 transcript_id "g1.1"; gene_id "g1";

The columns (fields) contain:

    seqname source feature start end score strand frame transcript ID and gene ID

If the --makehub option was used and MakeHub is available on your system, a hub directory beginning with the name hub_ will be created. Copy this directory to a publicly accessible web server. A file hub.txt resides in the directory. Provide the link to that file to the UCSC Genome Browser for visualizing results.

Example data

An incomplete example data set is contained in the directory BRAKER/example. In order to complete the data set, please download the RNA-Seq alignment file (134 MB) with wget:

cd BRAKER/example
wget http://topaz.gatech.edu/GeneMark/Braker/RNAseq.bam

In case you have trouble accessing that file, there's also a copy available from another server:

cd BRAKER/example
wget http://bioinf.uni-greifswald.de/augustus/datasets/RNAseq.bam

The example data set was not compiled in order to achieve optimal prediction accuracy, but in order to quickly test pipeline components. The small subset of the genome used in these test examples is not long enough for BRAKER training to work well.

Data description

Data corresponds to the last 1,000,000 nucleotides of Arabidopsis thaliana's chromosome Chr5, split into 8 artificial contigs.

RNA-Seq alignments were obtained by VARUS.

The protein sequences are a subset of OrthoDB v10 plants proteins.

List of files:

• genome.fa - genome file in fasta format
• RNAseq.bam - RNA-Seq alignment file in bam format (this file is not a part of this repository, it must be downloaded separately from http://topaz.gatech.edu/GeneMark/Braker/RNAseq.bam)
• RNAseq.hints - RNA-Seq hints (can be used instead of RNAseq.bam as RNA-Seq input to BRAKER)
• proteins.fa - protein sequences in fasta format

The below given commands assume that you configured all paths to tools by exporting bash variables or that you have the necessary tools in your $PATH.

The example data set also contains scripts tests/test*.sh that will execute below listed commands for testing BRAKER with the example data set. You find example results of AUGUSTUS and GeneMark-EX in the folder results/test*. Be aware that BRAKER contains several parts where random variables are used, i.e. results that you obtain when running the tests may not be exactly identical. To compare your test results with the reference ones, you can use the compare_intervals_exact.pl script as follows:

# Compare CDS features
compare_intervals_exact.pl --f1 augustus.hints.gtf --f2 ../../results/test${N}/augustus.hints.gtf --verbose
# Compare transcripts
compare_intervals_exact.pl --f1 augustus.hints.gtf --f2 ../../results/test${N}/augustus.hints.gtf --trans --verbose

Several tests use --gm_max_intergenic 10000 option to make the test runs faster. It is not recommended to use this option in real BRAKER runs, the speed increase achieved by adjusting this option is negligible on full-sized genomes.

We give runtime estimations derived from computing on Intel(R) Xeon(R) CPU E5530 @ 2.40GHz.

Testing BRAKER with RNA-Seq data

The following command will run the pipeline according to Figure 3:

braker.pl --genome genome.fa --bam RNAseq.bam --softmasking --cores N

This test is implemented in test1.sh, expected runtime is ~20 minutes.

Testing BRAKER with proteins of any evolutionary distance

The following command will run the pipeline according to Figure 4:

braker.pl --genome genome.fa --prot_seq proteins.fa --softmasking --cores N

This test is implemented in test2.sh, expected runtime is ~20 minutes.

Testing BRAKER with proteins of any evolutionary distance and RNA-Seq

Please consider using TSEBRA instead of running BRAKER with both RNA-Seq and protein data: https://github.com/Gaius-Augustus/TSEBRA

The following command will run a pipeline that first trains GeneMark-ETP with protein and RNA-Seq hints and subsequently trains AUGUSTUS on the basis of GeneMark-ETP predictions. AUGUSTUS predictions are also performed with hints from both sources, see Figure 5:

braker.pl --genome genome.fa --prot_seq proteins.fa --bam ../RNAseq.bam --etpmode --softmasking --cores N

This test is implemented in test3.sh, expected runtime is ~20 minutes.

