A small R app designed to visualize mzML data from past Ingalls Lab experiments.
It shouldn't be hard to open up a mass-spec file and browse around the data inside. But it is. Often the libraries required are difficult to install, the functions are slow and memory-inefficient, and the process produces arcane R objects from which it's difficult to extract m/z, retention time, or intensity data.
This R Shiny-based app is meant to resolve some of these problems by providing a simple user-interface to our mass-spec data, allowing the user to browse intuitively through files on their computer or connected drives, load those into R automagically, and search in those files by mass. The app is optimized for MSMS data, with scans that collected fragmentation information at a given mass denoted by points on the chromatogram, and clicking on those will pull up the fragments and their relative intensities in the bottom plot.
Loading the files into R is the slowest part of this process, but at a few seconds per file it's still
much faster than the typical readMSdata used by xcms or MSnbase. This is done via a custom
XML parser written around Hadley's xml2 library that also has the memory advantage of using external
pointers until the data itself is needed. Switching between masses should be super fast thanks to data.table's
subsetting and binary search.
Several R libraries are required, but none of which should be difficult to install.
- shiny
- shinyDirectoryInput (handles intuitive file selection)
- plotly (for making interactive plots)
- data.table (for super-fast indexing)
- xml2 (for parsing the mzML file)
- base64enc (for parsing m/z and intensity data encoded as binary arrays)
The app makes a brief check that these libraries are installed and has a small script to walk the user
through the process if they're not found, but it's probably easiest to install them yourself if they're
missing. shinyDirectoryInput is not currently on CRAN and must instead be
installed from Github.
Additionally, this app sources the custom script file "RaMS_custom.R" which
borrows several functions from the upcoming RaMS package and optimizes
them for MSMS data extraction.
Also, shinyDirectoryInput requires Powershell 3.0 or newer.
The app can be launched on Windows simply by double-clicking the "runDataBrowser.cmd"
command. This should initialize both an R script in the background and a browser
window to interact with the app. The default port is localhost:1313, but this
can be changed within the shinyApp() options to another if necessary.
Support for Mac and Linux upcoming shortly.
First, locate the mzML files of interest by clicking on the "..." button of the directory input window. Files ending in .mzML are automatically detected and rendered below. Clicking on a file name will "stage" it for loading, and multiple files can be loaded simultaneously this way.
Once the files render, click around the upper plot or enter a new mass in the sidebar. If a given file includes MSMS data, points on the upper plot will appear on the chromatogram and clicking on those will render the lower plot of fragments collected near that scan.
Additional files can be added at any time in the same way, and Plotly should assign them new colors. The legends in Plotly charts are also interactive - click on a single trace to hide it, and double click to select only that one. Zooming is also supported.
Absolutely full of 'em. Let me know if you find some, and what you were doing to cause it!