Single Virus Genomics: A New Tool for Virus Discovery
Lisa Zeigler Allen
, Thomas Ishoey
, Mark A. Novotny
, Jeffrey S. McLean
, Roger S. Lasken
, Shannon J.
1Microbial and Environmental Genomics, J. Craig Venter Institute, San Diego, California, United States of America, 2Scripps Institution of Oceanography, University of
California San Diego, La Jolla, California, United States of America
Whole genome amplification and sequencing of single microbial cells has significantly influenced genomics and microbial
ecology by facilitating direct recovery of reference genome data. However, viral genomics continues to suffer due to
difficulties related to the isolation and characterization of uncultivated viruses. We report here on a new approach called
‘Single Virus Genomics’, which enabled the isolation and complete genome sequencing of the first single virus particle. A
mixed assemblage comprised of two known viruses; E. coli bacteriophages lambda and T4, were sorted using flow
cytometric methods and subsequently immobilized in an agarose matrix. Genome amplification was then achieved in situ
via multiple displacement amplification (MDA). The complete lambda phage genome was recovered with an average depth
of coverage of approximately 437X. The isolation and genome sequencing of uncultivated viruses using Single Virus
Genomics approaches will enable researchers to address questions about viral diversity, evolution, adaptation and ecology
that were previously unattainable.
Citation: Allen LZ, Ishoey T, Novotny MA, McLean JS, Lasken RS, et al. (2011) Single Virus Genomics: A New Tool for Virus Discovery. PLoS ONE 6(3): e17722.
Editor: Jean-Pierre Vartanian, Institut Pasteur, France
Received October 7, 2010; Accepted February 12, 2011; Published March 23, 2011
Copyright: ß2011 Allen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was supported by the J. Craig Venter Institute and the Office of Science (BER), U.S. Department of Energy, Cooperative Agreement
No. De-FC02-02ER63453. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: Patent application number 12/543,046 titled "Amplification of Single Viral Genomes". This does not alter the authors’ adherence to all the
PLoS ONE policies on sharing data and materials.
* E-mail: email@example.com
Whole genome amplification and sequencing of single microbial
cells is a powerful new tool in the field of microbial genomics,
enabling direct examination of the genomic contents of individual
cells without the need of cultivation [1–3]. Microbes are found in
nearly all environments (e.g., human microbiome, rhizosphere,
aquatic ecosystems, soils, air) performing essential roles in
processes such as biogeochemical cycling , degradation ,
metabolism , and forming the foundation of environmental
food webs . Sequencing of single cells permits the study of
previously uncharacterized microbes from virtually any environ-
ment, thus enabling the direct and comprehensive analysis of a
microbe’s genetic and metabolic repertoire. Flow cytometry [1,8]
and micromanipulation [9,10] have aided in the advent of single
cell isolation and sequencing by providing access to individuals
from naturally occurring consortia or pure cultures. A reaction
called multiple displacement amplification (MDA) [11–13], which
uses the high-fidelity processive capabilities of the phi29 DNA
polymerase, can amplify the genome of a bacterial cell more than
a billion-fold generating the micrograms of genomic DNA
typically required for DNA sequencing either via Sanger
sequencing , 454 pyrosequencing , and/or Illumina
platforms . While some sequence is lost due to non-specific
amplification or damage to the single genome copy, as much as
.90% of the genome has been recovered from single cell
sequencing [16,17]. Small MDA reaction volumes were shown
to improve amplification from single viral DNA molecules 
and single cells . Recently, Rodrigue et al., showed a consistent
increase in the total genome coverage of Prochlorococcus single-celled
amplified genomes by using a duplex-specific nuclease to degrade
highly abundant sequences apparent after amplification; thereby
increasing the coverage of low abundant sequences.
While most single cell studies have focused on bacteria and
cyanobacteria, single virions have yet to be isolated and
genomically described using similar mechanisms. Viruses are
ubiquitous and the most numerous and diverse biological entities
on our planet . Nearly all aspects of our lives are influenced by
viruses through shaping the environments that surround us ,
our immune responses  and even our genomes . The field
of environmental viral metagenomics has gained momentum over
the past several years [23–28]; however, sequencing of individual
environmental viral genomes is currently dependent on the
establishment of cultivable virus-host systems. With this in mind,
if less than one percent of microbial populations can be cultured
using standard microbiological techniques due to incongruencies
in direct counts versus cultivatable microbes [29–32], then only a
very small number of viruses have the likelihood of being
genomically described. Currently, viral genomic sequences are
lacking in public databases, with the exception of human viruses
and those of agricultural and industrial significance (e.g.
