A unified modeling framework for metabonomic profile development and covariate selection for acute trauma subjects

Abstract

SUMMARY Acute trauma is often associated with progressive deterioration of multiple organ systems in humans and is the leading cause of death in trauma care units. Identification of specific organ failure in non-invasive fashion and the contribution of different demographic factors on the casual progression of acute trauma is of supreme interest for successful diagnosis, prognosis or monitoring of trauma status. Recently electrospray ionization (ESI) and matrix assisted laser desorption (MALDI) time-of-flight (TOF) mass spectrometry have been used to identify biomarkers both in Proteomics and Metabonomics studies. Data sets generated from mass spectrometers in such studies are generally very large in size and thus require the use of sophisticated statistical techniques to glean useful information. In a recent development Ghosh et al. [1] suggested a unified semi-parametric approach to distinguish urinary metabolic profiles in a group of traumatic subjects from those of a control group consisting of normal individuals. In this study we have extended their approach by combining available covariate information in the development of metabonomic profile of acute trauma. We have shown that age is a statistically significant covariate across trauma and control group, thus pointing out the fact that prognosis of trauma may be acutely linked with subjects age.

Complexity in Systems Level Biology and Genetics: Statistical Perspectives

“…The reconstruction of NMR spectra is a considerably more challenging area than those for which mixture modelling is conventionally used, as many more individual components are required. Flexible semi-parametric mixture models have been utilized in [75,76], whilst fully non-parametric mixture models similar to those described in section §2.5.2 can also be used [73].…”