1、Review Single Cell Genomics: Advances and Future Perspectives Iain C. Macaulay 1 , Thierry Voet 1,2 * 1Single Cell Genomics Centre, Wellcome Trust Sanger Institute, Hinxton, Cambridge, United Kingdom, 2Laboratory of Reproductive Genomics, Department of Human Genetics, KU Leuven, Belgium Abstract: Ad
2、vances in whole-genome and whole-tran- scriptome amplification have permittedthesequencing of the minute amounts of DNA and RNA present in a single cell, offering a window into the extent and nature of genomic and transcriptomic heterogeneity which occurs in both normal development and disease. Sing
3、le-cell approaches stand poised to revolutionise our capacity to understand the scale of genomic, epigenomic, and transcriptomic diversity that occurs during the lifetime of an individual organism. Here, we review the major technological and biological breakthroughs achieved, describe the remaining
4、challenges to overcome, and provide a glimpse into the promise of recent and future developments. Introduction Thecellisafundamental unitofbiology,inwhichtheblueprint of the genome is transcribed and translated into biological form and function. Almost all of our current understanding of the genome
5、and its regulation has been derived from studies carried outatthe populationleveltypicallythousandsormillionsofcells analysed in bulk. The resulting analysis, although unquestionably informative, oftenneglects anyheterogeneity that occurs within the population of cells. The genome, despite being wid
6、ely thought of as stable throughout normal development, has a small probability of acquiring genetic mutations with every cell division 1,2. Over sufficient divisions, genomic heterogeneity within the organism known as somatic variationis a certainty. While such variation liesattherootofmanydisorder
7、s3,4,includingcancer5,recent studies revealed unexpected levels of genomic variation in normal and diseased tissue, suggesting higher rates of genetic lesion than previously expected 612.Still,littleisknownaboutthe rateand nature of DNA mutation and how this is influenced by genetic background, life
8、style, and many other factors. The transcriptome is naturally more dynamic than the genome, reflecting the functionor typeof the cell. There is considerable evidence indicating that cell-to-cell variability in gene expression is ubiquitous, even within a phenotypically homogeneous popula- tionofcell
9、s13.Theextentoftranscriptionalheterogeneityandthe diversity of cell types in tissues remain, however, largely unknown. The genomic and transcriptomic composition of individual cells is lost in conventional sequencing studies, which analyse DNA and/or RNA extracted from large populations of cells; an
10、d de novo genome mutation and transcriptomic variations in cells will be largely concealed in the bulk signal. Clear insights into many biological processesfrom normal development to tumour evolutionwill thus only be gained froma detailed understanding of genomic, epigenomic, and transcriptional var
11、iation at the single-cell level. Furthermore, some cell types are so rare that single-cell approaches become paramount to their identification and characterisation. Advances in techniques for the isolation of single cells (Figure 1A), whole genome or transcriptome amplification, and genome-wide anal
12、ysis platformsprimarily next-generation se- quencing (NGS) devicespaved the way for high-resolution analysis of the genome or transcriptome from one cell, which reveals previously obscured biological complexity. Single-Cell Whole-Genome Amplification: Methods and Limitations A diploid human cell con
13、tains approximately 7 pg genomic DNA; necessitating amplification prior to microarray- or NGS- based analyses to detect various classes of genetic variation (Figure 1B1F). Current whole-genome amplification (WGA) principles are based on Multiple Displacement Amplification (MDA), Polymerase Chain Rea
14、ction (PCR), or a combination of both (Figure 2A2C). Unfortunately, no WGA method is faultless, and their various imperfections can considerably affect the interpretation of the microarray or NGS readout 14. The breadth of genomic coverage,theamplification biasduetolocaldifferencesinrichness for gua
15、nine and cytosine bases (%GC-bias), the prevalence of allelic drop outs (ADO), preferential allelic amplifications (PA), chimeric DNA-molecules, and nucleotide copy errors can vary significantly between different WGA approaches, making some methods better suited than others for detecting specific cl
16、asses of genetic variation 1417 (Figure 2D). A comparative analysis of all WGAmethods,including the investigation ofthe advantageous effectsofreducingthereactionvolumetoananoliterscale18,19, against a benchmark case is acute. AdvancesinNGSandBioinformaticsPermitHigh- Resolution Screening of a Single
17、-Cell Genome Single-cellWGAproductshavebeenanalysedusingavarietyof high-throughput platforms, including DNA-microarrays, SNP- arrays, and NGS (Figure 2D). A key difficulty in the interpretation Citation: Macaulay IC, Voet T (2014) Single Cell Genomics: Advances and Future Perspectives. PLoS Genet 10
18、(1): e1004126. doi:10.1371/journal.pgen.1004126 Editor: Nancy Maizels, University of Washington, United States of America Published January 30, 2014 Copyright: 2014 Macaulay, Voet. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits u
19、nrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: We acknowledge the Wellcome Trust (UK; http:/www.wellcome.ac.uk/ and http:/www.sanger.ac.uk/), the Research Foundation Flanders (FWO; http:/ www.fwo.be/; Belgium) FWO-G.A093.