Lactococcal phages). Clearly, a better understanding of virus
diversity and evolution will not be achieved until the genomes of a
broad range of viruses are available. Here we introduce an
approach for isolating and characterizing the genomes of viruses
called ‘‘Single Virus Genomics’’ (SVG) (Figure 1). The benefits of
SVG will be far-reaching, enabling novel virus discovery in a
variety of clinical and environmental settings, altering our
understanding of virus evolution, adaptation and ecology and
PLoS ONE | www.plosone.org 1 March 2011 | Volume 6 | Issue 3 | e17722
facilitating the interpretation of viral genomic and metagenomic
data by providing suitable reference genomes.
Single virus isolation
Flow cytometric methods have been optimized for  and
used on natural viral populations for enumeration purposes [34–
36]. Therefore, for this study, flow cytometry was used to sort a
mixed viral assemblage consisting of two known viruses; E. coli
bacteriophages lambda and T4. To increase the accuracy of
detecting a single viral particle, a fluorescence-activated cell sorter
(FACS) AriaII with a forward scatter photo multiplier tube (PMT)
was used, which enabled more sensitive detection and lower size
thresholds (Figure 2). While 96- and 384-well microtiter plates
would have been optimal for high-throughput processing of viral
assemblages, we were unable to reliably recover single virions from
plate wells. The majority of wells (98%) contained no viruses
evidenced via polymerase chain reaction (PCR) amplification of
specific loci for each bacteriophage. Therefore, as an alternative
strategy, viruses stained with SYBR Green (Invitrogen) were sorted
directly onto cooled agarose beads applied to ‘multi-well’
polytetrafluoroethylene (PTFE) microscope slides to increase virus
capture efficiency, as well as, to maximize the recovery of high-
quality template DNA required for the MDA reaction. PTFE
slides were chosen due to the ability of defining each sorting event
since they possess distinct regions (wells) where agarose beads were
positioned. Overlaying the nanoliter droplet containing the virions
with additional agarose simultaneously embedded and stabilized
the sorted viral particles in preparation of downstream processing.
Single virus validation: Confocal microscopy and loci-
To confirm isolation of single viruses, Confocal Laser Scanning
Microscopy (CLSM) was performed to detect the fluorescently
stained virions embedded in agarose [37,38]. CLSM was chosen to
obtain greater confidence that only a single viral particle was
contained within an agarose bead through 3D rendering of stacked
images surrounding the viral particle (Figure 3A). Additionally,
Figure 3B demonstrates the utility of CLSM to detect the relative
fluorescence of a single stained virus particle above background.
Once successful candidates were identified, whole genome
amplification via MDA was performed in situ in order to minimize
the potential for virus loss, reduce genomic shearing, and
contamination. Multiplex PCR using T4 (gp23, major capsid
gene) and lambda (integrase gene) specific primers was performed
on the amplified genomic material to confirm the genotype of the
virus and indicated that the isolated viral particle used in this study
was phage lambda (Figure 4A). To further confirm this result, two
additional loci specific to lambda were targeted including the
superinfection exclusion protein B and repressor genes (Figure 4B).
The results confirmed our initial finding that we had isolated and
amplified the genome of phage lambda.
Figure 1. Flow diagram depicting SVG methodology. Viral
suspensions are sorted via flow cytometry onto PTFE slides with 24
distinct wells containing agarose beads. Viral particles are then
embedded within the agarose bead by overlaying with an additional
layer of agarose. Lastly, MDA is performed in situ.
Figure 2. Flow cytometric bi-plot showing SYBR-stained T4 and
lambda phage mixture. Gates were chosen to highlight T4/lambda
assemblages (green), and background (blue). Particles not gated (black)
were not sorted.
Single Virus Genomics
PLoS ONE | www.plosone.org 2 March 2011 | Volume 6 | Issue 3 | e17722
A subsequent experiment to quantify virus particles within
agarose droplets using CLSM indicated that 75% contained 1 or
.1 (1–5), viruses (Table S2); and amplification of genomic
material via MDA was successful in 92% of virus-containing
droplets. Multiplex PCR using T4 and lambda-specific primers on
amplified genomic material was successful for 25% of the droplets
and positive results were only found for those droplets containing
one or more virus.
Sequencing, reference mapping and De novo assembly
The 48,502 bp double stranded DNA phage lambda was
sequenced using 454 GS FLX titanium pyrosequencing to an
Figure 3. Confocal laser scanning micrograph of sorted viral particle embedded in agarose bead. A) Three dimensional reconstruction of
syber green I stained viral particle within depth of agarose bead verifying a single sorted event. Inset: higher magnification of viral particle. B) Profile
plot of relative fluorescence for a stained viral particle in an agarose bead. The blue line through the viral particle (green) is the reference for the inset
graph showing the relative fluorescence.
