For each method, data analysis was restricted to i) probes for which the detection score was greater than 95% in all samples or ii) probes with SD > 0.5 in all samples. We compared the number of detected probes between methods by using a McNemar test. For investigation of the biological reproducibility and the concordance between methods, we calculated spearman correlations between visits for each probe for each method. To compare biological reproducibility between two methods, we averaged the expression values of each probes across visits and calculated spearman correlations between methods.
