Python (>= 3.6.8)
pytorch(tested on 1.2.0)
Numpy(tested on 1.16.3)
tqdm
Umap-learn(tested on 0.3.10)
pybloomfiltermmap3
unzip the data file under SPRITE folder before proceeding, the hyperedges with occurrence frequency 2 are not included due to the file size.
But they can be generated by the script (preprocess_SPRITE.py and analysis_SPRITE.py)
The required files for these two scripts are:
-
SPRITE cluster file from 4DN data portal.
-
The .mcool file from 4DN data portal. (Dumped into pair-wise contact named by SPRITE_contact.txt)
python main_drop.py -f adj -w hyper
for running on ChIA-Drop data
python main_SPRITE.py -f adj -w hyper
for running on SPRITE data