mlgenotype

The mlgenotype Python package can be used to train machine learning models to genotype structural variants using unaligned short read data (fastq or bam-formatted files), as well as to predict genotypes for samples using whole genome short read datasets.

The software was written by Nancy Fisher Hansen, a staff scientist in the Computational and Statistical Genomics Branch of NHGRI, beginning with code written by Gracelyn Hill and Jennifer C Lin. Nancy can be reached at nhansen@mail.nih.gov.

Install

The easiest ways to install mlgenotype are from PyPi with Python’s pip installer, or by using conda to install the bioconda mlgenotype package.

Pip/PyPi

To install mlgenotype with Python’s pip installer, first create a virtual environment. Then use pip install to install the latest version of mlgenotype:

python3 -m venv mlgeno_env
python3 -m pip install mlgenotype

Conda

The mlgenotype package is also hosted on anaconda and available through the bioconda channel:

conda create -n mlgeno -c bioconda -c conda-forge mlgenotype
conda activate mlgeno

From github

If you prefer not to use a package manager, it also works to clone the github repository and run Python’s setuptools installer:

git clone git://github.com/nhansen/mlgenotype
cd mlgenotype
python3 setup.py install

Note that installing from github requires you to first satisfy mlgenotype’s software dependencies:

• pandas >= 1.0
• scikit-learn == 1.0.2