Methods for combining multiple correlated biomarkers with application to the study of low-grade inflammation and muscle mass in senior horses

The simplest analysis of biomarker data is based on a series of single biomarker hypothesis tests, followed by correction for multiple testing. However, it is intuitively plausible that a joint analysis of multiple biomarkers will have higher statistical power and promise improved discrimination over tests based on single markers. In this article, we study analytical properties of the approach for joint analysis of correlated summary statistics based on the test for quadratic forms (TQ). Based on the derivation of the TQ-distribution, we proposed a scale-location approximation of the TQ statistic, which we call approximate TQ. We show that the approximate TQ has very similar power to the traditional TQ test under varying correlation structures among biomarkers. Our application of both the TQ and the approximate TQ test to data on biomarkers for inflamm-aging – an age-related low-grade chronic inflammation – reveals an association between the percentage of IFNγ positive lymphocytes and overall muscle condition in senior horses.