A Ali Heydari1,2, Oscar A Davalos3,2, Lihong Zhao1, Katrina K Hoyer4,2, Suzanne S Sindi1,2. 1. Department of Applied Mathematics, University of California, Merced, 95343, USA. 2. Health Sciences Research Institute, University of California, Merced, 95343, USA. 3. Quantitative and Systems Biology Graduate Program, University of California Merced, 95343, USA. 4. Department of Molecular and Cell Biology, University of California, Merced, 95343, USA.
Abstract
MOTIVATION: Single-cell RNA sequencing (scRNAseq) technologies allow for measurements of gene expression at a single-cell resolution. This provides researchers with a tremendous advantage for detecting heterogeneity, delineating cellular maps or identifying rare subpopulations. However, a critical complication remains: the low number of single-cell observations due to limitations by rarity of subpopulation, tissue degradation or cost. This absence of sufficient data may cause inaccuracy or irreproducibility of downstream analysis. In this work, we present ACTIVA (Automated Cell-Type-informed Introspective Variational Autoencoder): a novel framework for generating realistic synthetic data using a single-stream adversarial variational autoencoder conditioned with cell-type information. Within a single framework, ACTIVA can enlarge existing datasets and generate specific subpopulations on demand, as opposed to two separate models (such as scGAN and cscGAN). Data generation and augmentation with ACTIVA can enhance scRNAseq pipelines and analysis, such as benchmarking new algorithms, studying the accuracy of classifiers and detecting marker genes. ACTIVA will facilitate analysis of smaller datasets, potentially reducing the number of patients and animals necessary in initial studies. RESULTS: We train and evaluate models on multiple public scRNAseq datasets. In comparison to GAN-based models (scGAN and cscGAN), we demonstrate that ACTIVA generates cells that are more realistic and harder for classifiers to identify as synthetic which also have better pair-wise correlation between genes. Data augmentation with ACTIVA significantly improves classification of rare subtypes (more than 45% improvement compared to not augmenting and 4% better than cscGAN) all while reducing run-time by an order of magnitude in comparison to both models. AVAILABILITY OF DATA AND CODE: The codes and datasets are hosted on Zenodo (https://doi.org/10.5281/zenodo.5879639). Tutorials are available at https://github.com/SindiLab/ACTIVA. SUPPLEMENTARY INFORMATION: Supplementary material is available online.
MOTIVATION: Single-cell RNA sequencing (scRNAseq) technologies allow for measurements of gene expression at a single-cell resolution. This provides researchers with a tremendous advantage for detecting heterogeneity, delineating cellular maps or identifying rare subpopulations. However, a critical complication remains: the low number of single-cell observations due to limitations by rarity of subpopulation, tissue degradation or cost. This absence of sufficient data may cause inaccuracy or irreproducibility of downstream analysis. In this work, we present ACTIVA (Automated Cell-Type-informed Introspective Variational Autoencoder): a novel framework for generating realistic synthetic data using a single-stream adversarial variational autoencoder conditioned with cell-type information. Within a single framework, ACTIVA can enlarge existing datasets and generate specific subpopulations on demand, as opposed to two separate models (such as scGAN and cscGAN). Data generation and augmentation with ACTIVA can enhance scRNAseq pipelines and analysis, such as benchmarking new algorithms, studying the accuracy of classifiers and detecting marker genes. ACTIVA will facilitate analysis of smaller datasets, potentially reducing the number of patients and animals necessary in initial studies. RESULTS: We train and evaluate models on multiple public scRNAseq datasets. In comparison to GAN-based models (scGAN and cscGAN), we demonstrate that ACTIVA generates cells that are more realistic and harder for classifiers to identify as synthetic which also have better pair-wise correlation between genes. Data augmentation with ACTIVA significantly improves classification of rare subtypes (more than 45% improvement compared to not augmenting and 4% better than cscGAN) all while reducing run-time by an order of magnitude in comparison to both models. AVAILABILITY OF DATA AND CODE: The codes and datasets are hosted on Zenodo (https://doi.org/10.5281/zenodo.5879639). Tutorials are available at https://github.com/SindiLab/ACTIVA. SUPPLEMENTARY INFORMATION: Supplementary material is available online.
Authors: Mohamed Marouf; Pierre Machart; Vikas Bansal; Christoph Kilian; Daniel S Magruder; Christian F Krebs; Stefan Bonn Journal: Nat Commun Date: 2020-01-09 Impact factor: 14.919
Authors: Aviv Regev; Sarah A Teichmann; Eric S Lander; Ido Amit; Christophe Benoist; Ewan Birney; Bernd Bodenmiller; Peter Campbell; Piero Carninci; Menna Clatworthy; Hans Clevers; Bart Deplancke; Ian Dunham; James Eberwine; Roland Eils; Wolfgang Enard; Andrew Farmer; Lars Fugger; Berthold Göttgens; Nir Hacohen; Muzlifah Haniffa; Martin Hemberg; Seung Kim; Paul Klenerman; Arnold Kriegstein; Ed Lein; Sten Linnarsson; Emma Lundberg; Joakim Lundeberg; Partha Majumder; John C Marioni; Miriam Merad; Musa Mhlanga; Martijn Nawijn; Mihai Netea; Garry Nolan; Dana Pe'er; Anthony Phillipakis; Chris P Ponting; Stephen Quake; Wolf Reik; Orit Rozenblatt-Rosen; Joshua Sanes; Rahul Satija; Ton N Schumacher; Alex Shalek; Ehud Shapiro; Padmanee Sharma; Jay W Shin; Oliver Stegle; Michael Stratton; Michael J T Stubbington; Fabian J Theis; Matthias Uhlen; Alexander van Oudenaarden; Allon Wagner; Fiona Watt; Jonathan Weissman; Barbara Wold; Ramnik Xavier; Nir Yosef Journal: Elife Date: 2017-12-05 Impact factor: 8.140