This code is still under active initial development, so major backwards-incompatible changes might occur

mohawk

Neural nets for taxonomic profiling of microbial communities from metagenomic shotgun sequencing data.

Install

Currently this repo can be installed

git clone https://github.com/gwarmstrong/mohawk.git
cd mohawk
conda env create -n mohawk -f resources/environment.yml
conda activate mohawk
pip install -e .

Usage

Command Line Interface (CLI)

An example of using the mohawk API to train a neural network on some pre-specified genomes with designated classes on a given number of reads, you can use a command similar to the one shown below:

mohawk train seq-by-seq-pytorch \
    --genome-ids example_input/min-red_0.1__n_4__group_0__genome-ids.txt \
    --model-name ConvNetAvg \
    --log-dir example_output \
    --data-dir example_input/genomes \
    --metadata example_input/wol_supplemental_metadata.tsv \
    --train-ratio 0.8 \
    --gpu True

An example of using the interface for using a pre-trained mohawk model to classify reads from a fastq file is included below:

mohawk characterize classify-seq-by-seq-pytorch \
    --model /path/to/model \
    --sequence-file /path/to/input/fastq \
    --output-file /path/to/ouput/file \
    --length 150

Metadata

An up-to-date assembly summary from RefSeq can be obtained from: ftp://ftp.ncbi.nlm.nih.gov/genomes/ASSEMBLY_REPORTS/assembly_summary_refseq.txt