Figure 4. Phage identification using PCR. A) Multiplex PCR using T4 and lambda-specific primers to genotype, Lanes: 1. TrackIt 1 kb plus ladder
(Invitrogen), 2. Lambda integrase (750 bp), 3. T4 major capsid protein (1050 bp), 4. Mix of lambda integrase and T4 major capsid protein. B)
Subsequent lambda specific PCR with additional loci to further confirm phage genome isolation, Lanes: 1. Lambda integrase (750 bp), 2. Lambda
repressor (356 bp) 3. Lambda sie (superinfection exclusion) (456 bp) 4. TrackIt 1 kb plus ladder (Invitrogen).
Single Virus Genomics
PLoS ONE | www.plosone.org 3 March 2011 | Volume 6 | Issue 3 | e17722
average coverage of 437X across the genome (Figure 5C), (ranging
from 0–2000X). With the exception of the first 5 bp, the complete
genome of lambda was recovered (Table S3). Lacking the first
5 bp is likely due to a reported artifact of MDA reactions where
the ends of linear DNA segments are underrepresented [12,39]. It
has been reported that MDA is biased against genomic areas of
high GC content , however, our data suggests otherwise as
there was higher coverage in the regions of greater %GC, shown
in Figure 5A where the bars indicate the GC above or below the
average (average GC of phage lambda is 49%). We expected to
achieve .600X coverage from the 99,911 sequencing reads (mean
read length of 361.6 bp) that were generated if all sequences
produced were from the lambda template DNA. Reference
mapping to the lambda genome (NCBI Accession J02459)
indicated that 38,505 (38.5%) reads did not recruit to the genome
and were further termed ‘unmapped’. BLASTX analysis was
performed on these sequences, which resulted in 22,411 reads
(58.2% of the unmapped set) annotated based on homology to
public sequences (Table S4). The stringent settings used during
reference mapping prevented the recruitment of 116 sequences
classified as E. coli lambda phage through BLAST. The majority of
the annotated unmapped sequences were classified within the
Pseudomonas genera (12.9% of total reads). To assess errors within
the amplified lambda genome SNP analysis was completed, which
resulted in the detection of 17 SNPs across the genome (Table S6),
however, it is difficult to determine if these errors arose during
amplification, 454 pyrosequencing, or maintenance of ATCC
cultures. Deletion-Insertion Polymorphisms (DIPs) are also given
in Table S5; Two DIPs corresponding to reference positions
31619 and 39143 are deletions causing a frameshift within
essential phage proteins. The deletions (-) are found in 37.5 and
38.6 percent of the reads covering the corresponding positions in
the lambda repressor (cl) and DNA replication proteins respec-
tively. It is likely, therefore, that the polymorphism would not be
present in the phage population but perhaps are a result of MDA
and/or 454 pyrosequencing artifacts.
De novo assembly was performed with the GS De Novo
Assembler Software (i.e., Newbler, 454 Life Sciences) to assess
the utility of these methods for use on unknown SVGs (Table S6).
Optimal coverage depth for assembly is between 15–25X
(personal communication 454 Life Sciences; Newbler manual);
therefore the number of reads was randomly reduced to yield 30X
(4,700 reads) and 22X (3,400 reads) coverage, with the latter
generating the highest quality assembly (Figure 6A). Although
reducing coverage resulted in the highest quality assembly based
on assembly metrics (i.e., fewer contigs, longer length, greater N50
-see methods for details), using all reads (Figure 6B) resulted in
near complete coverage of the genome, however with shorter
contigs. In an effort to increase contig size while retaining genome
coverage, the redundancy of reads among overrepresented contigs
from this assembly was reduced and the data re-assembled, as seen
in Figure 6C . This method of bioinformatic normalization of
the data resulted in larger contigs coupled with almost complete
coverage of the genome (.99%). The utility of SVG approaches
for the study of uncultivated viruses will ultimately depend on the
success of de novo assembly due to the lack of suitable reference
genomes. Recently, SVG approaches were applied to virioplank-
ton samples collected from the Southern California Current
(Zeigler Allen, et al., in prep) followed by bioinformatic
normalization during de novo assembly procedures, which similarly
improved assembly statistics.
