Identifies significant gene expression ratios given an expression matrix and labels to predict. Currently supports binary outcome variables (i.e. responder vs nonresponder for drug response prediction).
docker run pfeiljx/taux -h
taux -x expression-data.tsv -l label-data.tsv:
-h [ --help ] Display this help message
-x [ --expression ] arg Path to expression matrix in TSV format
-l [ --label ] arg Path to sample labels in TSV format
-r [ --strata ] arg Path to strata for cross validation
sampling
--mode arg Runs analysis in sensitive mode
-n [ --normalize ] MinMax normalize expression data
-t [ --threshold ] arg (=3) T-statistic threshold for filtering
ratios
-p [ --splits ] arg (=1) Number of parallel splits to use.
-c [ --cpus ] arg (=16) Number of parallel splits to use.
-u [ --mean-filter ] arg (=1) mean expression filter. Default (1.0)
-v [ --var-filter ] arg (=80) Expression variance filter. Take top %
of most variable genes. Default (80)
-m [ --min-subtype ] arg (=5) Absolute number of samples to be
considered a subtype.
-o [ --output ] arg (=taux-output.tsv)
Path to output file
-d [ --debug ] Debug mode.
Content type
Image
Digest
sha256:714fef027…
Size
414.8 MB
Last updated
over 2 years ago
docker pull pfeiljx/taux