Gen
Gen is a version control system for genetic sequences. It efficiently stores genome-length sequences and sequence variations, with native support for polyploid genomes and pooled genotypes. Each project is organized into a repository, where collections of sequences and associated data are stored and tracked over time. Within a repository, branches can be created to explore different modifications or variations without affecting the main project. These branches can later be merged to integrate results from different experiments or collaborators.
The gen client can import standard sequence file formats from sources like NCBI and genetic design tools. The sequence model takes the form of a graph structure as shown in the figure below. Each molecule is made up out of a network …
Gen
Gen is a version control system for genetic sequences. It efficiently stores genome-length sequences and sequence variations, with native support for polyploid genomes and pooled genotypes. Each project is organized into a repository, where collections of sequences and associated data are stored and tracked over time. Within a repository, branches can be created to explore different modifications or variations without affecting the main project. These branches can later be merged to integrate results from different experiments or collaborators.
The gen client can import standard sequence file formats from sources like NCBI and genetic design tools. The sequence model takes the form of a graph structure as shown in the figure below. Each molecule is made up out of a network of nodes that represent sequence fragments, and edges that define how sequence fragments are connected. Multiple molecules are organized into collections that could represent the different chromosomes in a reference genome, proteins in a proteome, or pieces of DNA in a cloning reaction mixture. Molecules generally start out as a node that holds the reference sequence, and new edges and nodes are added for every sequence variant that is designed or observed. To reconstitute a linear sequence, the client walks from node to node along a defined path. This data model allows the representation of a wide range of biological complexity – from a single molecule to complex cultivars and cell lines, including any naturally occurring variation in addition to intended engineering.
Figure 1: Block graph model representation of a sequence variant where two nucleotides AT are replaced by TG; the modified sequence (shown in bold) is stored as a path over a list of edges that address specific coordinates.
The block graph model is designed to be additive: new sequence variants only add to the graph, existing nodes are stable and do not have to be split to accomodate the new topology. This is different from the more commonly used segment graph model shown in Figure 2, and gen automatically converts between formats as needed.
Figure 2: Segment graph model corresponding to the sequence variant in Figure 1. The original sequence was split into 3 parts; the modified sequence path is defined by a list of nodes that refer to these segments. Nodes are labeled by their parent ID and starting coordinate, new edges internal to the reference are shown as dashed arrows.
Individual chromosomes or contigs are stored as block groups that refer to a specific grouping of edges from the main graph. Block groups are adressed on three facets: name, sample, and collection. The name is an identifier like "chromosome II", and the sample could refer to a real individual or a virtual outcome of an experiment. The meaning of a collection is flexible, and for example could refer to the type of data (genomic, protein, ...) or experiment.
Prebuilt binaries
Prebuilt binaries exist for various platforms and architectures on the Release page. The latest version can be found here.
Installing from Source
Make sure you have a Rust compiler installed on your system. You can install the Rust toolset using the rustup installer.
Clone the source with git:
git clone https://github.com/ginkgobioworks/gen.git
cd rust
Compile the gen package and its dependencies:
cargo build --release
You can find the gen executable in ./target/release/ or execute it via cargo:
cargo run -- <arguments>
To cross-compile gen to run on a different architecture, you need to first add a target to the Rust toolchain and install a linker. For macOS to Linux this can be done as follows:
rustup target add x86_64-unknown-linux-gnu brew install SergioBenitez/osxct/x86_64-unknown-linux-gnu cargo build –release –target=x86_64-unknown-linux-gnu
The executable will be placed in ./target/x86_64-unknown-linux-gnu/release/
Usage
Gen is a command line tool with multiple subcommands that each have their own flags and arguments. The currently available commands are listed below and described in more detail here or built-in help pages (gen [COMMAND] --help).
Usage: gen [OPTIONS] [COMMAND]
Commands:
import Import a new sequence collection
update Update a sequence collection with new data
init Initialize a gen repository
branch Manage and create branches
checkout Migrate a database to a given operation
reset
operations View operations carried out against a database
apply
export Export a sequence collection
defaults Set a default database and collection to use
help Print this message or the help of the given subcommand(s)
Options:
-d, --db <DB> The path to the database you wish to utilize
-h, --help Print help
-V, --version Print version
Typical workflows are illustrated in these example workflows: