DeepMHCI

DeepMHCI: An Anchor Position-Aware Deep Interaction Model for Accurate MHC-I peptide Binding Affinity Prediction

Requirements

• python == 3.8.8
• pytorch == 1.7.1
• numpy == 1.19.2
• scipy == 1.6.1
• scikit-learn == 0.24.1
• click == 7.1.2
• ruamel.yaml == 0.16.12
• tqdm == 4.56.0
• logzero == 1.6.3

Experiments

The commands corresponding to the different experiments are shown below.

1. Train 10 models to ensemble for five-fold cross-validation.
2. Test on testsets with 10 models (after 5cv training).
3. Test on the epitope dataset with 10 models (after 5cv training).
4. Output the top 1% predicted binders to draw sequence logos.

python main.py -d config/data.yaml -m config/model.yaml --mode 5cv -s 0 -n 10 --eval_len
python main.py -d config/data.yaml -m config/model.yaml --mode test-5cv -s 0 -n 10 --eval_len
python main.py -d config/data.yaml -m config/model.yaml --mode epitope -s 0 -n 10 --eval_len
python main.py -d config/data.yaml -m config/model.yaml --mode seq2logo -s 0 -n 10 --allele HLA-A1101

Declaration

It is free for non-commercial use. For commercial use, please contact Mr.Wei Qu and Prof.Shanfeng Zhu ([email protected]).