Fig. 1: Flow cytometry variation across cytometers and days. | Communications Biology

Fig. 1: Flow cytometry variation across cytometers and days.

From: SwiftReg cluster registration automatically reduces flow cytometry data variability including batch effects

Fig. 1: Flow cytometry variation across cytometers and days.

From JMW090 and JMW092, a SWIFT cluster template containing 104 clusters was derived from a concatenate of influenza peptide-stimulated and negative control samples from one subject across eight bleeds. Cells from influenza-stimulated samples from 6 subjects, 8 bleeds each, were assigned to the cluster template. a Heatmap representation of low (blue) to high (yellow) Pearson correlations of cluster centroid values between all pairs of samples. The channels shown are CD45RA and CD4, with rows/columns sorted by assay day, cytometer, subject, and bleed, i.e., the smallest squares represent the eight bleeds. b QC plot: each dot within a channel row represents one cluster from one sample. A cluster’s vertical position within the channel row is the log ratio of its MFI value to the reference standard MFI value in that channel. The reference standard MFI values for each cluster were the averages of the MFI values for that cluster in all eight samples from subject 5. The horizontal lines indicate a 1:1 log ratio in each channel. Each column represents one sample from bleed 1 (ordered as in a). Dot colors are varied for ease of viewing between channels. c Full QC plot: same plot style as in b but with all channels and samples (ordered as in a); the small scale is intended for rapid qualitative assessment of the variability between samples.

Back to article page