In the PCA data analysis section of the original manuscript [1
Y. H. Ong, M. Lim, and Q. Liu, “Comparison of principal component analysis and biochemical component analysis in Raman spectroscopy for the discrimination of apoptosis and necrosis in K562 leukemia cells,” Opt. Express 20(20), 22158–22171 (2012). [CrossRef] [PubMed]
], the 2D cell spectra matrix was transposed by mistake. This error resulted in inaccurate scores and loadings of the principal components and has affected the cell death classification accuracies based on PCA scores therefore the data analysis was repeated.
In this correction, the data dimension mistake was rectified and the data analysis algorithm has been improved. PCA was performed on the processed Raman spectra using a custom developed MATLAB code. In this code, the data is not centered by subtracting the mean values of cell spectra. Fig. 6
, Fig. 7
, and Table 1
in the original paper have been revised to show the corrected results. Minor changes to the discussion regarding the corrected results were presented after each figures and table below.
Fig. 6 (a) 2-D and (b) 3-D PCA plots show the separation of data based on different modes of cell death. The percent variance captured by each PC is shown in parenthesis along each axis in (b).
Fig. 7 The spectra of first three principal components in PCA, where (a) is PC 1, (b) PC 2 and (c) PC 3.
Table 1 Classification accuracies using two principal component scores obtained from PCA
|PC 1||PC 2||PC 3||PC 4||PC 5|
It was found that only the means of PC 2 scores show significant differences among the three groups of cells when Kruskal-Wallis one-way analysis of variance was performed on the scores on first ten principal components. Although PC 1 accounts for most variance in the data sets, it does not show significant differences across the groups in Kruskal-Wallis analysis. From Fig. 6(a)
, it was found that necrotic cells can be easily distinguished from live and apoptotic cells by the first pair of PC scores or the combination of the first and third PC scores. The main discriminant was the score of PC 1, where necrotic cells have positive scores while live and apoptotic cells have negative scores. For PC 2, all live cells have positive scores and most apoptotic cells have negative scores, while necrotic cells have scores around zero intersecting apoptotic and necrotic cells. The cells are inseparable by the scores of PC 3 as all of them have broad score distribution over the range of the third PC scores. There is not any pair in the first three PC scores that could be used to separate three groups of cells effectively. A three-dimensional PCA plots employing all the first three PC scores was then constructed as in Fig. 6(b)
and showed excellent separation of live, apoptotic and necrotic cells.
All the percentage accuracies cited in section 4.3 have been corrected as shown in Table 1
below. Since the changes in percentages are small, no text is revised.
The general message of the original paper is unaffected. The authors sincerely regret for these errors.