You can add UTRs from RNA-Seq data (no AUGUSTUS training) to results of a BRAKER run in ETP-mode the following way:

braker.pl --genome=../genome.fa --addUTR=on --softmasking \
    --bam=../RNAseq.bam --workingdir=$wd \
    --AUGUSTUS_hints_preds=augustus.hints.gtf --cores=8 \
    --skipAllTraining --species=somespecies

This is implemented in test3_add_utrs.sh, expected runtime is ~1 minute.

Testing BRAKER with proteins of close homology

The following command will run the pipeline according to Figure 6:

braker.pl --genome genome.fa --prot_seq proteins.fa --prg gth \
    --trainFromGth --softmasking --cores N

This test is implemented in test4.sh, expected runtime is ~7 minutes. The fast runtime of this test is mostly caused by generating a low number of training genes. Note that this approach does not scale well with increasing genome size and the number of proteins in a protein database. The runtime on a full genome will be much slower than with the command used in test2.sh.

Testing BRAKER with proteins of close homology and RNA-Seq data (RNA-Seq supported training)

Please consider using TSBERA instead of running BRAKER with proteins and RNA-Seq data at the same time: https://github.com/Gaius-Augustus/TSEBRA

The following command will run the pipeline according to Figure 7:

braker.pl --genome genome.fa --prot_seq proteins.fa --prg gth \
    --bam RNAseq.bam --softmasking --cores N

This test is implemented in test5.sh, expected runtime is ~20 minutes.

Testing BRAKER with proteins of close homology and RNA-Seq data (RNA-Seq and protein supported training)

Please consider using TSBERA instead of running BRAKER with proteins and RNA-Seq data at the same time: https://github.com/Gaius-Augustus/TSEBRA

The following command will run the pipeline according to Figure 8:

braker.pl --genome genome.fa --prot_seq prot.fa --prg gth --bam RNAseq.bam \
    --gth2traingenes --softmasking --cores N

This test is implemented in test6.sh, expected runtime is ~20 minutes.

Testing BRAKER with pre-trained parameters

The training step of all pipelines can be skipped with the option --skipAllTraining. This means, only AUGUSTUS predictions will be performed, using pre-trained, already existing parameters. For example, you can predict genes with the command:

    braker.pl --genome=genome.fa --bam RNAseq.bam --species=arabidopsis \
        --skipAllTraining --softmasking --cores N

This test is implemented in test7.sh, expected runtime is ~1 minute.

Testing BRAKER with genome sequence

The following command will run the pipeline with no extrinsic evidence:

braker.pl --genome=genome.fa --esmode --softmasking --cores N

This test is implemented in test8.sh, expected runtime is ~20 minutes.

Testing BRAKER with RNA-Seq data and --UTR=on

The following command will run BRAKER with training UTR parameters from RNA-Seq coverage data:

braker.pl --genome genome.fa --bam RNAseq.bam --softmasking --UTR=on --cores N

This test is implemented in test9.sh, expected runtime is ~20 minutes.

Testing BRAKER with RNA-Seq data and --addUTR=on

The following command will add UTRs to augustus.hints.gtf from RNA-Seq coverage data:

braker.pl --genome genome.fa --bam RNAseq.bam --softmasking --addUTR=on --cores N

This test is implemented in test10.sh, expected runtime is ~20 minutes.

Starting BRAKER on the basis of previously existing BRAKER runs

There is currently no clean way to restart a failed BRAKER run (after solving some problem). However, it is possible to start a new BRAKER run based on results from a previous run -- given that the old run produced the required intermediate results. We will in the following refer to the old working directory with variable ${BRAKER_OLD}, and to the new BRAKER working directory with ${BRAKER_NEW}. The file what-to-cite.txt will always only refer to the software that was actually called by a particular run. You might have to combine the contents of ${BRAKER_NEW}/what-to-cite.txt with ${BRAKER_OLD}/what-to-cite.txt for preparing a publication. The following figure illustrates at which points BRAKER run may be intercepted.

Figure 10: Points for intercepting a BRAKER run and reusing intermediate results in a new BRAKER run.

Option 1: starting BRAKER with existing hints file(s) before training

If you have access to an existing BRAKER output that contains hintsfiles that were generated from extrinsic data, such as RNA-Seq or protein sequences, you can recycle these hints files in a new BRAKER run. Also, hints from a separate ProtHint run can be directly used in BRAKER.