20、11 to TV, and the KU Leuven http:/www. kuleuven.be/; Belgium; SymBioSys, PFV/10/016 to TV. The funders had no role in the preparation of the article. Competing Interests: I have read the journals policy and have the following conflicts: TV is a named inventor on submitted patent applications PCT/EP2
21、011/ 060211; PCT/EP2013/070858; ZL913096. * E-mail: Thierry.Voetmed.kuleuven.be PLOS Genetics | www.plosgenetics.org 1 January 2014 | Volume 10 | Issue 1 | e1004126Figure1.Detectionofvariousclassesofgeneticvariationusingsingle-cellWGA-NGSapproaches. A) The most prominent methods for (iii) isolating
22、individual cells (including (i) creation of single-cell suspensionsusually by enzymatic tissue disaggregationand subsequent cell isolationthroughmanualmicro-pipetting37,38,57,105,fluorescence-activatedcellsorting106,107ormicrofluidicsdevices18,81,108,and(ii)laser capture microdissection 109,110) as
23、well as (iii) isolating single nuclei 12,32,56,111 are indicated, accompanied with particular advantages and disadvantages. A comprehensive review of single-cell isolation methods is presented by Shapiro et al. 112. BD) Subsequently, the cell is lysed and its genomeamplified. A standardsequencinglib
24、rarycan be preparedfrom the WGA product for paired-end(or single-end)sequencing.Theresulting PLOS Genetics | www.plosgenetics.org 2 January 2014 | Volume 10 | Issue 1 | e1004126of single-cell WGA data on any platform is the separation of the numerous WGA artifacts from the genuine genetic variants p
25、resent in the template genome. Standard DNA-microarrays can detect copy number variations (CNVs) larger than 2.5 Mb from a single-cell genome 2022, while targeted array comparative genomic hybridizations can discover approximately 1 Mb-sized DNA imbalances 23, although remarkably, CNVs as small as 5
26、6 kb in single-cell PCR-based WGA products have been detected 24. Similarly, SNP-arrays can find copy number aberrations encompassing millions of bases in a cell 2528, but have the advantage of enabling the discovery of copy neutral DNA anomalies and regions of loss-of-heterozygosity (LOH), and allo
27、w inferring genome-wide haplotypes 2931. NGS has a number of advantages over microarrays enabling improved resolution and accuracy in variant calling 14. First, NGS can examine every nucleotide amplified from the cell and allows genome-wide discovery of the full spectrum of DNA mutations (Figure 1E)
28、, while microarrays only probe for certain CNV loci (Figure 2D). Secondly, sequencing provides digital precision, with one digital unit representing a mapped sequence read. Finally, paired-end sequencing and mapping discloses the linkage between both ends of each linear DNA-molecule in a sequencing
29、library of a single-cell WGA product, allowing the identification of structural variations via read-pairs mapping discordantly to the reference genome (Figure 1Ei). Analytical challenges remain in interpreting single-cell NGS data for the full spectrum of genetic variants. Although WGA imperfections
30、duetogenomebasecomposition(e.g.%GC-bias) can be computationally corrected for 17,32, the potential for PA and ADO can still generate local distortion in copy number, requiring distinct analyses to distinguish genuine copy number changes from WGA artefacts. Allelic fractions of heterozygous SNPs 26,3
31、3,34 or aberrantly mapping read pairs following paired-end sequencing of the WGA product 17 can be used to increase confidence in CNV measurements (Figure 1B1F). For instance, a real deletion of a diploid locus should show LOH and discordantly mapping read-pairs that explain the DNA loss. Furthermor
32、e, the cell cycle stage of the isolated cell must be considered, further complicating the analysis, as cells in S-phase demonstrate a dynamic copy number profile, leading to false structural DNA-imbalance discoveries 35. The identification of the full spectrum of intra- and inter- chromosomal (un)ba
33、lanced structural variants in a single-cell WGA product is still in its infancythe main difficulty being to filtertruestructuralvariantsfromchimericDNAgeneratedduring WGA, as well as issues with genome coverage (Figure 1Ei, 1F). Although filters have been designed to permit the detection of the stru
34、ctural architecture of DNA copy number variation 17 and even to detect L1-retrotransposition 36, many structural variants are still missed in single-cell analyses. Base alterations, such as SNPs, can be detected in single-cell WGA products (Figure 1Eiii). However, to call accurate and reliable base
