Large-scale surveys of single-cell gene expression have the potential to reveal rare cell populations and lineage relationships but require efficient methods for cell capture and mRNA sequencing1,2,3,4. Although cellular barcoding strategies allow parallel sequencing of single cells at ultra-low depths5, the limitations of shallow sequencing have not been investigated directly. By capturing 301 single cells from 11 populations using microfluidics and analyzing single-cell transcriptomes across downsampled sequencing depths, we demonstrate that shallow single-cell mRNA sequencing (∼50,000 reads per cell) is sufficient for unbiased cell-type classification and biomarker identification. In the developing cortex, we identify diverse cell types, including multiple progenitor and neuronal subtypes, and we identify EGR1 and FOS as previously unreported candidate targets of Notch signaling in human but not mouse radial glia. Our strategy establishes an efficient method for unbiased analysis and comparison of cell populations from heterogeneous tissue by microfluidic single-cell capture and low-coverage sequencing of many cells.
Access optionsAccess options
Subscribe to Journal
Get full journal access for 1 year
only $20.83 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Sequence Read Archive
NCBI Reference Sequence
We thank M. Bershteyn (University of California, San Francisco (UCSF)), E. Di Lullo (UCSF), C. Gertz (UCSF), C. McLean (23andMe), F. Ng (Fluidigm Corporation), J. Liu (UCSF), C. Lowe (Stanford University), M. Oldham (UCSF), A. Diaz (UCSF) and H. Retallack (UCSF) for reading the manuscript and for useful suggestions, S. Wang and Y. Wang for technical assistance, K. Davies (University of Oxford) for providing the antibody to syncoilin, A. Simeone (Institute of Genetics and Biophysics, Adriano Buzzati-Traverso) for providing the Emx2 probe, S. Yao and S. Ku (Allen Institute for Brain Research) for assistance in the neural progenitor induction protocol, G. Wang (Stanford University) for providing human induced pluripotent stem cells (hiPSCs) and the staff at the San Francisco General Hospital for providing access to donated fetal tissue. L.S., N.L. and M.A.U. are partially supported by California Institute for Regenerative Medicine Tools and Technologies II grant RT2-02052 to M.A.U., and the generation and maintenance of the hiPSCs and neural progenitor cells was also supported by RT2-02052 to M.A.U. A.A.P. is supported by a Damon Runyon Cancer Research Foundation postdoctoral fellowship (DRG-2166-13). This research was also supported by National Institute of Neurological Disorders and Stroke awards R01NS075998 and R01NS072630 to A.R.K.
Overview of cell types studied, sequencing depth, and alignment rates
Single cell capture efficiency data
Biological properties of PCA gene clusters
Comparison of the top 500 PCA genes identified in low- and high-coverage mRNA Seq data across 301 cells.
Top 500 PCA genes explaining variation across 65 neural cells
About this article
Nature Reviews Neuroscience (2019)