The hints can be given to BRAKER with --hints ${BRAKER_OLD}/hintsfile.gff option. This is illustrated in the test files test1_restart1.sh, test2_restart1.sh, test3_restart1.sh, test5_restart1.sh, and test7_restart1.sh. The other modes (for which this test is missing) cannot be restarted in this way.

Option 2: starting BRAKER after GeneMark-EX had finished, before training AUGUSTUS

The GeneMark result can be given to BRAKER with --geneMarkGtf ${BRAKER_OLD}/GeneMark-EX/genemark.gtf option. This is illustrated in the test files test1_restart2.sh, test2_restart2.sh, test3_restart2.sh, test5_restart2.sh, and test8_restart2.sh. The other modes (for which this test is missing) cannot be restarted in this way.

Option 3: starting BRAKER after AUGUSTUS training

The trained species parameters for AGUSTUS can be passed with --skipAllTraining and --species $speciesName options. This is illustrated in test*_restart3.sh files.

Bug reporting

Before reporting bugs, please check that you are using the most recent versions of GeneMark-EX, AUGUSTUS and BRAKER. Also, check the list of Common problems, and the Issue list on GitHub before reporting bugs. We do monitor open issues on GitHub. Sometimes, we are unable to help you, immediately, but we try hard to solve your problems.

Reporting bugs on GitHub

If you found a bug, please open an issue at https://github.com/Gaius-Augustus/BRAKER/issues (or contact [email protected] or [email protected]).

Information worth mentioning in your bug report:

Check in braker/yourSpecies/braker.log at which step braker.pl crashed.

There are a number of other files that might be of interest, depending on where in the pipeline the problem occurred. Some of the following files will not be present if they did not contain any errors.

• braker/yourSpecies/errors/bam2hints.*.stderr - will give details on a bam2hints crash (step for converting bam file to intron gff file)

• braker/yourSpecies/hintsfile.gff - is this file empty? If yes, something went wrong during hints generation - does this file contain hints from source “b2h” and of type “intron”? If not: GeneMark-ET will not be able to execute properly. Conversely, GeneMark-EP+ will not be able to execute correctly if hints from the source "ProtHint" are missing.

• braker/yourSpecies/(align_gthalign_exoneratealign_spaln)/*err - errors reported by the alignment tools gth/exonerate/spaln

• braker/yourSpecies/errors/GeneMark-{ET,EP}.stderr - errors reported by GeneMark-ET/EP+

• braker/yourSpecies/errors/GeneMark-{ET,EP).stdout - may give clues about the point at which errors in GeneMark-ET/EP+ occured

• braker/yourSpecies/GeneMark-{ET,EP}/genemark.gtf - is this file empty? If yes, something went wrong during executing GeneMark-ET/EP+

• braker/yourSpecies/GeneMark-{ET,EP}/genemark.f.good.gtf - is this file empty? If yes, something went wrong during filtering GeneMark-ET/EP+ genes for training AUGUSTUS

• braker/yourSpecies/genbank.good.gb - try a “grep -c LOCUS genbank.good.gb” to determine the number of training genes for training AUGUSTUS, should not be low

• braker/yourSpecies/errors/firstetraining.stderr - contains errors from first iteration of training AUGUSTUS

• braker/yourSpecies/errors/secondetraining.stderr - contains errors from second iteration of training AUGUSTUS

• braker/yourSpecies/errors/optimize_augustus.stderr - contains errors optimize_augustus.pl (additional training set for AUGUSTUS)

• braker/yourSpecies/errors/augustus*.stderr - contain AUGUSTUS execution errors

• braker/yourSpecies/startAlign.stderr - if you provided a protein fasta file and --prg option and this file is not empty, something went wrong during protein alignment

• braker/yourSpecies/startAlign.stdout - may give clues on at which point protein alignment went wrong

Common problems

• BRAKER complains that the RNA-Seq file does not correspond to the provided genome file, but I am sure the files correspond to each other!

