geNorm

A Python package for RNAseq housekeeping (or, reference) normalisation.

Read the docs.

You can:

• Run the geNorm algorithm [1], which automatically selects reference genes by recursively eliminating genes with high $M$ value.
• Compute the gene-stability measure $M$ for reference genes in a given set of samples.

Here, $M$ is defined in terms of the average variation in log-ratio expression:

M_j = \sum_{k=1}^n\frac{V_{jk}}{n-1}

where

A_{jk}^{(i)} = \log_2 \frac{a_{ij}}{a_{jk}}; V_{jk} = \sqrt{\mathrm{Var}(A_{jk})};

with expression $a_{ij}$ referring to gene $j$ in sample $i$.

Installation

You can grab geNorm from the Python Package Index:

pip3 install rna-genorm

Example

from pandas import DataFrame
from genorm import m_measure, genorm


# Expression data for three control genes.
counts = DataFrame(
    [[ 1,  2,  1],
    [ 3,  6,  5],
    [ 5, 10,  9],
    [ 3,  6,  5]],
    columns=['gene_a', 'gene_b', 'gene_c'],
    index=[f'sample_{i}' for i in range(1, 5)],
)

# Compute `M` value for this set of control genes.
m_measure(counts)

# Select top 2 control genes with lowest `M`.
gene_names, m_values = genorm(counts, n_stop=2)

Acknowledgements

Made by Hylke Donker & Bram van Es and open sourced under the Apache 2 license.

References:

[1]: Vandesompele, Jo, et al. "Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes." Genome biology 3.7 (2002): 1-12.