New method for data mining of metabolic biomarker

November 4, 2015

Finding hidden information of metabolic balance variations

RIKEN CSRS and RIKEN BioResource Center have jointly developed the Cluster-Aided MCR-ALS Method based on the multivariate curve resolution–alternating least squares (MCR-ALS) method. The new method focuses on component reliability rather than the number of calculation components for data mining of metabolic biomarker.

Results from this modified MCR-ALS method were confirmed as practically useful, based on analysis of NMR data from known standard mixtures. Additionally, analysis of urine and feces from aged and high-fat diet mice showed higher precision than conventional methods and a higher number of components with metabolic variation, reflecting and further supporting previously reported research.

Cluster-aided MCR-ALS allows the selection of reliable components even with little information, making it possible to find metabolic biomarkers resulting from hidden information at disease onset. It also provides opportunity to quantify and evaluate ambiguous information about health through the integration/analysis of routinely recorded data such as body weight and nutrition with information such as medical examination results, which will contribute to reexamination of social infrastructures and lifestyles for improved quality of life.

Original article

Scientific Reports doi: 10.1038/srep15710

H. Motegi, Y. Tsuboi, A. Saga, T. Kagami, M. Inoue, H. Toki, O. Minowa, T. Noda, J. Kikuchi,

"Identification of Reliable Components in Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS): a Data-Driven Approach across Metabolic Processes".

Contact

Jun Kikuchi; Team Leader

Yuuri Tsuboi; Technical Staff I

Environmental Metabolic Aanalysis Research Team