The Single Virus Genomics approach described here enabled,
for the first time, isolation and whole genome sequencing of an
individual virus; a significant technical achievement that has the
potential to alter the course of virological research. Further
optimization of SVG will pave the road to high throughput
processing of uncultivated viral assemblages, advancing studies of
viral diversity, evolution, adaptation and ecology. These include
efforts to improve the occurrence of single virus particles in
agarose droplets. Although genotyping of the single lambda phage
particle that yielded the sequence data for this study was successful
(Figure 2) and can clearly be accomplished, the overall low success
rate (25%) of specific PCR post-MDA evidenced was possibly due
Figure 5. Lambda genome attributes and coverage. A) GC plot with bars indicating %GC above or below the average of 49%, B) Genome map
of lambda (adapted from http://img.jgi.doe.gov), and C) Reference mapping of SVG reads to phage lambda, x-axis is genome position, y-axis is
Single Virus Genomics
PLoS ONE | www.plosone.org 4 March 2011 | Volume 6 | Issue 3 | e17722
Single Virus Genomics
PLoS ONE | www.plosone.org 5 March 2011 | Volume 6 | Issue 3 | e17722
to a lack of purification of the MDA products prior to genotyping
with T4 and lambda-specific primers, which is recommended (see
materials and methods). In addition, the high success rate of MDA
(92%), as evidenced by gel electrophoresis, could represent non-
specific amplification in addition to amplified viral DNA.
Optimization of flow-sorting parameters should increase the
likelihood of capturing individual viruses.
While complete coverage of the lambda phage genome was our
goal, the first 5 bp of the genome were missing, perhaps due to
DNA breakage or the linear nature of the molecule. To date,
complete coverage of bacterial genomes has not been reported,
suggesting that single cell genomics projects suffer from similar
obstacles. We make an effort to reduce DNA breakage through the
immobilization of viral particles in an agarose matrix which
minimizes DNA damage during viral particle isolation and
genome amplification. When applying SVG techniques to
unknown viruses, it may be difficult to determine if the ends of
linear genomes have been captured. However, approaches for
genome closure such as primer walking could be attempted on the
amplified viral genomic material if complete coverage is critical. A
recent study of MDA on phage lambda genomic DNA also
showed underrepresentation of DNA termini and reported using a
ligation reaction prior to MDA to generate circular molecules,
thereby overcoming this bias. A similar approach can be
adopted if future data suggests it is necessary.
Background DNA synthesis or nonspecific amplification is
commonly reported during amplification using the MDA reaction
. Nonspecific amplification has been attributed to contami-
nating DNA emerging from reaction reagents and/or through a
mechanism that enables amplification from the random hexamers
within the reaction mixture. The average coverage retrieved here
was lower than expected, most likely due to non-specific
amplification. As mentioned previously, steps were taken to
reduce the likelihood of contaminating DNAs being introduced
into our sample following flow cytometric sorting. However,
during the sorting process we acknowledge that free DNA as well
as multiple viral particles may be co-transported. Treatment of
viral assemblages with DNase I and examination of virus
containing agarose beads using confocal microscopy was used to
address these issues. Additionally, there is a higher likelihood of
nonspecific DNAs preferentially amplified due to the lower
quantity of template viral DNA as opposed to single bacterial
cells as a result of the significant difference in particle (cell) size and
genomic DNA content (25–100 nm; ,1.5femtograms for viruses,
as opposed to 0.2–1.5 um; ,14femtograms for bacteria). To
address this potential shortcoming the incubation time of genome
amplification was reduced and we took advantage of the massively
parallel, high-throughput capabilities of pyrosequencing to ensure
both adequate coverage of the lambda genome and to examine the
nature of any nonspecific amplification. A potential source of
contaminants is high molecular weight DNA fragments present in
commercially available phi29 polymerases. A recent report found
no manufactured enzyme to be contaminant-free and that levels of
contamination varied among enzyme and buffer reaction lots .
Specific 16S bacterial DNA sequences were detected as contam-
inants in our process. The identity of the microbial contaminants
present in no template control (NTC) MDA reactions using 16s
rDNA PCR and sequencing were determined and the most
abundant taxa are similar to those found in our taxonomic
classification of unmapped reads, in particular Burkholdaria and
Pseudomonas (Table S7). We have not attempted to distinguish
between the particle handling, MDA, 16S PCR and PCR product
sequencing steps as the potential source of the contaminants.
While bioinformatic curation of data can be performed to identify
potential contaminants that are not related to the target viral
molecule, conservative approaches are necessary in their removal
so that pertinent data is not lost. Therefore, steps to reduce the
amount of exogenous DNA-based contamination prior to
sequencing are imperative and are especially relevant when
working with unknown viral isolates. For example, testing new
enzyme and reagent lots prior to use and the reduction of free
DNA through nuclease treatment should help to reduce
A number of important factors must be taken into consideration
when applying SVG approaches to natural, unknown assemblages
of viruses. Although it is possible to capture and immobilize RNA-
containing viruses using flow cytometry, the MDA reaction will
not work on RNA templates. However, a reverse-transcription
step prior to amplification would address this issue and we are
currently evaluating the utility of whole transcriptome amplifica-
tion (WTA) to amplify individual RNA viral genomes. Genotyping
of previously unknown viruses is another topic that deserves
careful consideration since PCR using virus-specific primers
conserved across all viral groups is not option. While certain
techniques such as randomly amplified polymorphic (RAPD) PCR
may successfully produce unique viral genomic fingerprints ,
we are also evaluating alternative strategies such as optical
mapping  and automated artificial neural networks using
known morphological characteristics and fluorescence data
gathered from reference phage for training .
