Figure 5

Comparison of spectral quality before and after low quality spectra are detected by the correlation matrix and discarded. The quality of spectra from biological replicates of quality control samples is illustrated in the Pearson correlation matrix surfaces that were generated using the software DMVS 2.0 (Shenzhen Chipscreen Biosciences, Ltd, Shenzhen, P. R. China, http://www.chipscreen.com). The x and y axes represent spectra and the z-axes indicate the values of Pearson correlation coefficients. The bright red color indicates very good correlation with values of R close to 1, and the green color indicates that a low correlation was observed between two spectra. (A) Correlation matrix was applied to the dataset prior to quality control. Green valleys caused by several low quality spectra that had very low correlations to most of other spectra were detected. (B) After the low quality spectra were detected by the correlation matrix and discarded, the Pearson correlation matrix surface became clear and very red, indicating that the majority of the correlations among all spectra are very high.