RaMA

Get Started

RaMA can be installed using two methods: via conda or by compiling the source code.

• If you want to quickly get started with RaMA, we highly recommend using the conda installation method.
• if you prefer to customize RaMA, such as enabling M64 mode to handle ultra-large-scale data (not generally recommended) or using different instruction sets to accelerate alignment, we suggest installing from the source code.

Install via conda

We highly recommand you install RaMA via conda. Simply run the following command:

conda create -n RaMA
conda activate RaMA
conda install -c malab rama
RaMA -h

Install via source code

Requirements

Before installing RaMA, ensure your Linux system meets the following requirements:

• GCC version 9 or higher. (Required)
• CMake version 3.1 or higher. (Required)

Clone the Repository

Start by cloning the RaMA repository along with its submodules. Ensure that Alignment/WFA2-lib is downloaded properly:

git clone --recursive https://github.com/metaphysicser/RaMA.git

Compilation

Navigate to the RaMA directory, create a build directory, and compile the program:

cd RaMA 
mkdir build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j$(nproc)
./RaMA -h

Compilation Flags Customize your build with additional flags if necessary:

# Enable 64-bit mode
cmake .. -DCMAKE_BUILD_TYPE=Release -DUSE_M64=ON

# Use extra compiler flags, e.g., for AVX2 support
cmake .. -DCMAKE_BUILD_TYPE=Release -DEXTRA_FLAGS="-mavx2"

How to use RaMA

General Usage

Execute RaMA by specifying the input FASTA file and the output directory:

RaMA -r /path/to/ref.fasta -q /path/to/query.fasta -o /path/to/output_dir

Detailed Usage

Control RaMA's behavior with the following arguments:

Arguments:
    -h, --help               Show this help message and exit 
    -r, --reference          Reference FASTA file path containing the reference sequences for alignment.
    -q, --query              Query FASTA file path containing the query sequences for alignment.
    -o, --output             Output directory path for saving alignment results and additional files.
    -t, --threads            Number of threads for the alignment process. Defaults to the number of available cores if unspecified.
    -s, --save               Saves anchor binary files to the output directory for future use, including SA, LCP, and Linear Sparse Table.
    -l, --load               Loads existing anchor binary files from the output directory to skip SA, LCP, and Linear Sparse Table construction.
    -c, --max_match_count    Maximum number of rare matches to use for anchor finding. Altering this value is generally not recommended.
    -m, --match              Match score for sequence alignment. Lower values favor matching characters. Default is 0.
    -x, --mismatch           Mismatch penalty. Higher values penalize mismatches more. Default is 3.
    -g, --gap_open1          Penalty for initiating a short gap. Key for handling different gap lengths. Default is 4.
    -e, --gap_extension1     Penalty for extending a short gap. Less severe than gap opening penalty. Default is 2.
    -G, --gap_open2          Penalty for initiating a long gap. Aims to manage long gaps strategically. Default is 12.
    -E, --gap_extension2     Penalty for extending a long gap. Provides a lenient approach to long gap management. Default is 1.

Note: The -r -q (input) and -o (output) parameters are required. For most users, the default settings are recommended, but feel free to adjust them as needed.

License

Apache 2.0 © [MALABZ_UESTC Pinglu Zhang]

Contact

RaMA is actively developed and maintained by ZOU's Lab. For any questions, feedback, or support, we encourage you to connect with us:

• Issue Tracker: For reporting bugs, suggesting enhancements, or requesting features, please use our GitHub Issues.
• Email: For direct communication, you can reach out to Pinglu Zhang at [email protected].