  Please check the headers of the genome FASTA file. If the headers are long and contain whitespaces, some RNA-Seq alignment tools will truncate sequence names in the BAM file. This leads to an error with BRAKER. Solution: shorten/simplify FASTA headers in the genome file before running the RNA-Seq alignment and BRAKER.

• There are duplicate Loci in the train.gb file (after using GenomeThreader)!

  This issue arises if outdated versions of AUGUSTUS and BRAKER are used. Solution: Please update AUGUSTUS and BRAKER from github (https://github.com/Gaius-Augustus/Augustus, https://github.com/Gaius-Augustus/BRAKER).

• GeneMark fails!

  (a) GeneMark requires a valid hidden key file in your home directory (~/.gm_key). The file expires after 200 days. Please check whether you have a valid key file before reporting an issue about this. Also, BRAKER may issue a WARNING that GeneMark is likely going to fail due to limited extrinsic evidence. If you see that warning, please don't open an issue but try a different approach towards annotating your genome. For example, you can add more evidence data, you can try the protein mapping pipeline approach, you can try running --esmode without extrinsic evidence, ...

  (b) GeneMark by default only uses contigs longer than 50k for training. If you have a highly fragmented assembly, this might lead to "no data" for training. You can override the default minimal length by setting the BRAKER argument --min_contig=10000.

  (c) see "[something] failed to execute" below.

• [something] failed to execute!

  When providing paths to software to BRAKER, please use absolute, non-abbreviated paths. For example, BRAKER might have problems with --SAMTOOLS_PATH=./samtools/ or --SAMTOOLS_PATH=~/samtools/. Please use SAMTOOLS_PATH=/full/absolute/path/to/samtools/, instead. This applies to all path specifications as command line options to braker.pl. Relative paths and absolute paths will not pose problems if you export a bash variable, instead, or if you append the location of tools to your $PATH variable.

• BRAKER cannot find the Augustus script XYZ...

  Update Augustus from github with git clone https://github.com/Gaius-Augustus/Augustus.git. Do not use Augustus from other sources. BRAKER is highly dependent on an up-to-date Augustus. Augustus releases happen rather rarely, updates to the Augustus scripts folder occur rather frequently.

• Does BRAKER depend on Python3?

  It does. The python scripts employed by BRAKER are not compatible with Python2.

• Why does BRAKER predict more genes than I expected?

  If transposable elements (or similar) have not been masked appropriately, AUGUSTUS tends to predict those elements as protein coding genes. This can lead to a huge number genes. You can check whether this is the case for your project by BLASTing (or DIAMONDing) the predicted protein sequences against themselves (all vs. all) and counting how many of the proteins have a high number of high quality matches. You can use the output of this analysis to divide your gene set into two groups: the protein coding genes that you want to find and the repetitive elements that were additionally predicted.

• I am running BRAKER in Anaconda and something fails...

  Update AUGUSTUS and BRAKER from github with git clone https://github.com/Gaius-Augustus/Augustus.git and git clone https://github.com/Gaius-Augustus/BRAKER.git. The Anaconda installation is great, but it relies on releases of AUGUSTUS and BRAKER - which are often lagging behind. Please use the current GitHub code, instead.

Citing BRAKER and software called by BRAKER

Since BRAKER is a pipeline that calls several Bioinformatics tools, publication of results obtained by BRAKER requires that not only BRAKER is cited, but also the tools that are called by BRAKER. BRAKER will output a file what-to-cite.txt in the BRAKER working directory, informing you about which exact sources apply to your run.

• Always cite:

  • Bruna, T., Hoff, K.J., Lomsadze, A., Stanke, M., & Borodovsky, M. (2021). BRAKER2: Automatic Eukaryotic Genome Annotation with GeneMark-EP+ and AUGUSTUS Supported by a Protein Database. NAR Genomics and Bioinformatics 3(1):lqaa108, doi: 10.1093/nargab/lqaa108.

  • Hoff, K.J., Lomsadze, A., Borodovsky, M. and Stanke, M. (2019). Whole-Genome Annotation with BRAKER. Methods Mol Biol. 1962:65-95, doi: 10.1007/978-1-4939-9173-0_5.