Single virus genomics has the potential to dramatically influence
a wide variety of fields that will benefit from whole genome
sequence data produced from previously uncultivated viruses;
including (but not limited to) viral and microbial ecology,
evolutionary biology, epidemiology, immunology and other
clinical and agricultural-based sciences. In addition to enabling
novel virus discovery and facilitating comparative genomic
analyses, SVG will also provide an ‘anchor’ for metagenomic
studies by supplying relevant reference genomes. Reference viral
genomes will not only assist assembly of metagenomic data, but
will help to address questions surrounding genetic and functional
biodiversity, as well as the representation of individual viruses
within a community. Lastly we anticipate that the production of
new reference viral genomes will improve our ability to classify
previously unidentified sequences within viral metagenomes,
effectively bridging the gap between genomic and metagenomic
Materials and Methods
Bacteriophage standards for T4 and lambda were obtained
from ATCC (ATCC 11303-B4 and 23724-B2, respectively).
Stocks were diluted in 0.1 mm-filtered TE (Tris-EDTA, pH 7.2,
Invitrogen) followed by filtration through a 0.22 mm Pall syringe
filter. The viral particles were not fixed prior to flow cytometry, as
is typical, due to insufficient evidence that the fixative would not
inhibit downstream reactions.
assembly of reads followed by reference mapping to evaluate assembly. A) Filtered sequences randomly to 3400 reads,
approximately 22X coverage of the lambda genome, B) All reads (99,911), C) Normalization of assembly by reducing redundancy of overrepresented
sequences from (B).
Single Virus Genomics
PLoS ONE | www.plosone.org 6 March 2011 | Volume 6 | Issue 3 | e17722
Flow cytometry parameters
Viral particle suspensions were sorted on a BD FACSAria II
Flow Cytometer equipped with a custom Forward Scatter PMT
(FSC PMT). The particles were diluted in 0.1 mm filtered TE
(Tris-EDTA, pH 7.2, Invitrogen) to an appropriate titer for an
event rate of 200 events s
. TE was used because it improves the
emission signal of stained viruses . Thresholds were set to FSC
PMT at 1000 and SSC at 200 for T4/lambda particles to
maximize signal-to-noise ratios. Prior to beginning the sorting,
blanks containing 0.1 mm-filtered TE were measured for back-
ground event recognition. In addition to blanks, unstained and
stained viral particles of the sample were measured to a total of
5,000 events each. Readings were measured on bi-exponential
plots, consisting of a lower linear scale and a higher exponential
scale. Viral particle suspensions were stained with SYBR Green I
(Invitrogen) and sorted onto polytetrafluoroethylene (PTFE)
printed microscope slides (Electron Microscopy Sciences). These
slides were chosen due to their hydrophobic feature, which
controls for cross-contamination and low microliter capacity. Data
was analyzed using the BD FACSDiva Software v.6.1 software
Twenty-four well PTFE slides were used for agarose immobi-
lization. To each well, 5 ml of low melting point (LMP) agarose,
cooled to 37uC, was added. The viral particle suspensions were
subsequently sorted directly onto the LMP agarose droplets at a
concentration of 1 event per well. Each well was then overlaid with
5ml of LMP agarose, cooled to 37uC, thus embedding the virions.
Visualization and whole genome amplification
The embedded virion(s) were stained with SYBR Green I and
visualized on the slide using Confocal Laser Scanning Microscopy
(CLSM) to determine that a single viral particle was present in
each agarose droplet. CLSM was performed with a Leica TCSP5
(Leica Microsystems) with 488 nm laser excitation. Image stacks
were obtained using a 636long working distance objective, which
enabled visualization of the viral particle in the depth of the
agarose plug. Simulated 3-D images and sections were generated
using the software Volocity and the plan views with side profile
slices using IMARIS (Bitplane AG, Zu˝rich, CH). Once a well was
identified as positive, the single viral particle was lysed using heat
(94uC) for 3 minutes and its genomic material amplified in situ
using the phi29 DNA polymerase and multiple displacement
amplification (MDA) reaction, as per manufacturers recommen-
dations (GenomiPhi kit, GE Healthcare). After amplification, the
genomic material was purified away from the agarose matrix using
ab-agarase (New England Biolabs) reaction followed by
purification using buffer-saturated phenol (Invitrogen) and ethanol
precipitation. Unincorporated dNTPs and random hexamer
primers were removed through column purification according to
manufacturer specifications, (PureLink Genomic DNA Purifica-
tion, Invitrogen) as they would be a source of contamination on
downstream reactions, such as sequencing. We highly recommend
the previous step as it was needed to reliably obtain successful
specific PCR results (see below). An additional round of MDA,
restricted to one hour, was performed in triplicate, pooled and the
amplified genomic DNA purified as described above.
