DbyDeep

DbyDeep: Exploration of MS detectable peptides via deep learning

Hardware

DbyDeep requires

• a GPU with CUDA support

Installation

DbyDeep requires GPU setting and conda environment.

1. GPU setting for using tensorflow
   DbyDeep was tested on Ubuntu 18.04, CUDA 11.1, CUDNN 8.0.5 with Nvidia RTX 8000 and RTX A6000 graphic cards with the dependencies above.

nvidia GPU driver (https://www.nvidia.co.kr/Download/Find.aspx)
CUDA >= 11.0 (https://developer.nvidia.com/cuda-toolkit-archive) cudnn >= 8.0 (https://developer.nvidia.com/rdp/cudnn-download) https://www.tensorflow.org/install/source#linux

2. conda environment.
   tensorflow = 2.4.0
   python >=3.6, <=3.8

conda env create -f environment.yml

Model

DbyDeep assumes your models are in directories that look like this:

DbyDeep.h5 - a saved keras model and weights

Usage

1. dataset
   use ./scripts/dataset.sh file or python script.
   Currently two output formats are supported: a COMET style db_result.tsv and a MSGF+ style db_result.tsv file.

bash dataset.sh

python dbydeep_data.py
--save-path /path/to/save/
--protein-fasta /path/to/proteinDB/
--peptide-tsv /path/to/SearchResult/
--tool-name msgfplus # [msgfplus, comet]

2. prediction
   use ./scripts/model.sh file or python script.

bash model.sh

python dbydeep_model.py
--retrain-flag False
--data-path ./data/data.csv
--model-path ./data/DbyDeep.h5
--save-path ./data/
--job-name data_result

3. Using DbyDeep on your data
   You can retrain DbyDeep model to your own needs.

python dbydeep_model.py
--retrain-flag True
--data-path ./data/data.csv
--model-path ./data/DbyDeep.h5
--save-path ./data/
--job-name data_result

Please note: Sequences except 20 amino acids are not supported. Modifications are not supported.

Example

Please find an example input file at ./data/test.csv.

python dbydeep_model.py
--retrain-flag False
--data-path ./data/test.csv
--model-path ./data/DbyDeep.h5
--save-path ./data/
--job-name test_result