Python库 | refine-contigs-0.0.12.tar.gz

# refineC [](https://github.com/genomewalker/refine-contigs/releases) [](https://github.com/genomewalker/refine-contigs/actions) [](https://pypi.org/project/refine-contigs/) [](https://anaconda.org/genomewalker/refine-contigs) refineC is a simple tool to identify potential misassemblies in contigs recovered from ancient metagenomes. The assembly of ancient metagenomics data is challenging due to the short length and post-mortem damage in reads. The short length of the reads pushes the limits of the assemblers, and the recovered contigs might contain misassemblies that can, for example, lead to misleading ecological insights, spurious pangenomic analyses, erroneous prediction of the functional potential, or impact the outcome of the binning process by mixing distantly related or unrelated phylogenetic gene markers. With the assembly of ancient metagenomics data, we can face different problems: - **Inter-genomic mosaic**: Chimeric contigs containing a mixture of sequence from multiple organisms - **Intra-genomic mosaic**: Chimeric contigs mixing different genomic regions from the same organism - **Temporal mosaic**: Chimeric contigs containing a mixture of sequence from different times (and organisms) At the moment, refineC can mitigate the effects of the first and second type of problems by exploiting the information of de Bruijn (Megahit/Spades) and overlap-based assemblers (PLASS/PenguiN). While the de Bruijn assemblers will assemble longer contigs, the overlap-based assemblers will recover more of the ancient sequence space present in the ancient sample. In any case, both types of assemblers will end up generating misassembled contigs, especially when we fine-tune the assemblers to recover as much as possible. RefineC follows a simple approach to identify the misassemblies: - Perform an all-vs-all contig comparison - Identify groups of contigs that share a certain amount of sequence identity and coverage - Find overlapping regions for each contig supported by other contigs and extract the longest one. Keep the leftover parts of the contigs if they are longer than a certain threshold - Remove redundancy of the overlapping fragments by sequence clustering - Add the new contig set to the original set of contigs without hits in the all-vs-all comparison - Remove redundancy by sequence clustering  In addition, `refineC` has a **merge** option where it tries to find contigs that might be merged using Minimus2. On some occasions, there are overlaps between contigs that are very well supported by other contigs in the sample, and refineC cannot collapse them. This usually happens when terminal regions of the contigs are overlapping other terminal regions. For this reason, although not usually recommended, we use Minimus to merge overlapping contigs. We apply a conservative approach to the already refined contigs, where we find overlaps in the same manner as in the `split` subcommand, but in this case, we select the maximum clique in a component to be merged by Minimus2. # Installation We recommend having [**conda**](https://docs.conda.io/en/latest/) installed to manage the virtual environments ### Using pip First, we create a conda virtual environment with: ```bash wget https://raw.githubusercontent.com/genomewalker/refine-contigs/master/environment.yml conda env create -f environment.yml ``` Then we proceed to install using pip: ```bash pip install refine-contigs ``` ### Using conda ```bash conda install -c conda-forge -c bioconda -c genomewalker refine-contigs ``` ### Install from source to use the development version Using pip ```bash pip install git+ssh://git@github.com/genomewalker/refine-contigs.git ``` By cloning in a dedicated conda environment ```bash git clone git@github.com:genomewalker/refine-contigs.git cd refine-contigs conda env create -f environment.yml conda activate refine-contigs pip install -e . ``` # Usage refineC only needs a contig file. For a complete list of option ``` $ refineC --help usage: refineC [-h] [--version] [--debug] {split,merge} ... Finds misassemblies in ancient data positional arguments: {split,merge} positional arguments split Find misassemblies merge Merge potential overlaps optional arguments: -h, --help show this help message and exit --version Print program version --debug Print debug messages (default: False) ``` For the split mode: ``` $refineC split --help usage: refineC split [-h] [--tmp DIR] [--threads INT] [--keep-files] [--output OUT] [--prefix PREFIX] --contigs FILE [--min-id FLOAT] [--min-cov FLOAT] [--glob-cls-id FLOAT] [--glob-cls-cov FLOAT] [--frag-min-len INT] [--frag-cls-id FLOAT] [--frag-cls-cov FLOAT] optional arguments: -h, --help show this help message and exit --tmp DIR Temporary directory (default:./tmp) --threads INT Number of threads (default: 2) --keep-files Keep temporary data (default: False) --output OUT Fasta file name to save the merged contigs (default: contigs) --prefix PREFIX Prefix for contigs name (default: contig) required arguments: --contigs FILE Contig file to check for misassemblies overlap identification arguments: --min-id FLOAT Minimun id to use for the overlap (default: 0.9) --min-cov FLOAT Minimun percentage of the coverage for the overlap (default: 0.25) global clustering arguments: --glob-cls-id FLOAT Minimum identity to cluster the refined dataset (default: 0.99) --glob-cls-cov FLOAT Minimum coverage to cluster the refined dataset (default: 0.9) fragment refinement arguments: --frag-min-len INT Minimum fragment length to keep (default: 500) --frag-cls-id FLOAT Minimum identity to cluster the fragments (default: 0.9) --frag-cls-cov FLOAT Minimum coverage to cluster the fragments (default: 0.6) ``` And for the `merge` mode: ``` $refineC merge --help usage: refineC merge [-h] [--tmp DIR] [--threads INT] [--keep-files] [--output OUT] [--prefix PREFIX] --contigs FILE [--min-id FLOAT] [--min-cov FLOAT] [--glob-cls-id FLOAT] [--glob-cls-cov FLOAT] [--mnm2-threads INT] [--mnm2-overlap INT] [--mnm2-minid FLOAT] [--mnm2-maxtrim INT] [--mnm2-conserr FLOAT] optional arguments: -h, --help show this help message and exit --tmp DIR Temporary directory (default:./tmp) --threads INT Number of threads (default: 2) --keep-files Keep temporary data (default: False) --output OUT Fasta file name to save the merged contigs (default: contigs) --prefix PREFIX Prefix for contigs name (default: contig) required arguments: --contigs FILE Contig file to check for misassemblies overlap identification arguments: --min-id FLOAT Minimun id to use for the overlap (default: 0.9) --min-cov FLOAT Minimun percentage of the coverage for the overlap (default: 0.25) global clustering arguments: --glob-cls-id FLOAT Minimum identity to cluster the refined dataset (default: 0.99) --glob-cls-cov FLOAT Minimum coverage to cluster the refined dataset (def