Multiplex PCR was used for validation of model bacteriophage
isolation and genotyping. Primer sets specific to the lambda
integrase and T4 bacteriophage gp23 major capsid genes were
mixed and used in gradient PCR to identify the annealing
temperature for subsequent reactions (Table S1). PCR was
performed using PlatinumHTaq DNA polymerase, HiFi (Invitro-
gen). Additional genes were PCR amplified to verify the lambda
genotype, which included the lambda repressor (rep) and
superinfection exclusion (sie).
Library construction and sequencing
Purified amplified genomic DNA was randomly sheared using
nebulization and ends polished using BAL31 nuclease (New
England Biolabs) and T4 DNA polymerase (New England Biolabs)
reactions. Fragmented DNA was size selected using gel electro-
phoresis and 1% low melting point agarose. The DNA was
purified from the gel using b-agarase (New England Biolabs)
followed by buffer-saturated phenol extraction and ethanol
precipitation. Libraries for 454 pyrosequencing were constructed
using the sheared DNA. AMPure size fractionation was used to
purify the above reactions, followed by ligation of 454 adaptors
and emulsion PCR (ePCR). Sequencing was performed using the
The Nucleotide sequences were deposited as raw reads in
GenBank under the accession number SRA029358.
Reference mapping and De novo assembly
Reference mapping was conducted using CLC Genomics
Workbench, using the Enterobacteria phage lambda (ACC
J02459.1). Assembly parameters were as follows: local alignment
with mismatch cost 2, insertion cost 3, deletion cost 3, length
fraction 0.5, and similarity 0.9. Therefore, 50% of the read
needed to be aligned at 90% similarity. Parameters for SNP
analysis using CLC Genomics Workbench: max #of gaps and
mismatches 2, minimum average of quality of surrounding bases
15, minimum quality of central base 20, minimum coverage 1,
minimum variant frequence 35%. DIP analysis parameters using
CLC Genomics Workbench: minimum coverage 4, minimum
variant frequency 35%. De novo assembly was completed using
Newbler (454 Life Sciences Corporation, release 2.3). Default
settings were used for De novo assembly and reducing the amount
of sequences to gain 22X coverage of a ,50 Kb genome was
performed randomly. Following De novo assembly, reference
assembly (as described above) was performed using all contigs
generated to assess genome coverage. Bioinformatic normaliza-
tion was performed by reducing the redundancy of reads in
genomic regions of high coverage via clustering using cd-hit-est
.The contig N50 (bp) was recorded as an assembly metric and
represents the length of the smallest contig in the set that contains
the fewest (largest) contigs whose combined length represents at
least 50% of the assembly.
Table S1 Primers specific for phages T4 and lambda
loci used in multiplex PCR to identify phage isolated.
Table S2 Statistics following SVG methodology on 16
test samples. CLSM numbers corresponds to viruses detected
during microscopy, MDA refers to a positive (+) or negative (2)
when amplification was detected by gel electrophoresis of wells
containing viral particles. A positive specific PCR is denoted by the
genotype obtained after multiplex PCR of the amplified genomic
Single Virus Genomics
PLoS ONE | www.plosone.org 7 March 2011 | Volume 6 | Issue 3 | e17722
Table S3 Reference mapping statistics (all sequence
lengths are given in bp).
Table S4 BLAST analysis of unmapped read sequences
following reference mapping.
Table S5 Allele variation analysis using CLC Genomics
Table S6 De novo assembly statistics.
Table S7 Contaminants found in 16S PCR analysis of
We would like to thank Ken Nealson for his insight and advice throughout
the manuscript preparation process and Jasmine Pollard for assistance with
Conceived and designed the experiments: LZA TI SJW RSL. Performed
the experiments: LZA MAN JSM. Analyzed the data: LZA JSM SJW.
Contributed reagents/materials/analysis tools: SJW RSL JSM. Wrote the
paper: LZA SJW MAN JSM RSL.
1. Raghunathan A, Ferguson HR, Jr., Bornarth CJ, Song W, Driscoll M, et al.
(2005) Genomic DNA amplification from a single bacterium. Appl Environ
Microbiol 71(6): 3342–3347.