35、substitutions in non-haploid loci, one requires the data of at least three cells to discriminate the variant from a WGA or sequencing error 16,37,38, and as such, detailed characterisation of extremely rare cells or sub-clones within populations may not be possible. Despite these hurdles, several gr
36、oups have proven the efficacy of single-cell NGS to detect multiple classes of mutation within a genome and even to detect sister chromatid exchanges following single-cell Strand-seq 39. Step-by-step bioinformatics protocols for analyzing Strand-seq data 40 as well as for copy number profilingsingle
37、cellsthroughNGS32ormicroarrayanalysis34 and commercial solutions (e.g. platforms used within 21,41) are surfacing. Single-Cell Genomics Reveals the Extent of Somatic Variation in Development and Disease The study of multiple classes of mutations at the single-cell level revealed startling insights i
38、nto the genomic variation that can occur during the human life cycle. Following single-cell genome- wide analysis, up to 7% 18,42 and up to 70% 4345 of male and femalegametes, respectively, containnumerical chromosomal anomalies due to meiotic mis-segregations. Furthermore, sequenc- ing of haploid s
39、ingle sperm cells revealed a base mutation rate of 24610 28 which is severalfold higher than measurements from genome-sequenced pedigree data 46. Single-cell analyses of human embryos following in vitro fertilization (IVF) demonstrated that the very first cell cycles of human life are prone to numer
40、ical and structural chromosome instability 17,2527,44,4751. Var- ious observations indirectly hint that an in vivo conceptus faces a similar period of increased genomic vulnerability 5255, suggesting that the first cell divisions may represent a spring of DNA mutation, which does not necessarily und
41、ermine normal development 8, but can lead to a spectrum of conditions, including loss of conception, genetic disorders, and genetic variation development. Several studies sequenced and dissected cancer genomes to single-cell resolution, with the aim of understanding tumour development and progressio
42、n of the disease. Copy number landscapes of single nuclei from primary mammary ductal carcinomas and a paired metastatic liver tumour were generated following low-coverage sequencing. This revealed various chro- mosomal rearrangements, followed by distinct phases of clonal expansion during tumour ev
43、olution and metastasis 56. Subse- quent single-cell exome sequencing studies in bladder 57, kidney 38, and hematopoietic neoplasms 37 provided a detailed characterisation of base mutations in specific genes. Similarly, whole-genome sequencing of multiple MALBAC-amplified cells (Figure2C)revealedabas
44、emutationrateofacancercelllinetobe increased 10-fold when compared to estimated germ-line ciphers 16.Furthermore,bysequencingdaughtercellsofasinglemitotic division,theacquisitionofnewCNVscouldbedemonstratedfora breast cancer cell line 17. (short) sequence reads of the cell are mapped against a refer
45、ence genome for variant discovery (EiEiii). In all steps (EiEiii towards F), various confounding factors resulting from the WGA process have to be considered in the analysis (indicated in red boxes). EiF) Structural variants can be detected by analysing read-pairs which map discordantly to the refer
46、ence genome, or by discovering split reads crossing a rearrangement. However, WGA can create various chimeric DNA molecules that resemble structuralvariants following paired-end sequenceanalysis of the WGA-product. EiiF) Copynumbervariantsarecalledbybinningreadsthatmaptoparticularregionsofthegenome.
47、Bycomparingthereadcountperbintothecounts obtained in a reference sample 17, or an average read count per bin 32, a copy number profile can be calculated. However, single-cell copy number profiles can be distorted by ADO, PA, and %GC-bias during the WGA process. EiiiF) Single nucleotide variants (SNV
48、s) can be detected in sequenced single-cell WGA products by aligning the reads with a reference genome. However, three cells carrying the same SNV are required to confidently call the variant. doi:10.1371/journal.pgen.1004126.g001 PLOS Genetics | www.plosgenetics.org 3 January 2014 | Volume 10 | Iss
49、ue 1 | e1004126Figure2.Overviewofsingle-cellWGAapproaches. A) Multiple displacement amplification (MDA) initiates with random priming of denatured single-cell DNA template, followed by a 30uC isothermal amplification using a DNA-polymerase with strand-displacement activity, typically phi29 113. When the 39-end of a newly synthesized fragment reaches the 59-end of an adjoining primed string of nucleotides, it will displace the latter, liberating sing
