Table 1 Accordance with manual gating for GvHD data set. For HDPGMM we also report the result when components are merged according to our merging procedure. When this procedure is applied to the results obtained by ASPIRE, no components are merged, i.e. the original result is identical to what is obtained after mergeing
From: BayesFlow: latent modeling of flow cytometry cell populations
Method | F-measure | Precision | Recall |
|---|---|---|---|
BayesFlow run 1 | 0.91 (0.86, 0.95) | 0.96 | 0.89 |
BayesFlow run 2 | 0.87 (0.82, 0.92) | 0.95 | 0.84 |
ASPIRE | 0.67 (0.63, 0.72) | 0.86 | 0.63 |
HDPGMM | 0.35 (0.30, 0.39) | 0.98 | 0.23 |
HDPGMM merged | 0.60 (0.54, 0.66) | 0.95 | 0.48 |
flowMeans | 0.88 (0.82, 0.93) | 0.93 | 0.86 |
SamSPECTRAL | 0.87 (0.81, 0.93) | 0.96 | 0.83 |
Ensemble FlowCAP | 0.88 | Â | Â |