Blood Gene Expression Fluctuates More Than Expected Over Time

Blood-based gene expression is widely used to explore disease biology, but temporal variability can complicate interpretation of single time-point measurements. Seasonal shifts, time of day, and subclinical infections can all alter transcriptional readouts within the same individual, making it harder to distinguish true disease signals from ordinary physiological fluctuation. This remains a central challenge for laboratory medicine and biomarker discovery. New findings demonstrate that time-linked within-person changes dominate most blood gene expression and should be systematically considered in study design.

Kiel University’s Excellence Cluster PMI, together with KU Leuven and the German Center for Neurodegenerative Diseases (DZNE) in Bonn, analyzed longitudinal blood transcriptomes to quantify temporal effects. Investigators tracked 333 volunteers in Flanders across six months, drawing blood three times and measuring the activity of roughly 14,000 genes at each visit. The observations were tested in two independent resources: a British twin study and a cross-sectional cohort of 3,480 DZNE participants.

The analysis found that in 85% of genes, within-person variation over time exceeded variation between different people. More than 4,000 genes shifted activity seasonally, with immune signaling predominating in winter and genes linked to the biological clock prevailing in summer. Time of day and minor infections also contributed to fluctuation, and alternative splicing emerged as an underrecognized layer of variation.

A subset of genes, about 15%, remained notably stable within individuals but varied more strongly across individuals, acting as a molecular “fingerprint.” These genes were enriched for immune pathways tied to T and B cells. Sex differences were also prominent: women showed higher temporal variability than men, with variability diminishing after menopause but largely persisting, while men were more stable within themselves yet differed more from one another. The study indicates that some signals previously attributed to disease may instead reflect ordinary physiological dynamics.

The authors note practical implications for laboratory and clinical studies, including prioritizing repeated sampling over single snapshots, accounting for blood draw timing, and re-evaluating genes flagged as putative markers in areas such as cardiovascular disease, neurodegeneration, and chronic inflammatory conditions such as IBD. The study was published in Nature Communications under the title "Large-scale analysis of temporal gene expression variation in peripheral blood".

“When we compare blood samples, we usually think in categories: sick versus healthy, this person versus that one. What we see here is that the greatest biological distance often lies not between two people, but between the same person at different points in time,” said Professor Philip Rosenstiel of the Institute for Clinical Molecular Biology (IKMB) at CAU and UKSH, Kiel Campus.

“This doesn't mean previous research is wrong. But it does mean we need to take the temporal dimension far more seriously—and treat sex, season and time of day as biological variables, not footnotes,” said Professor Jeroen Raes of KU Leuven and VIB, the study lead for the Flemish gut flora project.

Related Links
Kiel University
KU Leuven
DZNE