An application to extract meaningful information from datasets of ChIP-seq experiments.
The code in Python represents the core part of the application, which uses the random forest model in the sklearn library (scikit-learn) to detect differential features between two groups of ChIP-seq experiments binarised using ChromHMM.
This software has been tested to detect the human sex dimorphism between two groups (males and females). The output can be further analysed using the R library associated, chromatinsight.tools https://github.com/ricolab/chromatinsight.tools .
Input:
- The output text files of ChromHMM (one for each chromosome). At least ten samples in each group are needed to provide enough statistical power.
- A bed file with the genomic regions (such as TADs).
Output:
- A text file with the degree of dimorphism for each genomic region and each trial of the algorithm.
- A second text file (optional) with the degree of dimorphism of each genomic region after randomising the sample labels (important to display the random behaviour and calculating the FDR).
A basic unit test is provided, which also serves as example of how to run the code; it includes files for twenty samples (ten men and ten women). To process it, use the data in ./utest/input and use the code provided in chromatinsight_utest.py .
Requirements:
Python 2.7
pandas installed, see https://pandas.pydata.org/pandas-docs/stable/getting_started/install.html
sklearn installed, see https://scikit-learn.org/stable/install.html
Quick way to install both:
pip install pandas
pip install sklearn