Large-scale proteogenomic meta-analyses highlight N-linked glycosylation as important regulator of proteins in the blood

What the largest study of its kind tells us about health, disease, and new medicines

Blood carries thousands of proteins, tiny working molecules that do jobs like fighting infections, carrying signals between organs, and helping with clotting. Doctors already use a handful of blood proteins as tests (for example, markers of inflammation), but for many proteins we still don’t know what drives their levels up or down in different people.

In a worldwide project from the SCALLOP consortium, researchers analysed 1,161 blood proteins in nearly 80,000 participants across 38 cohorts (including Viking Genes). They looked for DNA differences linked to protein levels. These links are called “protein quantitative trait loci”, usually shortened to pQTLs, and you can think of them as genetic “dials” that nudge a protein’s level higher or lower. The team found 24,738 independent pQTL signals, including more than 4,000 that had not been reported before.

The study did more than count these genetic signals. It also asked how they work, and what they might mean for disease research. Some signals sit close to the gene for the protein they affect, these are called cis signals (think “nearby control”). Others act from far away in the genome, these are trans signals (more like “remote control”), often working through wider biological pathways.

Proteogenomic study cohorts

Figuring out what ‘remote control’ (trans) signals are really doing

Trans signals are powerful, but they are harder to interpret: if a DNA variant sits far from the protein’s own gene, it’s hard to tell which gene is the variant acting through. To tackle this, the researchers used a computer‑based scoring approach that pulls together several kinds of clues, for example, whether genes are known to interact in the same pathway, whether the proteins physically bind each other, whether rare genetic changes in a gene affect the same protein, and where in the body the gene is most active.

Using this approach, the team could suggest at least one likely “effector” gene: the gene doing the work. In about two‑thirds of the locations they studied, the evidence pointed to one clear best‑candidate gene. When the researchers looked for common themes among these effector genes, one stood out strongly: a type of glycosylation (that is adding sugars to proteins). This pathway showed up again and again as a likely driver of protein differences, appearing as a trans regulator for 143 proteins. In plain terms: for many proteins, differences in sugar‑adding machinery may be one of the main reasons people have different levels in their blood.

Hint at new uses for existing drugs

One striking example centred on a protein‑changing variant in the TYK2 gene, a gene involved in immune signalling. This variant was linked remotely to levels of several messengers of the immune system, including CXCL9, CXCL10, CXCL11, and PDCD1. Higher levels of these proteins were associated with higher risk of autoimmune conditions such as rheumatoid arthritis, hypothyroidism, and psoriasis, and the genetic evidence suggested these links share the same underlying signal at the TYK2 region.

Because medicines that inhibit TYK2 already exist, this kind of genetic pattern can help researchers spot where an existing drug might work in a new disease. It also suggests that groups of proteins influenced by a trans signal could one day help identify patients most likely to benefit from a particular treatment.

This paper was published in Cell. To read more visit:

Multi-cohort proteogenomic analyses reveal genetic effects across the proteome and diseasome