Bioinformatics Seminar
Time: 11AM
Venue: Davis Auditorium and Online
27 October 2026
This is a WEHI only event.Analysing Olink proteomics data
Mengbo LiWEHI
Olink's proximity extension assays measure hundreds to thousands of proteins from a few microlitres of plasma, and are increasingly used in clinical and translational studies. On the latest platforms (Explore HT and Reveal), each protein measurement is a sequencing read count. Most analyses, however, start from NPX, a normalised value that Olink derives from those counts. In this talk I will explain how Olink data are generated, and how NPX is calculated from the counts using extension controls, plate controls and dilution blocks. I will then show that the counts behave much like RNA-seq counts, so they can be analysed with the edgeR and limma tools we use for RNA-seq. Using datasets from sepsis, Kawasaki disease and myeloproliferative neoplasms, I will compare a count-based analysis (voomLmFit) with the standard NPX analysis. On the same samples, the count-based approach finds substantially more differentially expressed proteins while keeping false discoveries under control. Simulations suggest the main gain comes from normalising each dilution block by its own library size, rather than by the extension control used in NPX.