BANNER is still under development - features and improvements are being added, so please check back soon.
BANNER
is a tool that lives inside HULK and aims to make sense of hulk sketches. At the moment, it trains a Random Forest Classifier using a set of labelled hulk sketches. It can then use this model to predict the label of microbiomes as they are sketches by HULK
.
For example, you could train BANNER
using a set of microbiomes from patients that either have or haven't received antibiotic treatment. You can then use BANNER
to predict whether a new microbiome sample exhibits signs of antibiotic dysbiosis. I will post more information and examples soon...
conda install banner
note: if using Conda make sure you have added the Bioconda channel first
pip install banner
BANNER
is called by typing banner, followed by the subcommand you wish to run. There are two main subcommands: train and predict. This quick start will show you how to get things running but it is recommended to follow the HULK documentation.
# Train a random forest classifier
banner train -m hulk-banner-matrix.csv -o banner.rfc
# Predict the label for a hulk sketch
hulk sketch -f mystery-sample.fastq --stream -p 8 | banner predict -m banner.rfc
## Notes
-
only supports 2 labels at the moment
-
there is very limited checking and not many unit tests...