Skip to content

Latest commit

 

History

History
115 lines (95 loc) · 5.25 KB

Demo_Training_Detectability.md

File metadata and controls

115 lines (95 loc) · 5.25 KB
Raw

DeepDIA Demo: Training a New Model for Detectability Prediction

Training a new model for MS detectability prediction using data-dependent acquisition (DDA) data.

1. System Requirements

This demo has been tested on a workstation with Intel Xeon E5-2690 v3 CPU, 16 GB RAM, and Microsoft Windows Server 2016 Version 1607 (OS Build 14393.2430) operating system with the following softwares:

A GPU card with Compute Unified Device Architecture (CUDA) is recommended, e.g. NVIDIA GeForce GTX 1050 Ti.

2. Demo Data

LC-MS/MS DDA data of HeLa and HEK-293 cells on Q Exactive HF are available at ProteomeXchange (http://proteomecentral.proteomexchange.org/) with the data set PXD005573. (Bruderer, R. et al. Mol. Cell. Proteomics 2017, 16, 2296-2309.)

  • Fig4_HeLa-1m-HPRP-10perc_DDA_R01_T0.raw
  • Fig4_HeLa-1m-HPRP-15perc_DDA_R01_T0.raw
  • Fig4_HeLa-1m-HPRP-20perc_DDA_R01_T0.raw
  • Fig4_HeLa-1m-HPRP-25perc_DDA_R01_T0.raw
  • Fig4_HeLa-1m-HPRP-50perc_DDA_R01_T0.raw
  • Fig4_HeLa-1m-HPRP-5perc_DDA_R01_T0.raw
  • Fig4_HeLa-1m-HPRP-FT_DDA_R01_T0.raw
  • Fig4_HeLa-1m_DDA_R01_T0.raw
  • Fig4_HeLa-1m_DDA_R02_T0.raw
  • Fig4_HeLa-1m_DDA_R03_T0.raw
  • Fig4_HEK293-1m-HPRP-10perc_DDA_R01_T0.raw
  • Fig4_HEK293-1m-HPRP-15perc_DDA_R01_T0.raw
  • Fig4_HEK293-1m-HPRP-20perc_DDA_R01_T0.raw
  • Fig4_HEK293-1m-HPRP-25perc_DDA_R01_T0.raw
  • Fig4_HEK293-1m-HPRP-50perc_DDA_R01_T0.raw
  • Fig4_HEK293-1m-HPRP-5perc_DDA_R01_T0.raw
  • Fig4_HEK293-1m-HPRP-FT_DDA_R01_T0.raw
  • Fig4_HEK293-1m_DDA_R01_T0.raw
  • Fig4_HEK293-1m_DDA_R02_T0.raw
  • Fig4_HEK293-1m_DDA_R03_T0.raw

SwissProt Homo sapiens database (FASTA) can be downloaded from UniProt (https://www.uniprot.org/). The FASTA file (2018-04 version, 20,301 entries) has been deposited to ProteomeXchange via the iProX partner repository with the data set identifier PXD014108/IPX0001628000.

  • swissprot_human_201804_validated.fasta

The saved project and exported results from SpectroMine are also available at ProteomeXchange/iProX with identifier PXD014108/IPX0001628000.

  • HeLa_4h_DDA.psar.zip
  • HeLa_4h_DDA.csv.zip
  • HEK293_DDA.psar.zip
  • HEK293_DDA.csv.zip

3. Prepare Training Data

In this demo, SpectroMine reports are used, which should be exported with the schema provided in the misc/SpectroMine_Report_Schema folder.

  • PeptideReport.rs
  • ProteinReport.rs

The file name of peptide report should end with .PeptideReport.csv, and that of protein report should end with .ProteinReport.csv e.g. HEK293.PeptideReport.csv and HEK293.ProteinReport.csv.

SpectroMine Manual is available at https://biognosys.com/shop/spectromine.

Start RStudio, ensure package readr has been installed.

install.packages("readr")

Open deepdetect/R/init.R and run the script by clicking Source.

source("{PATH_TO_CODE}/deepdetect/R/init.R")

Set the working directory to the reports and run deepdetect/R/get_peptides_from_ProteinDigestionSimulator_result.R.

setwd("{PATH_TO_DATA}")
source("{PATH_TO_CODE}/deepdetect/R/get_detectability_from_SpectroMine.R")

Single hits are excluded and only proteins with sequence coverage >= 25% are taken into considerasion. The protein list file are and the peptide detectability file are generated in the working directory.

  • HEK293_excludeSingleHit_coverage25.proteinAccession.txt
  • HEK293_excludeSingleHit_coverage25.detectability.csv

Open Windows PowerShell and run deepdetect/Filter-Fasta.ps1. A FASTA file containing the filtered proteins is generated in the working directory. Rename the filtered FASTA file.

{PATH_TO_CODE}/deepdetect/Filter-Fasta.ps1 swissprot_human_201804_validated.fasta HEK293_excludeSingleHit_coverage25.proteinAccession.txt
mv swissprot_human_201804_validated.filtered.fasta HEK293_excludeSingleHit_coverage25.fasta

Perform in silico digestion using Protein Digestion Simulator. Tryptic (no Proline Rule) is selected as digestion enzyme. Digestion is performed with the following parameters:

  • Max Miss Cleavages: 2
  • Minimum Residue Count: 7
  • Maximun Fragment Mass: 6000
  • Minimun Fragment Mass: 0

DeepDIA only supports peptide sequences with standard amino acids (ACDEFGHIKLMNPQRSTVWY) and length <= 50.

Open deepdetect/R/get_negative_peptides.R and run the script by clicking Source.

source("{PATH_TO_CODE}/deepdetect/R/get_negative_peptides.R")

The negative peptide file is generated in the working directory.

  • HEK293_excludeSingleHit_coverage25_negative.detectability.csv

Open deepdetect/R/get_cleavage_window.R and run the script by clicking Source.

source("{PATH_TO_CODE}/deepdetect/R/get_cleavage_window.R")

The training data are generated in the working directory.

  • HEK293_excludeSingleHit_coverage25.detectability.csv
  • HEK293_excludeSingleHit_coverage25_negative.detectability.csv

4. Train a MS Detectability Model

Run deepdetect/py/train_hard_negative.py to start training.

python {PATH_TO_CODE}/deepdetect/py/train_hard_negative.py

Expected run time depends on the number of peptides and the performance of the computer. In this demo, this command may take several hours to a day.

In the training_{round}/models folder, we find the trained model (with checkpoints during training).