In oganm/allenBrain: Acquiring puctures from allen brain atlas

knitr::opts_chunk$set(echo = TRUE)
library(knitr)
library(memoise)
library(dplyr)
# listGenes = memoise(listGenes)
# genes = listGenes()
# saveRDS(listGenes,'listGenes.rds')
devtools::load_all()
listGenes = readRDS('listGenes.rds')
dir.create('README_files',showWarnings = FALSE)

Table of Contents

• allenBrain
  • Example usage
    • Image acquisition
    • Region expression data
    • Incomplete stuff

allenBrain

This R package acquires pictures from Allen Brain Atlas.

Install allenBrain to R

devtools::install_github('oganm/allenBrain')

Example usage

Image acquisition

You can use downloadImage and downloadAtlas functions to get images. Output of these functions are magick-image objects

library(dplyr)


# get a list of structure names and ids
IDs = getStructureIDs()
IDs %>% head

# get the id of the desired region
granuleID = IDs['Dentate gyrus, granule cell layer' == IDs$name,]$id

# get the dataset for the desired gene (the first saggital experiment that did not fail)
datasetID = getGeneDatasets(gene = 'Prox1',
                            planeOfSection = 'sagittal',
                            probeOrientation = 'antisense')[1]


# get the slide that has the desired brain region and coordinates of the center of the region
imageID = structureToImage(datasetID = datasetID, regionIDs = granuleID)

# get the closest atlas image. 
atlasID = imageToAtlas(imageID$section.image.id,imageID$x,imageID$y,planeOfSection ='sagittal')

# decide how much to you wish to downsample
downsample = 2

# download the slide
downloadImage(imageID = imageID$section.image.id, 
             view = 'projection',
             outputFile = 'README_files/image.jpg',
             downsample = downsample)

# download the atlas
downloadAtlas(imageID = atlasID$section.image.id, 
             outputFile = 'README_files/atlas.jpg',
             downsample = downsample)

Images can be centered by providing center coordinates of a brain region. Input is either a file path or a magick-image object

# crop the slide so that the desired brain region is in the center
centerImage(image = 'README_files/image.jpg', 
            x = imageID$x,
            y= imageID$y,
            xProportions = c(.1,.1),
            yProportions =c(.1,.1),
            outputFile = 'README_files/cropped.jpg',
            downsample = downsample)

centerImage(image = 'README_files/atlas.jpg', 
            x = atlasID['x'],
            y= atlasID['y'],
            xProportions = c(.1,.1),
            yProportions =c(.1,.1),
            outputFile = 'README_files/croppedAtlas.jpg',
            downsample = downsample)

Adding a scale to the image

getSectionImage function returns all available information about an image. We can use this to access the resolution information which lists micron/pixel ratio in a given image.

(image_res = getSectionImage(imageID$section.image.id)$resolution)

We can then use the add_scale function to add a scale

downloadImage(imageID = imageID$section.image.id, 
             view = 'projection',
             downsample = downsample) %>% 
    centerImage(x = imageID$x,
            y= imageID$y,
            xProportions = c(.1,.1),
            yProportions =c(.1,.1),
            outputFile = 'README_files/cropped.jpg',
            downsample = downsample) %>% 
    add_scale(image_res,downsample)

Image syncronization

You can get closest points of other slides from the same dataset to get other slides depicting the region

# gel all images for Prox1 experiment
allImages = listImages(datasetID)  %>% arrange(as.numeric(`section.number`))

# get coordinates that are closest to the center of the brain region
closeSections = imageToImage2D(imageID$section.image.id,imageID$x,imageID$y,allImages$id)

# download and crop them all
croppedImage = closeSections %>% apply(1,function(x){
    # download and crop the images
    image = downloadImage(imageID = x['section.image.id'], 
             view = 'projection',
             # outputFile = file.path('README_files/allProx1',x['section.image.id']),
             downsample = downsample)

    centerImage(image = image, 
            x = x['x'],
            y= x['y'],
            xProportions = c(.1,.1),
            yProportions =c(.1,.1),
            # outputFile = file.path('README_files/allProx1',x['section.image.id']),
            downsample = downsample)
}) %>% do.call(c,.)

# some magick
animation = magick::image_animate(croppedImage, fps = 1)
magick::image_write(animation, "README_files/Prox1.gif")

Region expression data

Region expression can be acquired by datasetID. Data displayed in ABA web portals is expression.energy.

head(getStructureExpressions(datasetID))

If you want to get all genes, use listGenes to get all available genes for the species. Then do getGeneDatasets.

genes = listGenes()

geneDatasets = genes$acronym[1:10] %>% lapply(getGeneDatasets)

You may want to limit your search space as getting the data for all genes is a slow process.

Incomplete stuff

Grid data of a dataset can be downloaded by gridData function

gridData(datasetID = datasetID,
         outputFile ='README_files/Prox1_data.zip',
         include = c('energy','density','intensity'))
unzip(zipfile = 'README_files/Prox1_data.zip',exdir = "README_files")

oganm/allenBrain documentation built on March 28, 2024, 11:16 p.m.