2. Lasken RS (2007) Single-cell genomic sequencing using Multiple Displacement
Amplification. Curr Opin Microbiol 10(5): 510–516.
3. Ishoey T, Woyke T, Stepanauskas R, Novotny M, Lasken RS (2008) Genomic
sequencing of single microbial cells from environmental samples. Curr Opin
Microbiol 11(3): 198–204.
4. Falkowski PG, Fenchel T, Delong EF (2008) The microbial engines that drive
Earth’s biogeochemical cycles. Science 320(5879): 1034–1039.
5. Amon RMW, Benner R (1996) Bacterial utilization of different size classes of
dissolved organic matter. Limnology and Oceanography 41(1): 41–51.
6. Lindell D, Jaffe JD, Johnson ZI, Church GM, Chisholm SW (2005)
Photosynthesis genes in marine viruses yield proteins during host infection.
Nature 438(7064): 86–89.
7. Azam F, Fenchel T, Field JG, Gray JS, Meyer Reil LA, et al. (1983) The
ecological role of water-column microbes in the sea. Mar Ecol Prog Ser 10(3 ):
8. Podar M, Abulencia CB, Walcher M, Hutchison D, Zengler K, et al. (2007)
Targeted access to the genomes of low-abundance organisms in complex
microbial communities. Appl Environ Microbiol 73(10): 3205–3214.
9. Lasken RS, Raghunathan A, Kvist T, Ishoy T, Westermann P, et al. (2005)
Multiple displacement amplification from single bacterial cells. In: Huges S,
Lasken R, eds. Whole Genome Amplification: Methods Express. UK: Scion
Publishing Ltd. pp 119–147.
10. Ishey T, Kvist T, Westermann P, Ahring BK (2006) An improved method for
single cell isolation of prokaryotes from meso-, thermo- and hyperthermophilic
environments using micromanipulation. Applied Microbiology and Biotechnol-
ogy 69(5): 510–514.
11. Dean FB, Nelson JR, Giesler TL, Lasken RS (2001) Rapid amplification of
plasmid and phage DNA using Phi 29 DNA polymerase and multiply-primed
rolling circle amplification. Genome Res 11(6): 1095–1099.
12. Dean FB, Hosono S, Fang L, Wu X, Faruqi AF, et al. (2002) Comprehensive
human genome amplification using multiple displacement amplification. Proc
Natl Acad Sci U S A 99(8): 5261–5266.
13. Hosono S, Faruqi AF, Dean FB, Du Y, Sun Z, et al. (2003) Unbiased whole-
genome amplification directly from clinical samples. Genome Res 13(5):
14. Zhang K, Martiny AC, Reppas NB, Barry KW, Malek J, et al. (2006)
Sequencing genomes from single cells by polymerase cloning. Nat Biotechnol
15. Marcy Y, Ishoey T, Lasken RS, Stockwell TB, Walenz BP, et al. (2007)
Nanoliter reactors improve multiple displacement amplification of genomes from
single cells. PLoS Genet 3(9): 1702–1708.
16. Rodrigue S, Malmstrom RR, Berlin AM, Birren BW, Henn MR, et al. (2009)
Whole genome amplification and de novo assembly of single bacterial cells.
PLoS One 4(9): e6864.
17. Woyke T, Xie G, Copeland A, Gonz!ulez JM, Han C, et al. (2009) Assembling
the Marine Metagenome, One Cell at a Time. PLoS One 4(4): e5299.
18. Hutchison CA, Smith HO, Pfannkoch C, Venter JC (2005) Cell-free cloning
using phi 29 DNA polymerase. Proc Natl Acad Sci U S A 102(48):
19. Edwards RA, Rohwer F (2005) Viral metagenomics. Nature Reviews
Microbiology 3(6): 504–510.
20. Rohwer F, Prangishvili D, Lindell D (2009) Roles of viruses in the environment.
Environ Microbiol 11(11): 2771–2774.
21. Fauci AS (1988) THE HUMAN IMMUNODEFICIENCY VIRUS - INFEC-
TIVITY AND MECHANISMS OF PATHOGENESIS. Science 239(4840):
22. Lower R, Lower J, Kurth R (1996) The viruses in all of us: Characteristics and
biological significance of human endogenous retrovirus sequences. Proc Natl
Acad Sci U S A 93(11): 5177–5184.
23. Breitbart M, Salamon P, Andresen B, Mahaffy JM, Segall AM, et al. (2002)
Genomic analysis of uncultured marine viral communities. Proc Natl Acad
Sci U S A 99(22): 14250–14255.
