https://github.com/khanovico/neural-struct-detect

deep neural network for detecting structure of protein
https://github.com/khanovico/neural-struct-detect

artificial-intelligence biological-data-analysis deep-learning python

Last synced: about 1 month ago
JSON representation

deep neural network for detecting structure of protein

• Host: GitHub
• URL: https://github.com/khanovico/neural-struct-detect
• Owner: khanovico
• Created: 2024-06-19T12:01:55.000Z (7 months ago)
• Default Branch: master
• Last Pushed: 2024-06-26T18:23:47.000Z (6 months ago)
• Last Synced: 2024-08-05T00:57:16.676Z (5 months ago)
• Topics: artificial-intelligence, biological-data-analysis, deep-learning, python
• Language: Python
• Homepage:
• Size: 31.8 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Deep Nerual Network approach for Sturcture Detection

## Setup

The software is ready for Docker: the image can be created from `Dockerfile` by running `docker build -t deeplytough .` (image size ~4.7GB so you may have to increase the disk space available to docker). The DeeplyTough tool is then accessible

conda environment inside the container with `source activate deeplytough`.

```bash

# create new python 3 env and activate

conda create -y -n deeplytough python=3.6

# create python2 env used for protein structure preprocessing

conda create -y -n deeplytough_mgltools python=2.7

conda install -y -n deeplytough_mgltools -c bioconda mgltools=1.5.6

```

### Environment setup

To run the evaluation and training scripts, please first set the `DEEPLYTOUGH` environment variable to the directory containing this repository and then update the `PYTHONPATH` and `PATH` variables respectively:

```bash

export DEEPLYTOUGH=/path_to_this_repository

export PYTHONPATH=$DEEPLYTOUGH/deeplytough:$PYTHONPATH

export PATH=$DEEPLYTOUGH/fpocket2/bin:$PATH

```

## Training

Training requires a GPU with >=11GB of memory and takes about 1.5 days on recent hardware. In addition, at least a 4-core CPU is recommended due to volumetric input pre-processing being an expensive task.

Note that due to non-determinism inherent to the currently established process of training deep networks, it is nearly impossible to exactly reproduce the pre-trained networks in `networks` directory.

Also note the convenience of an output directory containing "TTTT" will afford this substring being replaced by the current `datetime`.