Recovering Gene Interactions from Single-Cell Data Using Data Diffusion

Abstract

<p class="first" id="P5">Single-cell RNA-sequencing technologies suffer from many sources of technical noise, including under-sampling of mRNA molecules, often termed ‘dropout’, which can severely obscure important gene-gene relationships. To address this, we developed MAGIC (Markov Affinity-based Graph Imputation of Cells), a method that shares information across similar cells, via data diffusion, to denoise the cell count matrix and fill in missing transcripts. We validate MAGIC on several biological systems and find it effective at recovering gene-gene relationships and additional structures. MAGIC reveals a phenotypic continuum, with the majority of cells residing in intermediate states that display stem-like signatures and uncovers known and previously uncharacterized regulatory interactions, demonstrating that our approach can successfully uncover regulatory relations without perturbations. </p><p class="first" id="P6"> <b>One Sentence Summary:</b> Graph diffusion-based imputation method recovers missing transcripts in scRNA-seq data, yielding insight into the epithelial-to-mesenchymal transition. </p><p id="P7">Abstract highlights:</p><p id="P8">1. MAGIC restores noisy and sparse single-cell data using diffusion geometry.</p><p id="P9">2. Corrected data is amenable to myriad downstream analyses.</p><p id="P10">3. MAGIC enables archetypal analysis and inference of gene interactions.</p><p id="P11">4. Transcription factor targets can be predicted without perturbation after MAGIC. In brief - A new algorithm overcomes limitations of data loss in single cell sequencing experiments </p><p id="P12"> <div class="figure-container so-text-align-c"> <img alt="" class="figure" src="/document_file/2b76c424-d481-41af-bcef-3dca4d1d1c24/PubMedCentral/image/nihms-977470-f0001.jpg"/> </div> </p>