Note

I created this repo of biogrinder mainly for reading the README.md in a better formatted and easier way.

The source was originally downloaded from

https://sourceforge.net/projects/biogrinder/files/biogrinder/Grinder-0.5.3/Grinder-0.5.3.tar.gz.

A newer version is avialable now at

https://sourceforge.net/projects/biogrinder/files/biogrinder/.

This repo is not for properly tracking development history of biogrinder, its code history is at

https://sourceforge.net/p/biogrinder/code/ci/master/tree/.

Please see the <a href="#bugs">end</a> of this document for how to clone it.

Grinder

A versatile omics shotgun and amplicon sequencing read simulator

Description

Grinder is a versatile program to create random shotgun and amplicon sequence libraries based on DNA, RNA or proteic reference sequences provided in a FASTA file.

Grinder can produce genomic, metagenomic, transcriptomic, metatranscriptomic, proteomic, metaproteomic shotgun and amplicon datasets from current sequencing technologies such as Sanger, 454, Illumina. These simulated datasets can be used to test the accuracy of bioinformatic tools under specific hypothesis, e.g. with or without sequencing errors, or with low or high community diversity. Grinder may also be used to help decide between alternative sequencing methods for a sequence-based project, e.g. should the library be paired-end or not, how many reads should be sequenced.

Grinder features include:

• shotgun or amplicon read libraries
• omics support to generate genomic, transcriptomic, proteomic, metagenomic, metatranscriptomic or metaproteomic datasets
• arbitrary read length distribution and number of reads
• simulation of PCR and sequencing errors (chimeras, point mutations, homopolymers)
• support for paired-end (mate pair) datasets
• specific rank-abundance settings or manually given abundance for each genome, gene or protein
• creation of datasets with a given richness (alpha diversity)
• independent datasets can share a variable number of genomes (beta diversity)
• modeling of the bias created by varying genome lengths or gene copy number
• profile mechanism to store preferred options
• available to biologists or power users through multiple interfaces: GUI, CLI and API

Briefly, given a FASTA file containing reference sequence (genomes, genes, transcripts or proteins), Grinder performs the following steps:

1. Read the reference sequences, and for amplicon datasets, extracts full-length reference PCR amplicons using the provided degenerate PCR primers.

2. Determine the community structure based on the provided alpha diversity (number of reference sequences in the library), beta diversity (number of reference sequences in common between several independent libraries) and specified rank- abundance model.

3. Take shotgun reads from the reference sequences or amplicon reads from the full- length reference PCR amplicons. The reads may be paired-end reads when an insert size distribution is specified. The length of the reads depends on the provided read length distribution and their abundance depends on the relative abundance in the community structure. Genome length may also biases the number of reads to take for shotgun datasets at this step. Similarly, for amplicon datasets, the number of copies of the target gene in the reference genomes may bias the number of reads to take.

4. Alter reads by inserting sequencing errors (indels, substitutions and homopolymer errors) following a position-specific model to simulate reads created by current sequencing technologies (Sanger, 454, Illumina). Write the reads and their quality scores in FASTA, QUAL and FASTQ files.

Citation

If you use Grinder in your research, please cite:

Angly FE, Willner D, Rohwer F, Hugenholtz P, Tyson GW (2012), Grinder: a versatile amplicon and shotgun sequence simulator, Nucleic Acids Reseach

Available from http://dx.doi.org/10.1093/nar/gks251.

Author

Florent Angly <florent.angly@gmail.com>

Installation

You need to install these dependencies first:

• Perl (>= 5.6)
• make
• The following CPAN Perl modules are dependencies that will be installed automatically for you:
  • Bioperl modules (>=1.6.901). Note that some unreleased Bioperl modules have been included in Grinder.
  • Getopt::Euclid (>= 0.3.4)
  • List::Util. First released with Perl v5.7.3
  • Math::Random::MT (>= 1.13)
  • version (>= 0.77). First released with Perl v5.9.0

To install Grinder globally on your system, run the following commands in a terminal or command prompt:

On Linux, Unix, MacOS:

perl Makefile.PL
make
# The following step needs administrator privileges
make install

On Windows:

perl Makefile.PL
nmake
# The following step needs administrator privileges
nmake install

If you don't have administrator privileges, then Grinder needs to be installed in your home directory.