  • Hoff, K.J., Lange, S., Lomsadze, A., Borodovsky, M. and Stanke, M. (2016). BRAKER1: unsupervised RNA-Seq-based genome annotation with GeneMark-ET and AUGUSTUS. Bioinformatics, 32(5):767-769.

  • Stanke, M., Diekhans, M., Baertsch, R. and Haussler, D. (2008). Using native and syntenically mapped cDNA alignments to improve de novo gene finding. Bioinformatics, doi: 10.1093/bioinformatics/btn013.

  • Stanke. M., Schöffmann, O., Morgenstern, B. and Waack, S. (2006). Gene prediction in eukaryotes with a generalized hidden Markov model that uses hints from external sources. BMC Bioinformatics 7, 62.

• If any kind of AUGUSTUS training was performed by BRAKER, check carefully whether you configured BRAKER to use NCBI BLAST or DIAMOND. One of them was used to filter out redundant training gene structures.

  • If you used NCBI BLAST, please cite:

    • Altschul, A.F., Gish, W., Miller, W., Myers, E.W. and Lipman, D.J. (1990). A basic local alignment search tool. J Mol Biol 215:403--410.

    • Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J., Bealer, K., and Madden, T.L. (2009). Blast+: architecture and applications. BMC bioinformatics, 10(1):421.

  • If you used DIAMOND, please cite:

    • Buchfink, B., Xie, C., Huson, D.H. (2015). Fast and sensitive protein alignment using DIAMOND. Nature Methods 12:59-60.

• If BRAKER was executed with a genome file and no extrinsic evidence, cite, then GeneMark-ES was used, cite:

  • Lomsadze, A., Ter-Hovhannisyan, V., Chernoff, Y.O. and Borodovsky, M. (2005). Gene identification in novel eukaryotic genomes by self-training algorithm. Nucleic Acids Research, 33(20):6494--6506.

  • Ter-Hovhannisyan, V., Lomsadze, A., Chernoff, Y.O. and Borodovsky, M. (2008). Gene prediction in novel fungal genomes using an ab initio algorithm with unsupervised training. Genome research, pages gr--081612, 2008.

• If BRAKER was run with proteins of any phylogenetic distance (--epmode or --etpmode), please cite all tools that are used by the ProtHint pipeline to generate hints:

  • Bruna, T., Lomsadze, A., & Borodovsky, M. (2020). GeneMark-EP+: eukaryotic gene prediction with self-training in the space of genes and proteins. NAR Genomics and Bioinformatics, 2(2), lqaa026.

  • Buchfink, B., Xie, C., Huson, D.H. (2015). Fast and sensitive protein alignment using DIAMOND. Nature Methods 12:59-60.

  • Lomsadze, A., Ter-Hovhannisyan, V., Chernoff, Y.O. and Borodovsky, M. (2005). Gene identification in novel eukaryotic genomes by self-training algorithm. Nucleic Acids Research, 33(20):6494--6506.

  • Iwata, H., and Gotoh, O. (2012). Benchmarking spliced alignment programs including Spaln2, an extended version of Spaln that incorporates additional species-specific features. Nucleic acids research, 40(20), e161-e161.

  • Gotoh, O., Morita, M., Nelson, D. R. (2014). Assessment and refinement of eukaryotic gene structure prediction with gene-structure-aware multiple protein sequence alignment. BMC bioinformatics, 15(1), 189.

• If BRAKER was executed with RNA-Seq alignments in bam-format, then SAMtools was used, cite:

  • Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., Durbin, R.; 1000 Genome Project Data Processing Subgroup (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25(16):2078-9.

  • Barnett, D.W., Garrison, E.K., Quinlan, A.R., Strömberg, M.P. and Marth G.T. (2011). BamTools: a C++ API and toolkit for analyzing and managing BAM files. Bioinformatics, 27(12):1691-2

• If BRAKER used RNA-Seq alignments for generating a training gene set, cite GeneMark-ET:

  • Lomsadze, A., Paul D.B., and Mark B. (2014) Integration of Mapped Rna-Seq Reads into Automatic Training of Eukaryotic Gene Finding Algorithm. Nucleic Acids Research 42(15): e119--e119

• If BRAKER was executed with proteins of closely related species, cite GenomeThreader:

  • Gremme, G. (2013). Computational Gene Structure Prediction. PhD thesis, Universität Hamburg.

