rawDIAtect

This script process the DIA-NN output of any DIA LC-MS run analyzed with the CARD AMR databases and a species background from UniProt (modified databases can be found in /database). It filters for unique AMR related peptides and only takes a selected amount of them. Furthermore it calculates the TOP-3 precursor quantity and potiental Drug Class spectra. The results are then collected into pdf reports. It can analyze multiple DIA-NN outputs at once and reports every sample individual.

Installation

[Tested on Windwos 10, RStudio v1.2.1335, R v4.2.1]

Dependecies

• readr
• tidyverse
• profvis
• tidyr
• ggplot2
• dplyr
• cowplot
• plyr
• devtools
• gameofthrones
• tm
• crayon
• progress
• Hmisc

Run

install.packages("devtools")
library("devtools")
devtools::install_github("https://github.com/CptChiler/rawDIAtect")

or from release https://github.com/CptChiler/rawDIAtect/releases

You may need to restart RStudio to see the new package.

Databases for DIA-NN and settings

In the folder /database are two zip files, those are the two protein CARD fastas and some proteomic background fastas. If your want to run DIA-NN with your MS data you need to unzip them to use them.

AMR Databases

1. “Full CARD” (196.072 entries) -> Modified CARD
2. “Whitelisted CARD” (22 entries) -> PCR (CMY-2, CTX-M-9/15, KPC-2/3, NDM-1, OXA-1/48, SHV-1/12, TEM-1/52, VIM-1, AAC(6)-lb-cr, VanA-R/S and VanB-R/S)

DIA-NN setup

1=Add your raw DIA file, 2=Add a species background and a CARD target fasta, 3= Check the boxes for Fasta digestion and Deep learning spectra prediction. 4= Uncheck batch process and check Unrelated runs. Maunual set Mass accuracy and MS1 accuracy to 10 and 20 ppm.

Usage example

Input needs to be a folder with "report".tsv files from DIA-NN and runs should have unique names. You can put multiple reports in one folder.

Minimal input:

rawDIAtect(path_in = "/Quants")

Options:

iso_diff = 2 (0 = Off), Allows a max difference to other protein isoforms
pep_filter = 3 (<1 = Off), Sets threshold for unique peptide per hit
path_out = (Default = path_in), Path were report and split data will be stored
Exp_name = "Induced", Experiment name (all files will get this name)

Expert input:

rawDIAtect(path_in = "/Quants", pep_filter = 3 ,iso_diff = 2, path_out = "/AMRs", Exp_name = "Induced")

Output

A=AMR gene families, B=expected Drug classs spectra, C= On top possibly protein isoform and bottom the count of unique peptides.

A=AMR gene families, B= All possible protein isoforms in relation.

A=TOP-3 preqursor quantity per AMR gene family, B= Drug spectra proportion of all peptides found.

Citation

Unbiased Antimicrobial Resistance Detection from Clinical Bacterial Isolates Using Proteomics
Christian Blumenscheit, Yvonne Pfeifer, Guido Werner, Charlyn John, Andy Schneider, Peter Lasch, and Joerg Doellinger
Analytical Chemistry 2021 93 (44), 14599-14608
DOI: 10.1021/acs.analchem.1c00594

Release History

*v0.5 (25.11.2022)
v0.4 (20.01.2021)
v0.3 (23.11.2020)

Meta

Christian Blumenscheit – @ChrisMiBiFlower Distributed under the GNU GENERAL PUBLIC LICENSE Version 3, 29 June 2007. See LICENSE for more information. https://github.com/CptChiler