Follow the instructions to install local::lib here. Then, every Perl module that you install manually or through the CPAN command-line application will be installed in your home directory. At last, install Grinder by following the instructions detailed as above.

Running Grinder

You can run Grinder using the

• command-line interface (CLI). see grinder --help
• programming interface (API). see perldoc Grinder
• or graphical user interface (GUI) in Galaxy. See the Galaxy documentation here

The utils folder

It's included in the Grinder package and contains the following utilities:

• average genome size: This calculates the average genome size (in bp) of a simulated random library produced by Grinder.

• change_paired_read_orientation: This reverses the orientation of each second mate-pair read (ID ending in /2) in a FASTA file.

Reference Sequence Database

A variety of FASTA databases can be used as input for Grinder. For example, GreenGenes database, RDP or Silva. They all provide many 16S rRNA sequences, but Silva includes eukaryotic sequences, as well.

While 16S rRNA is a popular gene, datasets containing any type of gene could can used in the same fashion to generate simulated amplicon datasets, provided that appropriate primers are used.

The over 2,400 curated microbial genome sequences in the NCBI RefSeq microbial collection would also be suitable for producing 16S rRNA simulated datasets (using the adequate primers). However, the lower diversity of this database compared to the previous two makes it more appropriate for producing artificial microbial metagenomes. Individual genomes from this database are also very suitable for the simulation of single or double-barreled shotgun libraries. Similarly, the RefSeq viral collection contains over 3,100 curated viral sequences which can be used to produce artificial viral metagenomes.

Quite a few eukaryotic organisms have been sequenced and their genome or genes can be the basis for simulating genomic, transcriptomic (RNA-seq) or proteomic datasets. For example, you can use

• the human genome available at ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/RefSeqGene/
• the human transcripts at ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/mRNA_Prot/human.rna.fna.gz
• or the human proteome at ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/mRNA_Prot/human.protein.faa.gz

CLI Examples

A shotgun DNA library with a coverage of 0.1X

grinder -reference_file genomes.fna -coverage_fold 0.1

Same thing but save the result files in a specific folder and with a specific name

grinder -reference_file genomes.fna -coverage_fold 0.1 -base_name my_name -output_dir my_dir

A DNA shotgun library with 1000 reads

grinder -reference_file genomes.fna -total_reads 1000

A DNA shotgun library where species are distributed according to a power law

grinder -reference_file genomes.fna -abundance_model powerlaw 0.1

A DNA shotgun library with 123 genomes taken random from the given genomes

grinder -reference_file genomes.fna -diversity 123

Two DNA shotgun libraries that have 50% of the species in common

grinder -reference_file genomes.fna -num_libraries 2 -shared_perc 50

Two DNA shotgun library with no species in common and distributed according to a exponential rank-abundance model. Note that because the parameter value for the exponential model is omitted, each library uses a different randomly chosen value:

grinder -reference_file genomes.fna -num_libraries 2 -abundance_model exponential

A DNA shotgun library where species relative abundances are manually specified

grinder -reference_file genomes.fna -abundance_file my_abundances.txt

A DNA shotgun library with Sanger reads

grinder -reference_file genomes.fna -read_dist 800 -mutation_dist linear 1 2 -mutation_ratio 80 20

A DNA shotgun library with first-generation 454 reads

grinder -reference_file genomes.fna -read_dist 100 normal 10 -homopolymer_dist balzer

A paired-end DNA shotgun library, where the insert size is normally distributed around 2.5 kbp and has 0.2 kbp standard deviation

grinder -reference_file genomes.fna -insert_dist 2500 normal 200

A transcriptomic dataset

grinder -reference_file transcripts.fna

A unidirectional transcriptomic dataset

grinder -reference_file transcripts.fna -unidirectional 1

Note: the use of -unidirectional 1 to prevent reads to be taken from the reverse-complement of the reference sequences.

A proteomic dataset

grinder -reference_file proteins.faa -unidirectional