README.md

In kieranrcampbell/SpatialPRo: SpatialPRo: methods for spacial proteomics data

SpatialPRo - spatial proteomics analysis

Installation

Using devtools call install_github("kieranrcampbell/SpatialPRo")

Usage

SpatialPRo contains two different classes - SPData that represents a single (tissue) sample and SPExp that represents the entire experiment, storing a list of SPDatas, the sample ids and the directory and filenames of the original files.

SPData

There is an example dataset included - to load it first load the library via library(SpatialPRo) then call data(sp5). This loads an example object of class SPData called sp. The following can then be used to access data elements:

Generic SPData methods

• channels(sp) returns the channel names
• nChannel(sp) returns the number of channels
• nCells(sp) returns the number of cells
• rawData(sp) returns a cell by channel matrix of the raw data
• cells(sp) returns a cell by channel matrix with the readouts for each cell, normalised wrt cell size & concentration
• neighbours(sp) returns a list of length nCells, where entry i is an n by channel matrix for cell i having n neighbours
• neighbourIDs(sp) returns a list of the nearest neighbour cell identifiers
• size(sp) returns a vector containing the cell sizes (in pixels)
• weight(sp) returns the boundary sizes for each cell to its nearest neighbours
• ID(sp) returns the sample ID
• cellClass(sp) returns a vector of length nCells where each cell is assigned a class. Here there are 2 classes, with 1 corresponding to stromal and 2 corresponding to tumour.
• xy(sp) returns a nCell by 2 matrix of centre-of-mass locations for each cell
• An sp object can be partitioned using the [ operator, so sp[1:2, c(3,5,7)] returns a new SPData object using cells 1 & 2 and channels 3, 5 and 7

Methods for dealing with different cell classes

• cellClass(sp) returns a vector of length nCells(sp) with a numeric classification for each cell (currently only 2 types are supported)
• neighbourClass(sp, cell.class) functions similarly to cellClass but returns a list similar to neighbours(sp) but with only neighbours of class cell.class remaining
• neighbourChannel(sp, c(1,3,5) returns the neighbour list as in neighbours(sp) but including only channels 1,3 and 5
• findBoundary(sp) returns the (numeric identifiers of) cells that lie on the boundary between 2 different classes

The class SPData also contains several plotting routines:

• boxplots(sp,4,8) plots the ranges of each of the 32 channels (in 4 rows by 8 columns)
• channelPlot(sp, 1:4) provides boxplots of the first 4 channels but highlighting differences between different classes of data (class1 vs class2)

SPExp

An SPExp has four slots:

1. dir Directory location of experiment files
2. files List of files used in the experiment
3. spdata List containing objects of class SPData
4. ids Vector of sample ids

Generic SPExp Methods

• files(spe) shows the experiment files
• getDir(spe) experiment directory
• IDs(spe) sample IDs
• spe[2:3] subset samples 2-3
• spe[[4]] return an SPData object of the fourth sample
• SPlist(spe) the underlying list holding the SPData objects
• length(spe) number of samples

Example workflows

Download all the original matlab files from https://s3.amazonaws.com/supplemental.cytobank.org/report_data/report_113/Figure_5/Figure_5_raw_image_files.zip and extract to ~/myfolder

We can then initialise an empty SPE with

spe <- SPExperimentfromDir("~/myfolder")

If the files option is left blank then all files are used.

We can then load all the files into the object using

spe <- loadExp(spe)

(this may take some time for matlab -> R conversion)

kieranrcampbell/SpatialPRo documentation built on May 20, 2019, 9:24 a.m.