Scalable optimal Bayesian classification of single-cell trajectories under regulatory model uncertainty

• Published in: BMC genomics Vol. 20; no. Suppl 6; pp. 435 - 11
• Main Authors: Hajiramezanali, Ehsan, Imani, Mahdi, Braga-Neto, Ulisses, Qian, Xiaoning, Dougherty, Edward R.
• Format: Journal Article
• Language: English
• Published: London BioMed Central 13.06.2019; BioMed Central Ltd; Springer Nature B.V; Springer; BMC
• Subjects: Accounting; Algorithms; Analysis; Animal Genetics and Genomics; BASIC BIOLOGICAL SCIENCES; Bayes Theorem; Bayesian analysis; Biochemistry; Bioinformatics; Biomedical and Life Sciences; Biotechnology & Applied Microbiology; Boolean algebra; Cancer; Cancer genetics; Classification; Complexity; Computational Biology - methods; Computer applications; Computer memory; Dynamical systems; Gene expression; Gene Expression Profiling; Gene Regulatory Networks; Genes; Genetic regulation; Genetics & Heredity; Genomics; Humans; Identification and classification; Leukemia; Leukemia, Large Granular Lymphocytic - genetics; Life Sciences; Lymphocytes; Lymphocytes T; Methods; Microarrays; Microbial Genetics and Genomics; Models, Biological; Models, Genetic; Noise; Optimal Bayesian classification; Parameter estimation; Particle filter; Plant Genetics and Genomics; Probabilistic Boolean networks; Proteomics; Signal processing; Single-Cell Analysis - methods; Single-cell trajectory classification; T cells; Trajectories; Uncertainty; United States; Optimal Bayesian classification; Probabilistic Boolean networks; Single-cell trajectory classification; Particle filter
• ISSN: 1471-2164; 1471-2164
• DOI: 10.1186/s12864-019-5720-3

Background Single-cell gene expression measurements offer opportunities in deriving mechanistic understanding of complex diseases, including cancer. However, due to the complex regulatory machinery of the cell, gene regulatory network (GRN) model inference based on such data still manifests significant uncertainty. Results The goal of this paper is to develop optimal classification of single-cell trajectories accounting for potential model uncertainty. Partially-observed Boolean dynamical systems (POBDS) are used for modeling gene regulatory networks observed through noisy gene-expression data. We derive the exact optimal Bayesian classifier (OBC) for binary classification of single-cell trajectories. The application of the OBC becomes impractical for large GRNs, due to computational and memory requirements. To address this, we introduce a particle-based single-cell classification method that is highly scalable for large GRNs with much lower complexity than the optimal solution. Conclusion The performance of the proposed particle-based method is demonstrated through numerical experiments using a POBDS model of the well-known T-cell large granular lymphocyte (T-LGL) leukemia network with noisy time-series gene-expression data.