Skip to contents

Using Webber

Source: vignettes/using-webber.Rmd
using-webber.Rmd

Introduction

This package is a utility belt for working with Western blot images in R. It’s primarily focused on making annotation easier.

library(webber)
library(magick)
#> Linking to ImageMagick 6.9.11.60
#> Enabled features: fontconfig, freetype, fftw, heic, lcms, pango, webp, x11
#> Disabled features: cairo, ghostscript, raw, rsvg
#> Using 4 threads
library(tibble)

Getting set up with webber

You need to introduce yourself to webber so that you can more easily pull files from Sharepoint. Fortunately, that’s rather simple: just run wb_setup(). It’ll tell you to paste a few lines into your .Rprofile (which it will handily open up for you), and you’ll need to add your folder name (which should just be lastname-firstname). After you restart R, you should be ready to go (provided you’ve introduced yourself to Sharepoint)

Looking through files

Most of the time I don’t really know the exact path of the file I’m looking for. webber makes this easy with wb_list:

wb_list("/") |> head()
#>                                                                   name    size
#> 1                                Raw Data/licor/aragaki-kai/2022-04-04 4480648
#> 2                        Raw Data/licor/aragaki-kai/2022-04-20_bupa-dr 2018128
#> 3                        Raw Data/licor/aragaki-kai/2022-04-20_erda-tc 1492864
#> 4                        Raw Data/licor/aragaki-kai/2022-04-20_pemi-tc  746432
#> 5                  Raw Data/licor/aragaki-kai/2022-04-28_rt112-igfr-kd 1064087
#> 6 Raw Data/licor/aragaki-kai/2022-05-11_uc14-nt-835_nt-irf1si_wo-w-dox 4424925
#>   isdir                                 id
#> 1  TRUE 013PDU6SSJG5PQELW2JZB22SXQOVX5UNEQ
#> 2  TRUE 013PDU6SWOQEF2EVTREZCKCPYEMYQAS2BY
#> 3  TRUE 013PDU6SQZBJKUT7YZHBGZHXKHALNKEPYU
#> 4  TRUE 013PDU6SSHCQEVJEMGJRGIPXSQV7UNJBEX
#> 5  TRUE 013PDU6SXTXPJ3Y5NEWJH3CTHSNYXQ6YMN
#> 6  TRUE 013PDU6SWZLCYWV3XAZJF2YVX2LTAQL2IW

Notice how I didn’t have to specify my name - that’s because of the introduction that we did earlier. If I wanted to root around in someone else’s folder, I could do this:

wb_list("/", user = "arbuiso-alyssa")
#>                                                         name     size isdir
#> 1 Raw Data/licor/arbuiso-alyssa/2023-02-14_rt112-100nm-cm272 41214201  TRUE
#>                                   id
#> 1 013PDU6SU3ESBH6BUQTBAKQ7JFLBKJ3V6R

If you want to drill deeper into a folder, you can do this:

wb_list("/2022-04-20_bupa-dr")
#>                                                                name    size
#> 1            Raw Data/licor/aragaki-kai/2022-04-20_bupa-dr/pakt.tif  996356
#> 2 Raw Data/licor/aragaki-kai/2022-04-20_bupa-dr/pakt-s473-actin.tif 1021772
#>   isdir                                 id
#> 1 FALSE 013PDU6SQNCRN5B6WB2ZBJJNIQGWNMRGBW
#> 2 FALSE 013PDU6SWOT55JIX5VCBGLU2ZW5OYDGUHC

Getting

When you’ve found your file or folder of interest, you download it with wb_get. I’m going to download a single file, but you can do this with whole folders:

file <- wb_get(
  path = "2022-04-20_bupa-dr/pakt-s473-actin.tif",
  dest = tempfile(),
  overwrite = TRUE
)

Editing

With our file in hand, we can now edit it. For this, we call upon the delightful magick R package, which uses imagemagick under the hood.

image <- image_read(file)
image

The first part - the hardest part - is cropping it down to size.

imagemagick (henceforth magick) specifies that the TOP LEFT corner is (0, 0). So if I do this:

image_crop(image, "100x300") |>
  plot()

I get the upper left hand portion.

If I wanted to offset where the area was (I still want it to be 100x300, say, but I want it closer to the bottom) I can specify an offset:

image_crop(image, "100x300+0+100") |>
  plot()

And if I want it closer to the middle:

image_crop(image, "100x300+100+0") |>
  plot()

So how do you make sure you crop the right portion? Unfortunately I don’t have a perfect answer for you. You can do trial and error, or you can bring it into a photo editor and measure the dimensions and coordinates of your crop there. The things we do for reproducibility.

cropped <- image_crop(image, "370x100+90+200")
plot(cropped)

Now it just so happens that this dose response is backward - higher concentrations on the left, lower on the right. I prefer it the other way.

flopped <- image_flop(cropped)
plot(flopped)

Great - we’re ready to annotate!

Annotating

Now, back to webber.

To annotate blot lanes with webber, you need to construct a tibble with one row per lane, and one column per line of annotation.

# Why a tibble and not a regular data.frame?
# tibbles preserve spaces in their column names, which makes them display nicer.
annot <- 
  tibble(
    "Cell line" = "UC6",
    "Buparlisib (nM)" = c("DMSO", "1", "10", "100", "1000", "10000")
  )
annot
#> # A tibble: 6 × 2
#>   `Cell line` `Buparlisib (nM)`
#>   <chr>       <chr>            
#> 1 UC6         DMSO             
#> 2 UC6         1                
#> 3 UC6         10               
#> 4 UC6         100              
#> 5 UC6         1000             
#> 6 UC6         10000

Then supply the tibble to wb_annot_lanes:

wb <- wb_annot_lanes(flopped, annot, n_lanes = 6, text_size = 4)
plot(wb)