关键词: 核磁共振, 代谢组学, 自适应分段积分, 统计差异性, 数据预处理

Abstract:

A novel adaptive binning method was proposed for NMR metabonomic data preprocessing. The statistical discrepancy of each spectral data point is estimated, then the contiguous data points are integrated adaptively based on the statistical discrepancy. Comparing to the fixed width binning, the proposed method can overcome the following negative effects on the subsequently statistical analysis. For example, signals with opposite statistical discrepancies may be superposition in a same region. Both simulated NMR data and experimental spectra from dietary intervention individuals were employed to validate the performance of the adaptive binning. The results show that the proposed method effectively mitigates disturbance from spectral noises and signals without statistical significance. It can increase the interpretability of PCA loading results so that the metabonomics results are more biological significant.

Key words: NMR spectroscopy, Metabonomics, Adaptive binning, Statistical discrepancy, Data preproces-sing