• If BRAKER called MakeHub for creating a track data hub for visualization of BRAKER results with the UCSC Genome Browser, cite:

  • Hoff, K.J. (2019) MakeHub: Fully automated generation of UCSC Genome Browser Assembly Hubs. Genomics, Proteomics and Bioinformatics, in press 2020, preprint on bioarXive, doi: https://doi.org/10.1101/550145.

• If BRAKER called GUSHR for generating UTRs, cite:

  • Keilwagen, J., Hartung, F., Grau, J. (2019) GeMoMa: Homology-based gene prediction utilizing intron position conservation and RNA-seq data. Methods Mol Biol. 1962:161-177, doi: 10.1007/978-1-4939-9173-0_9.

  • Keilwagen, J., Wenk, M., Erickson, J.L., Schattat, M.H., Grau, J., Hartung F. (2016) Using intron position conservation for homology-based gene prediction. Nucleic Acids Research, 44(9):e89.

  • Keilwagen, J., Hartung, F., Paulini, M., Twardziok, S.O., Grau, J. (2018) Combining RNA-seq data and homology-based gene prediction for plants, animals and fungi. BMC Bioinformatics, 19(1):189.

License

All source code, i.e. scripts/*.pl or scripts/*.py are under the Artistic License (see http://www.opensource.org/licenses/artistic-license.php).

Footnotes

[F1] EX = ES/ET/EP/ETP, all available for download under the name GeneMark-ES/ET/EP ↩

[F2] Please use the latest version from the master branch of AUGUSTUS distributed by the original developers, it is available from github at https://github.com/Gaius-Augustus/Augustus. Problems have been reported from users that tried to run BRAKER with AUGUSTUS releases maintained by third parties, i.e. Bioconda. ↩

[F3] Not tested in this release, we recommend using GenomeThreader, instead ↩

[F4] install with sudo apt-get install cpanminus ↩

[F5] GeneMark-EX is not a mandatory tool if AUGUSTUS is to be trained from GenomeThreader aligments with the option --trainFromGth. ↩

[F6] The binary may e.g. reside in bamtools/build/src/toolkit ↩

References

[R0] Tomas Bruna, Katharina J. Hoff, Alexandre Lomsadze, Mario Stanke and Mark Borodvsky. 2021. “BRAKER2: automatic eukaryotic genome annotation with GeneMark-EP+ and AUGUSTUS supported by a protein database." NAR Genomics and Bioinformatics 3(1):lqaa108.↩

[R1] Hoff, Katharina J, Simone Lange, Alexandre Lomsadze, Mark Borodovsky, and Mario Stanke. 2015. “BRAKER1: Unsupervised Rna-Seq-Based Genome Annotation with Genemark-et and Augustus.” Bioinformatics 32 (5). Oxford University Press: 767--69.↩

[R2] Lomsadze, Alexandre, Paul D Burns, and Mark Borodovsky. 2014. “Integration of Mapped Rna-Seq Reads into Automatic Training of Eukaryotic Gene Finding Algorithm.” Nucleic Acids Research 42 (15). Oxford University Press: e119--e119.↩

[R3] Stanke, Mario, Mark Diekhans, Robert Baertsch, and David Haussler. 2008. “Using Native and Syntenically Mapped cDNA Alignments to Improve de Novo Gene Finding.” Bioinformatics 24 (5). Oxford University Press: 637--44.↩

[R4] Stanke, Mario, Oliver Schöffmann, Burkhard Morgenstern, and Stephan Waack. 2006. “Gene Prediction in Eukaryotes with a Generalized Hidden Markov Model That Uses Hints from External Sources.” BMC Bioinformatics 7 (1). BioMed Central: 62.↩

[R5] Barnett, Derek W, Erik K Garrison, Aaron R Quinlan, Michael P Strömberg, and Gabor T Marth. 2011. “BamTools: A C++ Api and Toolkit for Analyzing and Managing Bam Files.” Bioinformatics 27 (12). Oxford University Press: 1691--2.↩

