Implementing High Dimensional Reduction Analysis on Histocytometric Data

摘要

In a previous protocol article, we demonstrated construction of a histocytometry pipeline that is capable of both segmenting highly aggregated cell populations and retaining the original intensity data range of the input microscopy images. In the protocol presented here, using the output from the aforementioned article, we demonstrate how to phenotype the data using the high dimensional reduction analysis technique optimized t-distributed stochastic neighbor embedding (opt-t-SNE) and compare it to traditional manual gating. Additionally, we present a protocol illustrating the advantage of the inclusion of cell junction/membrane markers for accurately segmenting highly aggregated cell populations in ilastik.