|Title||Utilizing biomarkers to forecast quantitative metabolite concentration profiles in human red blood cells|
|Publication Type||Conference Paper|
|Year of Publication||2017|
|Authors||Yurkovich JT, Yang L, Palsson BO|
|Conference Name||2017 IEEE Conference on Control Technology and Applications (CCTA)|
One of the major limitations in making experimental measurements of biological systems is the complexity of the network being investigated. Major efforts have been made to identify a subset of measurements (“biomarkers”) that can be used to provide information about the rest of the system. For red blood cells under cold storage conditions in a blood bank, a set of metabolite biomarkers have been identified that can reliably define the qualitative trend of cellular metabolism. Recently, it was shown that these biomarkers could also be used to train a model that quantitatively predicts the concentrations of other metabolites in the network over a 45 day time course. Here, we extend the utility of these methods by using a linear blackbox model to forecast future values of these concentrations. We show that 57 of the 70 metabolites measured in the red blood cell metabolic network (81%) can be accurately forecasted after 8 days of storage (5 time points) with a global median error of 18.36%. The ability to forecast metabolite profiles by only requiring a subset of measurements for the first few days of storage makes these methods immediately applicable in a clinical setting to assess the metabolic health of stored blood.
Utilizing biomarkers to forecast quantitative metabolite concentration profiles in human red blood cells