Comparison of clustering methods for high-dimensional single-cell flow and mass cytometry (CyTOF) data

This repository contains pre-processed data files to reproduce the analyses in our paper on comparing clustering methods for high-dimensional flow and mass cytometry (CyTOF) data. There are 6 main data sets, and 2 data sets from the FlowCAP-I challenges, as described in the paper. R scripts to reproduce all analyses and figures in the paper (including scripts to prepare the data files) are available from GitHub at: https://github.com/lmweber/cytometry-clustering-comparison. The final paper is published in Cytometry Part A, 89A: 1084-1096, available open access at: http://onlinelibrary.wiley.com/doi/10.1002/cyto.a.23030/full.
Notes:
For each of the 6 main data sets, there is one main .fcs file, plus an additional .fcs file labeled "notransform". The "notransform" files contain non-transformed data, where the standard arcsinh transform has not yet been applied. For the two FlowCAP data sets, non-transformed data is not available. Manually gated population labels are included as a column within the .fcs files. The "population_names" .txt files (attachments) provide meta-data of cell population names corresponding to each label (for the data sets with multiple populations of interest; see paper). Gating-ML 2.0 .xml files (attachments) provide the gating schemes for data sets Nilsson_rare and Mosmann_rare.