Variance Stabilizing Normalization (VSN)
App for VSN using Quarto
Support for VSN, another of those bioinformatics workhorses, as promised as “I'm thinking about adding…” in
is finally here at https://fuzzylife-vsn.share.connect.posit.cloud/
VSN is available at Bioconductor so all i had to do fire up Rstudio (going back felt like breath of fresh air!) and ask Posit Assistant to convert R-script using the package to even more “shiny” Quarto-notebook ready to be published in posit-cloud just with a click of “publish”!
All one need to do now is fire up the app and upload results like proteinGroups.txt, choose the columns to be VSN-ed (LFQ chosen by default) and click “Run VSN Normalization” (default parameters work fine but do check diagnostic plots!) then “Download VSN Normalized Data” should appear, this file containts the VSN-ed data in columns starting with “vsn_” and corresponding pow^2 transformed data in columns starting with “pow2_” which is “solution to my Laziness”/can be input into ani-Basic Differential Expression Analysis App (choose pow2_ columns) like
Back to basics
While MaxLFQ takes whole in one go, VSN is clever to normalize variance for both shift and scale, building upon the living legend in the field
I should derive it but i got exhausted deriving just this from the article abstract
Crazy how something i could do in few seconds in high school took me to exhaustion 🤪 I will blame it on the coding agent-driven attention deficit 😛 More details in the upcoming post 🤞
As usual, open-source/code is available as R-script
wget https://raw.githubusercontent.com/animesh/scripts/refs/heads/master/proteinGroupsVSN.r
wget "https://zenodo.org/records/14557756/files/proteinGroups.txt?download=1" -O proteinGroups.txt
Rscript proteinGroupsVSN.r proteinGroups.txt LFQAND now Quarto notebook
wget https://raw.githubusercontent.com/animesh/scripts/2574e10a84b90a3deab29b3180907ed3219fb18c/proteinGroupsVSN.r
quarto serve proteinGroupsVSN.qmd
Loading required namespace: shiny
processing file: proteinGroupsVSN.qmd
...
Browse at http://localhost:6490/So feel free to checkout, break and fix 😛