[R6] Li, Heng, Handsaker, Bob, Alec Wysoker, Tim Fennell, Jue Ruan, Nils Homer, Gabor Marth, Goncalo Abecasis, and Richard Durbin. 2009. “The Sequence Alignment/Map Format and Samtools.” Bioinformatics 25 (16). Oxford University Press: 2078--9.↩

[R7] Gremme, G. 2013. “Computational Gene Structure Prediction.” PhD thesis, Universität Hamburg.↩

[R8] Gotoh, Osamu. 2008a. “A Space-Efficient and Accurate Method for Mapping and Aligning cDNA Sequences onto Genomic Sequence.” Nucleic Acids Research 36 (8). Oxford University Press: 2630--8.↩

[R9] Iwata, Hiroaki, and Osamu Gotoh. 2012. “Benchmarking Spliced Alignment Programs Including Spaln2, an Extended Version of Spaln That Incorporates Additional Species-Specific Features.” Nucleic Acids Research 40 (20). Oxford University Press: e161--e161.↩

[R10] Osamu Gotoh. 2008b. “Direct Mapping and Alignment of Protein Sequences onto Genomic Sequence.” Bioinformatics 24 (21). Oxford University Press: 2438--44.↩

[R11] Slater, Guy St C, and Ewan Birney. 2005. “Automated Generation of Heuristics for Biological Sequence Comparison.” BMC Bioinformatics 6(1). BioMed Central: 31.↩

[R12] Altschul, S.F., W. Gish, W. Miller, E.W. Myers, and D.J. Lipman. 1990. “Basic Local Alignment Search Tool.” Journal of Molecular Biology 215:403--10.↩

[R13] Camacho, Christiam, et al. 2009. “BLAST+: architecture and applications.“ BMC Bioinformatics 1(1): 421.↩

[R14] Lomsadze, A., V. Ter-Hovhannisyan, Y.O. Chernoff, and M. Borodovsky. 2005. “Gene identification in novel eukaryotic genomes by self-training algorithm.” Nucleic Acids Research 33 (20): 6494--6506. doi:10.1093/nar/gki937.↩

[R15] Ter-Hovhannisyan, Vardges, Alexandre Lomsadze, Yury O Chernoff, and Mark Borodovsky. 2008. “Gene Prediction in Novel Fungal Genomes Using an Ab Initio Algorithm with Unsupervised Training.” Genome Research. Cold Spring Harbor Lab, gr--081612.↩

[R16] Hoff, K.J. 2019. MakeHub: Fully automated generation of UCSC Genome Browser Assembly Hubs. Genomics, Proteomics and Bioinformatics, in press, preprint on bioarXive, doi: https://doi.org/10.1101/550145.↩

[R17] Bruna, T., Lomsadze, A., & Borodovsky, M. 2020. GeneMark-EP+: eukaryotic gene prediction with self-training in the space of genes and proteins. NAR Genomics and Bioinformatics, 2(2), lqaa026. doi: https://doi.org/10.1093/nargab/lqaa026.↩

[R18] Kriventseva, E. V., Kuznetsov, D., Tegenfeldt, F., Manni, M., Dias, R., Simão, F. A., and Zdobnov, E. M. 2019. OrthoDB v10: sampling the diversity of animal, plant, fungal, protist, bacterial and viral genomes for evolutionary and functional annotations of orthologs. Nucleic Acids Research, 47(D1), D807-D811.↩

[R19] Keilwagen, J., Hartung, F., Grau, J. (2019) GeMoMa: Homology-based gene prediction utilizing intron position conservation and RNA-seq data. Methods Mol Biol. 1962:161-177, doi: 10.1007/978-1-4939-9173-0_9.↩

[R20] Keilwagen, J., Wenk, M., Erickson, J.L., Schattat, M.H., Grau, J., Hartung F. (2016) Using intron position conservation for homology-based gene prediction. Nucleic Acids Research, 44(9):e89.↩

[R21] Keilwagen, J., Hartung, F., Paulini, M., Twardziok, S.O., Grau, J. (2018) Combining RNA-seq data and homology-based gene prediction for plants, animals and fungi. BMC Bioinformatics, 19(1):189.↩