This repository demonstrates the application of the ARACNE algorithm for gene regulatory network reconstruction using a synthetic dataset. It includes the Python code for generating the synthetic dataset, implementing the ARACNE method, and analyzing the results.
This repository contains the code and results of the ARACNE network reconstruction method applied to a synthetic dataset. The ARACNE method is an algorithm for inferring gene regulatory networks from gene expression data.
-
Data Generation: A synthetic gene regulatory network of 100 genes and 500 interactions was randomly generated. The expression data for this network was simulated using a Gaussian distribution determined by the covariance matrix of the network structure.
-
Network Reconstruction: The ARACNE method was applied to the synthetic dataset to reconstruct the gene regulatory network. This involved calculating the Mutual Information (MI) between each pair of genes and applying the Data Processing Inequality (DPI) to remove indirect interactions.
-
Analysis: The reconstructed network was compared to the original synthetic network to evaluate the performance of the ARACNE method.
The code for this project is written in Python and makes use of scientific computing libraries such as NumPy and SciPy. The code is organized into several Jupyter notebooks, each corresponding to a different step in the process.
The results of the network reconstruction are presented in the form of graphs and tables. These include the reconstructed network, the comparison between the reconstructed and original networks, and various performance metrics.
Future work will involve applying the ARACNE method to real gene expression data to further evaluate its performance. Additionally, the impact of various parameters on the performance of the ARACNE method will be explored.
关于ARACNE的详细介绍请见 https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-7-S1-S7