2019/07/18 by Eustasio del Barrio, Hristo Inouzhe, del Barrio, Eustasio +7
Biochemistry, Genetics and Molecular Biology · Computer Science · #Single-cell and spatial transcriptomics #Gene expression and cancer classification #Bayesian Methods and Mixture Models
paper · pdf · doi:10.48550/arxiv.1907.08006
Data obtained from Flow Cytometry present pronounced variability due to\nbiological and technical reasons. Biological variability is a well-known\nphenomenon produced by measurements on different individuals, with different\ncharacteristics such as illness, age, sex, etc. The use of different settings\nfor measurement, the variation of the conditions during experiments and the\ndifferent types of flow cytometers are some of the technical causes of\nvariability. This mixture of sources of variability makes the use of supervised\nmachine learning for identification of cell populations difficult. The present\nwork is conceived as a combination of strategies to facilitate the task of\nsupervised gating.\n We propose optimalFlowTemplates, based on a similarity distance and\n\Wasserstein barycenters, which clusters cytometries and produces\nprototype cytometries for the different groups. We show that supervised\nlearning, restricted to the new groups, performs better than the same\ntechniques applied to the whole collection. We also present\noptimalFlowClassification, which uses a database of gated cytometries and\noptimalFlowTemplates to assign cell types to a new cytometry. We show that this\nprocedure can outperform state of the art techniques in the proposed datasets.\nOur code is freely available as optimalFlow a Bioconductor R package at\nhttps://bioconductor.org/packages/optimalFlow.\n optimalFlowTemplates+optimalFlowClassification addresses the problem of using\nsupervised learning while accounting for biological and technical variability.\nOur methodology provides a robust automated gating workflow that handles the\nintrinsic variability of flow cytometry data well. Our main innovation is the\nmethodology itself and the optimal-transport techniques that we apply to flow\ncytometry analysis.\n