Fig. 7: Visualization of data generation and processing.
From: AI-QuIC machine learning for automated detection of misfolded proteins in seed amplification assays
Each reaction is obtained from the RT-QuIC analysis of CWD positive/negative samples obtained through the experimental design process outlined in the Milstein, Gresch, et. al. paper43. The fluorescence readings from this analysis are then processed into summarized metrics to be used as a comparative dataset. Both datasets are split into a training and testing set. These datasets are used respectively to train and test the models, with the testing data providing the basis for evaluation.
