Tomas Bruna, Alexandre Lomsadze, Mark Borodovsky

Georgia Institute of Technology, Atlanta, Georgia, USA

Reference: GeneMark-EP+: eukaryotic gene prediction with self-training in the space of genes and proteins

ProtHint is a pipeline for predicting and scoring hints (in the form of introns, start and stop codons) in the genome of interest by mapping and spliced aligning predicted genes to a database of reference protein sequences.

• Installation
  • Perl dependencies
  • Python dependencies
  • GeneMark-ES
  • DIAMOND
  • Spaln
    • Spaln boundary scorer
• Usage
  • Input
  • Protein Database Preparation
  • Running ProtHint
  • Output
• About

To install, copy the content of this distribution to desired location. To verify the installation, run ProtHint with the sample input located in the example folder.

Running ProtHint requires a Linux system with Bash. The following dependencies need to be satisfied.

Perl 5.10 or higher is required.

The following non-Core Perl modules are required:

• MCE::Mutex
• threads
• YAML
• Math::Utils

Core module Thread::Queue needs to be updated to a version 3.11 or higher.

These modules are available at CPAN and can be installed/updated with

cpan MCE::Mutex threads YAML Thread::Queue Math::Utils

Python 3.3 or higher is required. No libraries outside of the Python Standard Library are required.

There are two ways of using GeneMark-ES in ProtHint:

1. Run ProtHint with --geneMarkGtf genemark.gtf option which specifies the path to a file with GeneMark-ES predictions. If this option is used, GeneMark-ES does not need to be installed as a part of ProtHint.

2. Install GeneMark-ES, ProtHint will run it automatically as a part of the pipeline.

   Download and extract the contents of the GeneMark-ES suite (versions 4.30 and up) into the ProtHint/dependencies/GeneMarkES folder. GeneMark-ES suite is available at https://exon.gatech.edu/GeneMark/license_download.cgi

   To verify that GeneMark-ES is installed correctly, run the following command: ProtHint/dependencies/GeneMarkES/check_install.bash.

DIAMOND local sequence aligner (available at https://github.com/bbuchfink/diamond) is included in this distribution package.

In case the included version is not working, install DIAMOND from https://github.com/bbuchfink/diamond and replace the diamond binary in ProtHint/dependencies folder.

Spaln, space-efficient spliced alignment program (available at https://github.com/ogotoh/spaln), is included in this distribution package.

In case the included version is not working, install Spaln from https://github.com/ogotoh/spaln and replace the spaln binary in ProtHint/dependencies folder.

Binary for parsing and scoring hints from Spaln's alignment output is included in this distribution package.

In case the included binary is not working, compile it from source at https://github.com/gatech-genemark/spaln-boundary-scorer and replace the spaln_boundary_scorer binary in ProtHint/dependencies folder.

ProtHint inputs consist of:

• Genomic sequence from the target species in multi-FASTA format
• Reference protein sequences in multi-FASTA format

The tool is applicable to complete as well as draft genome assemblies. Every sequence in each multi-FASTA input needs to have a unique ID (first word of a FASTA header is used for ID). Examples of valid FASTA headers:

>contig10
ID: contig10
> seq3  genome Z
ID: seq3
>IV contig 25
ID: IV

We recommend to use a relevant portion of OrthoDB protein database as the source of reference protein sequences. The pre-partitioned OrthoDB clades (prepared with https://github.com/tomasbruna/orthodb-clades) are available at https://bioinf.uni-greifswald.de/bioinf/partitioned_odb11/.

To run ProtHint, use the following command:

prothint.py genome.fasta proteins.fasta

See the example folder for a sample input and output.

To display a list of all available options, use:

prothint.py --help

Frequently used options are:

--workdir WORKDIR   Folder for results and temporary files. If not
                    specified, current directory is used
--geneMarkGtf GENEMARKGTF
                    File with GeneMark-ES predictions in gtf format. If
                    this file is provided, GeneMark-ES run is skipped.
--diamondPairs DIAMONDPAIRS
                    File with "seed gene-protein" hits generated by
                    DIAMOND. If this file is provided, DIAMOND search for
                    protein hits is skipped.

ProtHint generates two main outputs:

• prothint.gff Gff file with all reported hints (introns, starts and stops)
• evidence.gff High confidence subset of prothint.gff which is, for instance, suitable for the GeneMark-EP Plus mode. This set is generated using default thresholds in ProtHint/bin/print_high_confidence.py script. If you wish to use different filtering criteria, re-run print_high_confidence.py script with custom thresholds.

An output which is ready to be used in BRAKER and AUGUSTUS is also generated:

• prothint_augustus.gff

ProtHint is developed by Tomas Bruna and Alexandre Lomsadze at Dr. Mark Borodovsky's Bioinformatics Lab, Georgia Institute of Technology, Atlanta, USA.