regCNN: identifying Drosophila genome-wide cis-regulatory modules via integrating the local patterns between epigenetic marks and TF binding motifs

Related paper:

Tzu-Hsien Yang*, Ya-Chiao Yang# and Kai-Chi Tu#, "regCNN: identifying Drosophila genome-wide cis-regulatory modules via ntegrating the local summarizing patterns between epigenetic marks and TF binding motifs". (Submitting)
#: These authors contributed equally.

Available Datasets:

1. The training/validation set
2. The test set
3. The list of Drosophila TFs used in regCNN
4. The list of Histone modification ChIP used in regCNN
5. The list of chromatin binding protein ChIP used in regCNN
6. The list of nucleosome ChIP used in regCNN

Model Codes for regCNN

Suggested running environments: Linux Ubuntu 16.04.6, Python 3.6.12.

1. Download the codes from the following link: Download.
2. Unzip the file:
   unzip regCNN.zip
3. Change the working directory into the model codes:
   cd regCNN
4. Download the processed transcription-regulation related dataset: Download.
5. Unzip the file:
   tar -zxvf Dataset.tar.gz
6. Install the necessary packages:
   pip install -r requirements.txt
7. Prepare the input Drosophila chromosomal regions (ver. 6) in the folder name "Input".
   Multiple chromosomal regions are allowed in the same input file. The input format should followed the following examples:
   <chromosome>,<start>,<end>
   Note: The input chromosomal regions should all start from the 5' end.
   2L,6543468,6546487
   
8. Predict the existence probability of functional CRMs:
   python main.py -i <Input_file_name> -o <Output_file_name>
9. Output: Probability that the given chromosomal location conatins a functional CRM is written to <Output_file_name>.