GINGER: Gradual INtegrated GEne Reconstruction

Description

GINGER is a tool that is implemented an integrated method for gene structure prediction in higher eukaryotes. RNA-Seq-based methods, ab initio-based methods, homology-based methods are performed, and then gene structures are reconstructed via dynamic programming with appropriately weighted and scored exon/intron/intergenic regions. Different prediction processes and filtering criteria are applied to multi-exon and single-exon genes. This pipeline is implemented using Nextflow.

Web site

https://github.com/i10labtitech/GINGER https://anaconda.org/i10labtitech/ginger

Requirements

See Requirements section in INSTALL.

Installation

See Installation section in INSTALL.

Synopsis

Settings

Set nextflow.config properly. See Settings section in INSTALL

Inputs

A file containing genome sequences (multi FASTA format)
A file containing masked genome sequences by RpeatMasker (multi FASTA format)
A file containing repeat information by RpeatMasker (***.out)
A file containing RNA-Seq 1 (FASTQ format)
A file containing RNA-Seq 2 (FASTQ format)
Files containing closely related species

Commands A (using the Docker image)

The command to kick the docker image with the prepared Perl script is as follows. (Docker image: i10labtitech/tools:GINGER_v1.0.1)

generateSampleData_cel [Directory name for sample input data (e.g. sample_data_cel)]
# Copy and edit nextflow.config
runGINGER.pl nextflow.config.user

Edit the [The path to the directory that contains sample_data_cel/] part of the variables named INPUT_GENOME, INPUT_GENOME, INPUT_MASKEDGENOME, INPUT_REPOUT. INPUT_RNASEQR1, INPUT_RNASEQR2, PROTEIN
Edit the [The path to the output directory (e.g. /XXX/XXX/sample_data_cel/] part of the variable named PDIR.
Edit the variable (path to scrach directory) named SCRATCH.
Edit the variable (number of threads used) named N_THREAD.
Edit the variable (maximum memory size used) named MAX_MEMORY.
Edit the variables named RNASEQ_OTHER#, HOMOLOGY_OTHER#, ABINITIO_OTHER#, (paths) RNASEQ_OTHER#_WEIGHT, HOMOLOGY_OTHER#_WEIGHT, ABINITIO_OTHER#_WEIGHT, (weights) when you want to use your annotation data to predict gene structures in GINGER.
Note that # is the number of annotation data.

cd /XXX/XXX/output_cel/
runEvaluatePred.pl /XXX/XXX/sample_data_cel/GCF_000002985.6_WBcel235_genomic.gff ./ginger_phase2.gff RNA CDS RNA CDS

Commands B (using installed GINGER)

The command to run GINGER after installing it is as follows. Put nextflow.config into your working directory. If you want to run the preparation phase all at once, type like following. phase2.sh requires an argument specifying minimum CDS length. 50 is an example.

If you have installed the conda package using "conda" or "mamba" command, you can obatain the automatically configured nextflow.config file for tool paths in the current directory using the "gingerInitCfg" command.

mamba install -c i10labtitech ginger
gingerInitCfg
# Then, a nextflow.config file will be generated in the current directory.
# Edit the paths of input files, the output directory, the scratch directory path,
# the number of threads used, and the maximum memory capacity, among other settings,
# in the nextflow.config file.

nextflow -C nextflow.config run /path/to/pipeline/ginger.nf
/path/to/pipeline/phase0.sh nextflow.config
/path/to/pipeline/phase1.sh nextflow.config > phase1.log
/path/to/pipeline/phase2.sh 50
/path/to/pipeline/summary.sh nextflow.config

[Note] /path/to/pipeline/ a path to a directory <"pipeline/" in GINGER's source tree>
[Note] An automatically calculated threshold for the score is written like score threshold = 1.25 in phase1.log.

If you want to run each methods separetely in the preparation phase, type like following. abinitio.nf must be executed after mapping.nf. The order in which denovo.nf and homology.nf are executed do not depend on them.

nextflow -C nextflow.config run /path/to/pipeline/mapping.nf
nextflow -C nextflow.config run /path/to/pipeline/denovo.nf
nextflow -C nextflow.config run /path/to/pipeline/abinitio.nf
nextflow -C nextflow.config run /path/to/pipeline/homology.nf
/path/to/pipeline/phase0.sh nextflow.config
/path/to/pipeline/phase1.sh nextflow.config > phase1.log
/path/to/pipeline/phase2.sh 50
/path/to/pipeline/summary.sh nextflow.config

To set a threshold for reconstructed gene structure's scores in the merge phase, type like following (use phase1_manual.sh instead of phase1.sh). The first argument is the threshold. 1.0 is an example.

/path/to/pipeline/phase1_manual.sh nextflow.config 1.0

Final output

The final outputs:

ginger_phase2.gff : gene structures by GINGER (GFF3) [Note] See http://gmod.org/wiki/GFF3 for details.
ginger.pep : protein sequences of the gene structurs (FASTA)
ginger.cds : CDS of the gene structures (FASTA)
ginger_stats.tsv : statistical information of gene structures

colindaven / GINGER