Figure 1
From: Label-free macrophage phenotype classification using machine learning methods
Workflow of the macrophage intrinsic signal analysis. (a) Right block: six macrophage phenotypes were induced by applying different stimuli and macrophage polarization confirmed by measuring biological characteristics validated in accordance with previous studies. (a) Left block: a separate macrophage polarization experiment was used to investigate the autofluorescence signature for each phenotype. DAPI staining was used to exclude dead cells and autofluorescence of live cells was measured. (b) Label free polarized (DAPI-/live) macrophage samples were acquired in a Beckman Coulter CytoFLEX LX flow cytometer recording 45 parameters per cell = 20 fluorescence detector × 2 (height and area) + forward scatter 1 × 3 (height, area, width) + side scatter 1 × 2 (height and area). (c) The generated data was divided into three parts: training (60%), validation (20%), and testing (20%), where a supervised fully connected neural network (FCNN) classified the different phenotypes based on cell specific signatures.
