Individualized co-expression-like index (iCKI) enables gene-gene interactions as individual biomarkers for complex disease.

[UNLABELLED] When a single gene exhibits an insignificant association with complex disease, applying gene–gene interactions as biomarkers may achieve striking findings. However, it is still a barrier to measuring the strength of gene–gene interaction at the individual level and further applying gene–gene pairs as biomarkers. To overcome this challenge, we introduce iCKI, namely individualized co-expression-like index, quantifying the interaction strength of a gene–gene pair for each individual. The higher the absolute value of iCKI, the stronger the co-expression of two biomarkers. iCKI makes co-expression variations as novel individual biomarkers possible, enabling advanced applications in disease diagnosis, survival analysis, and more. Applying iCKI to rheumatoid arthritis early prediction and pancreatic cancer survival analysis, we demonstrated that co-expression achieved substantial improvements compared to single biomarkers. Overall, iCKI offers an innovative and efficient indicator for considering gene–gene interactions as biomarkers and represents a starting point for individual co-expression.

[SUPPLEMENTARY INFORMATION] The online version contains supplementary material available at 10.1186/s12859-025-06225-x.