A0568
Title: MODE: High resolution digital dissociation with deep multimodal autoencoder
Authors: Qian Li - St. Jude Children\'s Research Hospital (United States) [presenting]
Abstract: In single-cell biology, the complexity of tissues may hinder lineage cell mapping or tumor microenvironment decomposition, requiring digital dissociation of bulk tissues. Many deconvolution methods focus on transcriptomic assay, which is not easily applicable to other omics due to ambiguous cell markers and reference-to-target difference. The aim is to present MODE, a multimodal autoencoder pipeline linking multi-dimensional features to jointly predict personalized multi-omic profiles and cellular compositions, using pseudo-bulk data constructed by internal non-transcriptomic reference and external scRNA-seq data. The pseudo-bulk training data was generated by Dirichlet and Poisson distributions, using parameters initialized from the target data with a joint nonnegative matrix factorization (JNMF) model. MODE was evaluated through rigorous simulation experiments and real multi-omic data from multiple tissue types, outperforming nine deconvolution pipelines with superior generalizability and fidelity.