24. Breitbart M, Felts B, Kelley S, Mahaffy JM, Nulton J, et al. (2004) Diversity and
population structure of a near-shore marine-sediment viral community. Proc
Biol Sci 271(1539): 565–574.
25. Angly FE, Felts B, Breitbart M, Salamon P, Edwards RA, et al. (2006) The
marine viromes of four oceanic regions. PLoS Biol 4(11): e368.
26. Culley AI, Lang AS, Suttle CA (2006) Metagenomic analysis of coastal RNA
virus communities. Science 312(5781): 1795–1798.
27. Bench SR, Hanson TE, Williamson KE, Gosh D, Radosovich M, et al. (2007)
Metagenomic characterization of Chesapeake Bay virioplankton. Appl Environ
Microbiol 73(23): 7629–7641.
28. Williamson SJ, Rusch DB, Yooseph S, Halpern AL, Heidelberg KB, et al. (2008)
The Sorcerer II Global Ocean Sampling Expedition: metagenomic character-
ization of viruses within aquatic microbial samples. PLoS One 3(1): e1456.
29. Staley JT, Konopka A (1985) Measurement of in situ activities of nonphotosyn-
thetic microorganisms in aquatic and terrestrial habitats. Annu Rev Microbiol
30. Fuhrman JA, Campbell L (1998) Microbial microdiversity. Nature 393:
31. Whitman WB, Coleman DC, Wiebe WJ (1998) Prokaryotes: the unseen
majority. Proc Natl Acad Sci U S A 95(12): 6578–6583.
32. Rappe MS, Giovannoni SJ (2003) The uncultured microbial majority. Annu
Rev Microbiol 57: 369–394.
33. Brussaard CP (2004) Optimization of procedures for counting viruses by flow
cytometry. Appl Environ Microbiol 70(3): 1506–1513.
34. Marie D, Brussaard CPD, Thyrhaug R, Bratbak G, Vaulot D (1999)
Enumeration of marine viruses in culture and natur al samples by flow
cytometry. Appl Environ Microbiol 65(1): 45.
35. Brussaard CPD, Marieb D, Bratbak G (2000) Flow cytometric detection of
viruses. J Virol Meth 85(1-2): 175–182.
36. Brussaard CP (2009) Enumeration of bacteriophages using flow cytometry.
Methods Mol Biol 501: 97–111.
37. Noble RT, Fuhrman JA (1998) Use of SYBR Green I for rapid epifluorescence
counts of marine viruses and bacteria. Aquat Microb Ecol 14(2): 113–118.
38. Luef B, Neu TR, Peduzzi P (2009) Imaging and quantifying virus fluorescence
signals on aquatic aggregates: a new method and its implication for aquatic
microbial ecology. FEMS Microbiol Ecol 68(3): 372–380.
39. Tzvetkov MV, Becker C, Kulle B, Nurnberg P, Brockmoller J, et al. (2005)
Genome-wide single-nucleotide polymorphism arrays demonstrate high fidelity
of multiple displacement-based whole-genome amplification. Electrophoresis
40. Pinard R, de Winter A, Sarkis GJ, Gerstein MB, Tartaro KR, et al. (2006)
Assessment of whole genome amplification-induced bias through high-
throughput, massively parallel whole genome sequencing. BMC Genomics 7:
41. Miller JR, Koren S, Sutton G (2010) Assembly algorithms for next-generation
sequencing data. Genomics 95(6): 315–327.
42. Panelli S, Damiani G, Espen L, Sgaramella V (2005) Ligation overcomes
terminal underrepresentation in multiple displacement amplification of linear
DNA. Biotechniques 39(2): 174, 176, 178 passim.
43. Blainey PC, Quake SR. Digital MDA for enumeration of total nucleic acid
contamination. Nucleic Acids Research.
44. Winget DM, Wommack KE (2008) Randomly amplified polymorphic DNA
PCR as a tool for assessment of marine viral richness. Appl Environ Microbiol
Single Virus Genomics
PLoS ONE | www.plosone.org 8 March 2011 | Volume 6 | Issue 3 | e17722
45. Meng X, Benson K, Chada K, Huff EJ, Schwartz DC (1995) Optical mapping of
lambda bacteriophage clones using restriction endonucleases. Nature Genetics
46. Storrie-Lombardi MC, Irwin MJ, von Hippel T, Storrie-Lombardi LJ (1994)
Spectral classification with principal component analysis and artificial neural
networks. Vistas in Astronomy 38: 331–340.
47. Li WZ, Godzik A (2006) Cd-hit: a fast program for clustering and comparing
large sets of protein or nucleotide sequences. Bioinformatics 22(13): 1658–1659.
Single Virus Genomics
PLoS ONE | www.plosone.org 9 March 2011 | Volume 6 | Issue 3 | e17722