Latent periodic process inference from single-cell RNA-seq data

Abstract

Convoluted biological processes underlie the development of multicellular organisms and diseases. Advances in scRNA-seq make it possible to study these processes from cells at various developmental stages. Achieving accurate characterization is challenging, however, particularly for periodic processes, such as cell cycles. To address this, we developed Cyclum, a novel AutoEncoder approach that characterizes circular trajectories in the high-dimensional gene expression space. Cyclum substantially improves the accuracy and robustness of cell-cycle characterization beyond existing approaches. Applying Cyclum to removing cell-cycle effects leads to substantially improved delineations of cell subpopulations, which is useful for establishing various cell atlases and studying tumor heterogeneity. Cyclum is available at https://github.com/KChen-lab/cyclum. BackgroundConvoluted biological processes, which involve cell proliferation, differentiation, state transition, and cell-to-cell communication [1,2]. The course of development can be influenced by genetic (e.g., mutations), epigenetic, and environmental factors. Alterations to the genome, transcriptome, and proteome of individual cells also can result in pathogeneses [3]. Early efforts have been made to reconstruct the temporal ordering of biological samples using bulk data [4,5], although challenges associated with cellular heterogeneity make it difficult to infer accurate time series. Advances in single-cell RNA sequencing (scRNA-seq) enabled large-scale acquisition of singlecell transcriptomic profiles and provided an unprecedented opportunity to uncover latent biological processes that orchestrate dynamic expression of genes in single cells throughout the course of the development [6]. However, it is very challenging to deconvolute these processes from scRNA-seq data accurately. A sufficiently large number of cells across time, lineage, and space need to be sampled in order to capture detailed sub-populational features and reduce technological noise. Tremendous efforts have been made to develop trajectory inference methods from scRNA-seq data. Over 59 methods have been developed since 2014 [7], including the widely known Monocle and Wanderlust. These methods represent biological processes in linear, bifurcating, or other graph topologies.In many developmental processes, such as embryogenesis, organogenesis, and tumorigenesis [8], cell cycle plays a fundamental role. Distinct from processes that evoke linear changes in gene expression, cell cycle causes periodicity. A cycle starts from the G1 phase, goes through S and G2/M, and then returns to G1 within 24-hours for human cells [2]. This process is orchestrated elegantly by variable sets of genes (e.g., cyclins and cyclin-dependent kinases) that are turned on and off at relatively precise timings. As a result of such periodicity, the cycling cells at different transcriptomic states form a circular, non-linear trajectory in high-dimensional gene expression spaces. The positions of a cell alongside the circular trajec...

Temporal and spatial topography of cell proliferation in cancer

“…The time dynamics reconstruction by the ccD-CMD algorithm were found to be accurate, with 93% of cells being placed within 1% of their correct ordering along a canonical cell cycle trajectory (Figure S3H). When we compared the performance of ccD-CMD time inference with other published time inference algorithms (Cannoodt et al, 2016;Gut et al, 2015;Liang et al, 2020; . CC-BY-NC-ND 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.…”

“…We selected a subset of these time inference algorithms, processed our multiplexing imaging data and compared the results with the ccD-CMD representation and pseudotime ordering from the same datasets. We used the following three algorithms, all originally developed to process single cell RNA sequencing data: SCORPIUS (Cannoodt et al, 2016), Palantir (Setty et al, 2019) and Cyclum (Liang et al, 2020). We compared the ccD-CMD time inference output with SCORPIUS, Palantir and Cyclum on three exemplar CyCIF datasets from different experimental sources: 1) on tissue-based CyCIF data from a breast tumor tissue sample from Figures 3A-3F; 2) on plate-based CyCIF data from unperturbed MCF10a cells; 3) on synthetic data generated from a mathematical model of the mammalian cell cycle.…”

Cell cycle gene regulation dynamics revealed by RNA velocity and deep-learning

“…scRNA-seq provides a high-resolution approach to study the cell cycle without external perturbations, such as synchronization by drugs or engineered fluorescent reporters 3,4 . Many attempts to computationally assign cell cycle phases have been performed [5][6][7][8] , but they typically lack generalizability and fail in capturing the correct cell cycle dynamics. Thanks to the depth of the scRNA-seq datasets generated in this paper, cycling patterns in the unspliced-spliced RNA space for single genes (RNA velocity) can be observed clearly and exploited to naturally sort cells across the cell cycle 9 .…”

OmicPioneer-sc: an integrated, interactive visualization environment for single-cell sequencing data

“…To name one use-case, inclusion of PVMs can aid in the detective work required for identification and annotation of cell-types in the DRP and NGCHM. The ability to navigate interactively among cell, sample, feature, and pathway levels of information can also help address the complex molecular heterogeneity that often makes it challenging to isolate interesting cell phenotypes (Tirosh et al 2016;Izar et al 2020) or analyze cell functional status (Liang et al 2020) All-in-one analysis software, including Seurat (Butler et al 2018) and SimpleSingleCell in Bioconductor (Lun et al 2016), have been developed for processing and analysis of scRNA-seq data. However, they focus more on the data pre-processing and analysis steps (batch correction, differential expression analysis, clustering), and they require programming, typically in R. Unlike OmicPioneer-sc, however, they do not provide a navigable, dynamic environment for visualization of the data or for linking-out to external information resources.…”

“…To name one use-case, inclusion of PVMs can aid in the detective work required for identification and annotation of cell-types in the DRP and NGCHM. The ability to navigate interactively among cell, sample, feature, and pathway levels of information can also help address the complex molecular heterogeneity that often makes it challenging to isolate interesting cell phenotypes (Tirosh et al 2016; Izar et al 2020) or analyze cell functional status (Liang et al 2020)…”