ISSEC adopts deep learning to learn specific patterns within predicted inter-residue contacts and subsequently identifies the objects having these patterns as inter-SSE contacts.
- Set your config in
./libs/config/config_v1.py
. - Specify your raw data path in
read_into_tfrecord.py
, put your data in the path (Example in./data/traindata
) and runpython read_into_tfrecord.py
for tfrecord generation. - Run
python train.py
for training, model will be saved in./output
. - To test your model on your dataset, you should put the files (
.ccmpred
,.ss3
,.pdb
and.fasta
) of your dataset in./data/testdata/<dataset name>/
, and runpython test.py -m <model path> -d <dataset name> [options]
; example thatpython test.py -m output/new_train_ss3 -d psicov
.