This python CLI tool can rank genes based on their differential expression between a 'case' and a 'control' status, from most up-regulated to most down-regulated, with a variety of metrics.
This is not a differential expression analysis, as no p-values are computed since no statistical test is used. It is just meant to give each gene a 'rank' and a rank value in order to run other statistical methods such as pre-ranked GSEA.
Input data should be the base-2 logarithm of 1 + read counts (i.e log2(counts + 1)). The program un-logs the data when appropriate (e.g. running DESeq2).
Currently supported ranking methods:
- Fold Change: The
fold_changemethod computes a simple difference of average fold changes between the case and controls. - Cohen's d: The
cohen_dmetric computes Cohen's d between the different expression values of each gene. - DESeq2 Shrunk Log Fold Change: Uses
DESeq2's LFC shrinking method to compute LFCs, and uses them as ranking metric. This uses PyDESeq2, so the input data is always normalized in the process. - Signal to Noise ratio: Compute the signal to noise ratio between the control and case genes. This is roughly the mean divided by the variance of each gene.
- Baumgartner-Weiss-Schindler test statistic: Compute the
BWS statistic
for each gene. Uses a
scipyprimitive so it's much faster than usingbws_test.
Most of these methods come with a normalized version, where the input is first
normalized with the "mean of ratios" method
(as implemented by DESeq2).
They are usually named as norm_<method>.
You can use generanker --list-methods for a list of all the methods.
Install Python and cargo.
Install the prerequisite fast-cohen executable with:
cargo install --git https://github.com/MrHedmad/fast-cohen.gitThen, install the tool with:
# I suggest you do this in a virtual environment:
# python -m venv env && source env/bin/activate
python -m pip install git+https://github.com/TCP-Lab/gene_ranker.gitYou may then use generanker from the command line.
Use generanker --help for additional usage details.