fetchm 0.1.0

A Python tool for fetching metadata for bacterial genomes.

FetchM: Metadata Fetching and Analysis Tool

Overview

FetchM is a Python-based tool for fetching and analyzing genomic metadata from NCBI BioSample records. When you download ncbi_dataset.tsv from the NCBI genome database, the metadata fields such as 'Collection Date', 'Host', 'Geographic Location', and 'Isolation Source' are missing. This tool helps fetch the associated metadata for each BioSample ID. FetchM requires an input file (ncbi_dataset.tsv) from the NCBI genome database, retrieves additional annotations from NCBI, filters the data based on quality thresholds, and generates visualizations to help interpret the results. You can also download the filtered sequences.

Features

• Fetch metadata from NCBI BioSample API.
• Filter genomes based on CheckM completeness and ANI check status.
• Generate metadata summaries and annotation statistics.
• Create various visualizations for geographic distribution, collection dates, and gene counts.
• Download genome sequences (optional).

Installation

Using Conda

You can install FetchM in a Conda environment:

conda create -n fetchM_env python=3.8
conda activate fetchM_env
conda create -n fetchM_env -c conda-forge python=3.8 pandas requests xmltodict matplotlib seaborn scipy tqdm

Using pip

Ensure you have Python 3 installed. Install dependencies with:

pip install -r requirements.txt

Usage

Run FetchM with the following command:

fetchM --input input.tsv --outdir results/

Additional Options:

• --checkm 95 (Set CheckM completeness threshold, default: 95)
• --seq (Enable sequence download mode)

Output

FetchM creates multiple output files inside the results/ directory:

• Metadata summaries in metadata_output/
• Figures in figures/
• Filtered datasets for further analysis

Visualizations

Annotation Distributions

Assembly Statistics

Metadata Summaries

Scatter Plots

License

This project is licensed under the MIT License.

Author

Developed by Tasnimul Arabi Anik.

Contributions

Contributions and improvements are welcome! Feel free to submit a pull request or report issues.