Abstract
To understand the molecular pathogenesis of human disease, precision analyses to define alterations within and between disease-associated cell populations are desperately needed. Single-cell genomics represents an ideal platform to enable the identification and comparison of normal and diseased transcriptional cell populations. We created cellHarmony, an integrated solution for the unsupervised analysis, classification, and comparison of cell types from diverse single-cell RNA-Seq datasets. cellHarmony efficiently and accurately matches single-cell transcriptomes using a community-clustering and alignment strategy to compute differences in cell-type specific gene expression over potentially dozens of cell populations. Such transcriptional differences are used to automatically identify distinct and shared gene programs among cell-types and identify impacted pathways and transcriptional regulatory networks to understand the impact of perturbations at a systems level. cellHarmony is implemented as a python package and as an integrated workflow within the software AltAnalyze. We demonstrate that cellHarmony has improved or equivalent performance to alternative label projection methods, is able to identify the likely cellular origins of malignant states, stratify patients into clinical disease subtypes from identified gene programs, resolve discrete disease networks impacting specific cell-types, and illuminate therapeutic mechanisms. Thus, this approach holds tremendous promise in revealing the molecular and cellular origins of complex disease.
Highlights
Single-cell RNA-sequencing provides the unique ability to profile transcripts from diverse cell populations along a continuum of related or disparate states (Olsson et al 2016)
The software implements a unique comparison and aggregation approach to identify impacted gene and regulatory programs that may be restricted to individual cell-states, shared more broadly between specific states, or that are common to all populations (Fig. 1C-F)
The approach remains scalable to large datasets, due to its low memory footprint, and will be less sensitive to outlier effects, through the use of neighborhood matching. While this method can accurately align diverse cells from different tissues and across different single-cell technologies, its principle power is in the determination of global, regional, and local differences among cell-states at the transcriptional, pathway, and gene-network level. As demonstrated, such insights include the localization of critical disease transcriptional changes to specific cell types, improved understanding of the specificity of drug targets to specific cell types in disease, improved diagnostic biomarkers, and novel regulatory and signaling networks that can inform therapy
Summary
Single-cell RNA-sequencing (scRNA-Seq) provides the unique ability to profile transcripts from diverse cell populations along a continuum of related or disparate states (Olsson et al 2016). In addition to defining known and novel cell populations, single-cell technologies can identify disease-related gene regulatory programs which underlie molecular and cellular dysfunction. While diverse single-cell experimental platforms exist to facilitate such analyses, there is an urgent need for integrated and easy-touse computational approaches to identify discrete differences between comparable diseased and healthy cells. Given that most tissue scRNA-Seq analyses will potentially identify dozens of cell populations, such an exercise becomes non-trivial, as distinct cell populations will have different transcriptional, cellular, pathway, and gene regulatory network impacts. Cellular and molecular differences can occur in either a cell-state-specific manner or across a spectrum of related cell states, requiring new holistic analysis solutions. Given the complexity of the analyses required to achieve these goals, automated solutions that can be applied by both experienced bioinformaticians and conventional laboratory biologists are required
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
