bsun0802/eQTL-mapping

"Demos for eQTL-mapping in Python with Quantile Regression"

eQTL-mapping with Quantile Regression

This repository contains a runnable python script which performs eQTL mapping with OLS/Quantile Regression, with an optional inverse-normal-transformation (INT).

1. Input the expression matrix, covariates, variants genotype.
2. --int Turn on inverse-normal-transformation to normalize gene expression to a Gaussian distribution if needed
3. Select regressor: either OLS or QuantReg.
4. Output the p-value, slope, and t-value of coefficient for genotype.

This repository serves as a supplementary for the manuscripts submitted to Briefs in Bioinformatics. see: http://@OUR-PUBLICATION-URL

you can install the packages required by pip install -r requrirements.txt

Usage

Quantile Regression without inverse-normal-transformation.
python eQTL_mapping.py data/expr_TPM.csv data/covariates.csv data/variant_genotype.csv output/test_output.txt --model QuantReg

OLS with inverse-normal-transformation (OLS is sensitive to outliers, so it is wise to turn on INT)
python eQTL_mapping.py data/expr_TPM.csv data/covariates.csv data/variant_genotype.csv output/test_output.txt --model OLS --int

Use python eQTL_mapping.py -h to see complete options list.

Demo

See the demo in Example.ipynb The input files here are supposed to be well-structured in csv format, just for the purpose of demonstrate quantile regression.