2019/02/21 by Cassandra Burdziak, Elham Azizi, Burdziak, Cassandra +6 · 6 citations
Biochemistry, Genetics and Molecular Biology · Computer Science · Mathematics · #Artificial intelligence #Bayesian Methods and Mixture Models #Biology #Cell #Cell type #Computational biology #Computer science #Data type #Epigenetics #FOS: Biological sciences #FOS: Computer and information sciences #Gene #Gene Regulatory Network Analysis #Gene expression #Gene expression and cancer classification #Gene regulatory network #Genetics #Genomics (q-bio.GN) #Inference #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Molecular Networks (q-bio.MN) #Quantitative Methods (q-bio.QM) #Regulation of gene expression #Single-cell and spatial transcriptomics #cs.LG #q-bio.GN #q-bio.MN #q-bio.QM #stat.ML
paper · pdf · doi:10.48550/arxiv.1902.08138
published in arXiv (Cornell University) (Cornell University)
arxiv created 2019/02/21 · openalex publication_date 2019/02/21 · arxiv updated 2019/02/22 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/06
We present a Bayesian hierarchical multi-view mixture model termed Symphony\nthat simultaneously learns clusters of cells representing cell types and their\nunderlying gene regulatory networks by integrating data from two views:\nsingle-cell gene expression data and paired epigenetic data, which is\ninformative of gene-gene interactions. This model improves interpretation of\nclusters as cell types with similar expression patterns as well as regulatory\nnetworks driving expression, by explaining gene-gene covariances with the\nbiological machinery regulating gene expression. We show the theoretical\nadvantages of the multi-view learning approach and present a Variational EM\ninference procedure. We demonstrate superior performance on both synthetic data\nand real genomic data with subtypes of peripheral blood cells compared to other\nmethods.\n