Table 1. Effect of preprocessing steps on prediction accuracy.

	Original signal	+	Low filtering	+	Spatial detrending	+	Quantile normalization
Pearson	0.603		0.603		0.607		0.624
Spearman	0.618		0.620		0.623		0.624

This table shows the effect that different preprocessing steps have on the prediction accuracy of our model. The Pearson and Spearman correlations are averaged over all 86 PBM sample pairs. We used samples from the HK array as our training samples.