Massively parallel polymerase cloning and genome sequencing of single cells using nanoliter microwells - PubMed (original) (raw)

. 2013 Dec;31(12):1126-32.

doi: 10.1038/nbt.2720. Epub 2013 Nov 10.

Affiliations

Massively parallel polymerase cloning and genome sequencing of single cells using nanoliter microwells

Jeff Gole et al. Nat Biotechnol. 2013 Dec.

Abstract

Genome sequencing of single cells has a variety of applications, including characterizing difficult-to-culture microorganisms and identifying somatic mutations in single cells from mammalian tissues. A major hurdle in this process is the bias in amplifying the genetic material from a single cell, a procedure known as polymerase cloning. Here we describe the microwell displacement amplification system (MIDAS), a massively parallel polymerase cloning method in which single cells are randomly distributed into hundreds to thousands of nanoliter wells and their genetic material is simultaneously amplified for shotgun sequencing. MIDAS reduces amplification bias because polymerase cloning occurs in physically separated, nanoliter-scale reactors, facilitating the de novo assembly of near-complete microbial genomes from single Escherichia coli cells. In addition, MIDAS allowed us to detect single-copy number changes in primary human adult neurons at 1- to 2-Mb resolution. MIDAS can potentially further the characterization of genomic diversity in many heterogeneous cell populations.

PubMed Disclaimer

Figures

Figure 1

Figure 1

Microwell displacement amplification system. (a) Each slide contains 16 arrays of 255 microwells each. Cells, lysis solution, denaturing buffer, neutralization buffer and MDA master mix were each added to the microwells with a single pipette pump. Amplicon growth was then visualized with a fluorescent microscope using a real-time MDA system. Microwells showing increasing fluorescence over time were positive amplicons. The amplicons were extracted with fine glass pipettes attached to a micromanipulation system. (b) Scanning electron microscopy of a single E. coli cell displayed at different magnifications. This particular well contains only one cell, and most wells observed also contained no more than one cell. (c) A custom microscope incubation chamber was used for real time MDA. The chamber was temperature and humidity controlled to mitigate evaporation of reagents. Additionally, it prevented contamination during amplicon extraction by self-containing the micromanipulation system. An image of the entire microwell array is also shown, as well as a micropipette probing a well. (d) Complex three-dimensional MDA amplicons were reduced to linear DNA using DNA polymerase I and Ampligase. This process substantially improved the complexity of the library during sequencing.

Figure 2

Figure 2

Depth of coverage of assembled contigs aligned to the reference E. coli genome. Three single E. coli cells were analyzed using MIDAS. Between 88% and 94% of the genome was assembled from 2–8M paired-end 100bp reads. Each colored circle is a histogram of the log2 of average depth of coverage across each assembled contig for one cell. Gaps are represented by blank whitespace in between colored contigs

Figure 3

Figure 3

Genomic coverage of single bacterial (a,b) and mammalian (c,d) cells amplified by MDA in a tube and by MIDAS. The observed multi-peak profile for the MDA reactions implies that certain regions may have been amplified with exponentially greater bias compared to the majority of the genome. (a) Comparison of single E. coli cells amplified in a PCR tube for 10 hours (top), 2 hours (middle) and in a microwell (MIDAS) for 10 hours (bottom). Genomic positions were consolidated into 1 kb bins (x-axis), and were plotted against the log10 ratio (y-axis) of genomic coverage (normalized to the mean). (b) Distribution of coverage of amplified single bacterial cells. The x-axis shows the log10 ratio of genomic coverage normalized to the mean. (c) Comparison of single human cells amplified using traditional MDA in a PCR tube for 10 hours (top) or in a microwell (MIDAS) for 10 hours (middle) to a pool of unamplified human cells (bottom). Genomic positions were consolidated into variable bins of approximately 60 kb in size previously determined to contain a similar read count, and were plotted against the log10 ratio (y-axis) of genomic coverage (normalized to the mean). (d) Distribution of coverage of amplified single mammalian cells. The x-axis shows the log10 ratio of genomic coverage normalized to the mean.

Figure 4

Figure 4

Detection of copy number variants using MIDAS (a,c) and in-tube MDA (b,d). Genomic positions were consolidated into bins of approximately 60 kb in size which were previously determined to contain a similar read count. Estimated copy numbers below were rounded to the nearest whole number. (a) Copy number variation in a Down Syndrome single cell analyzed with MIDAS. The x-axis shows genomic position, while the y-axis shows (on a log2 scale) the estimated copy number as a red line. The arrow indicates trisomy 21, which is clearly visible in this single cell. (b) Copy number variation in a Down Syndrome single cell analyzed with traditional in-tube MDA. The x-axis shows genomic position, while the y-axis shows (in a log2 scale) the estimated copy number as a red line. The arrow marks the expected region of Trisomy 21, which is not detectable in this data. (c) Copy number variation in a Down Syndrome single cell with Trisomy 21 “spike-ins.” The x-axis shows genomic position, while the y-axis shows (in a log2 scale) the estimated copy number as a red line. At each arrow, prior to CNV calling, data from a randomly determined 2 Mb section of Trisomy chromosome 21 was computationally inserted into the genome, simulating a small gain of single copy event. At each location, a copy number variant was called, showing that MIDAS can detect 2 Mb copy number variation accurately. (d) Copy number variation in a Down Syndrome single cell with Trisomy 21 “spike-ins.” The x-axis shows genomic position, while the y-axis shows (on a log2 scale) the estimated copy number as a red line. At each arrow, prior to CNV calling, data from a randomly determined 2 Mb section of Trisomy chromosome 21 was computationally inserted into the genome, simulating a small gain of single copy event.

Figure 5

Figure 5

Comparison of MIDAS to previously published data for in-tube MDA, microfluidic MDA and MALBAC.for diploid regions of pools of two sperm cells and diploid regions of a single SW480 cancer cell processed using MALBAC. Genomic positions were consolidated into variable bins of approximately 60 kb in size previously determined to contain a similar read count, and were plotted against the log10 ratio (y-axis) of genomic coverage (normalized to the mean). For the cancer cell data, non-diploid regions have been masked out (white gaps between pink) to remove the bias generated by comparing a highly aneuploid cell to a primarily diploid cell.

Similar articles

Cited by

References

    1. Zhang K, et al. Sequencing genomes from single cells by polymerase cloning. Nat Biotechnol. 2006;24:680–686. - PubMed
    1. Rodrigue S, et al. Whole genome amplification and de novo assembly of single bacterial cells. PLoS One. 2009;4:e6864. - PMC - PubMed
    1. Fan HC, Wang J, Potanina A, Quake SR. Whole-genome molecular haplotyping of single cells. Nat Biotechnol. 2011;29:51–57. - PMC - PubMed
    1. Hou Y, et al. Single-cell exome sequencing and monoclonal evolution of a JAK2-negative myeloproliferative neoplasm. Cell. 2012;148:873–885. - PubMed
    1. Pan X, et al. A procedure for highly specific, sensitive, and unbiased whole-genome amplification. Proc Natl Acad Sci U S A. 2008;105:15499–15504. - PMC - PubMed

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources