alperenbolat0 / dcnn-dta

DCCN-DTA: deep drug-target binding affinity prediction

This document is prepared as a project for the Bioinformatic course at TOBB ETU University.this project uses DeepDTA's datasets and subsystems. The motivation of the project is to suggest a more successful and efficient deep learning model.

By Alperen Bolat

# Installation

Data

Please see the readme for detailed explanation.

Requirements

You'll need to install following in order to run the codes.

• Python 3.4 <=
• Keras 2.x
• Tensorflow 1.x
• numpy
• matplotlib

You have to place "data" folder under "source" directory.

Running baseline

python run_baseline.py --num_windows 32 \
                          --seq_window_lengths 12 \
                          --smi_window_lengths 8 \
                          --batch_size 256 \
                          --num_epoch 100 \
                          --max_seq_len 1000 \
                          --max_smi_len 100 \
                          --dataset_path 'data/kiba/' \
                          --problem_type 1 \
                          --log_dir 'logs/'

Running proposed model

python run_master.py --num_windows 32 \
                          --seq_window_lengths 12 \
                          --smi_window_lengths 8 \
                          --batch_size 256 \
                          --num_epoch 100 \
                          --max_seq_len 1000 \
                          --max_smi_len 100 \
                          --dataset_path 'data/kiba/' \
                          --problem_type 1 \
                          --log_dir 'logs/'