Fig. 2: Datasets description and overview of batch effect correction. | Nature Communications

Fig. 2: Datasets description and overview of batch effect correction.

From: BERNN: Enhancing classification of Liquid Chromatography Mass Spectrometry data with batch effect removal neural networks

Fig. 2

A Table summarizing the five datasets used in the study. For each dataset, 2 conditions are classified by BERNNs. B PCA visualization of the raw data for five datasets. The datasets are, from left to right, ordered by the strength of the initial batch effect. For each dataset, the middle row of images represents the transformation resulting in the best valid MCC. The images in the last row are from representations that were in the top methods purely for batch correction but performed badly for classification. The PCA visualization of the two datasets not represented here is available in Supplementary Fig. 13.

Back to article page