Open Access

The complete genome, structural proteome, comparative genomics and phylogenetic analysis of a broad host lytic bacteriophage ϕD3 infecting pectinolytic Dickeya spp.

  • Robert Czajkowski1Email author,
  • Zofia Ozymko1,
  • Joanna Siwinska1,
  • Adam Ossowicki1,
  • Victor de Jager2,
  • Magdalena Narajczyk3 and
  • Ewa Łojkowska1
Standards in Genomic Sciences201510:68

https://doi.org/10.1186/s40793-015-0068-z

Received: 1 May 2015

Accepted: 17 September 2015

Published: 24 September 2015

Abstract

Plant necrotrophic Dickeya spp. are among the top ten most devastating bacterial plant pathogens able to infect a number of different plant species worldwide including economically important crops. Little is known of the lytic bacteriophages infecting Dickeya spp. A broad host lytic bacteriophage ϕD3 belonging to the family Myoviridae and order Caudovirales has been isolated in our previous study. This report provides detailed information of its annotated genome, structural proteome and phylogenetic relationships with known lytic bacteriophages infecting species of the Enterobacteriaceae family.

Introduction

Pectinolytic Dickeya spp. can cause disease on a number of arable and ornamental crops worldwide including potato, tomato, carrot, onion, pineapple, maize, rice, hyacinth, chrysanthemum and calla lily resulting into severe economic losses [1]. Dickeya spp. are recognized to be among the top ten most important bacterial pathogens in agriculture [2]. To date there is no effective control of Dickeya spp. in agriculture due to the lack of practical measures and strategies [3].

Lytic bacteriophages have been proposed as potential biological control agents against various pathogenic bacterial species including plant pathogens [4]. Their potential to control plant bacterial diseases has been evaluated among others against Erwinia amylovora , Xanthomonas pruni, Ralstonia solanacearum and also were experimentally tested against Pectobacterium spp. and Dickeya spp. in different crop systems [4]. In the case of Pectobacterium spp. and Dickeya spp. lytic bacteriophages, only limited attempts have been made so far to isolate and characterize these bacteriophages in detail [5, 6] and to provide information on their genomes and structural proteomes [7].

At present, only two Dickeya spp. lytic bacteriophages: LimeStone1 and ϕD5 were characterized in detail, viz. their complete genomes are available in the Genbank (accessions: NC019925 and KJ716335, respectively) and information on other features (e. g. structural proteomes and host range, multiplicity of infection and adsorption to bacterial hosts) is also available [6, 7].

Virus information

Bacteriophage ϕD3 was isolated from garden soil collected in Kujawsko-Pomorskie region (Kuyavian-Pomeranian Province) in 2013 in Poland and it has been characterized in full for morphologic and phenotypic features [5]. It is a broad host lytic phage belonging to Myoviridae family and Caudovirales order and infecting isolates of D. solani , D. dadantii , D. dianthicola , D. zeae and D. chrysanthemi species. In transmission electron microscopy, this bacteriophage was characterized by the presence of a 130 nm long contractile tail, a head of 100 nm in diameter and of dodecahedral symmetry [5] (Fig. 1).
Fig. 1

Transmission electron micrograph of Dickeya spp. bacteriophage ϕD3 stained with uranyl acetate. Bacteriophage particle was purified four times by passaging individual plaques using the soft top agar method and D. solani IPO2222 as a host. Phage suspension of ca. 105 plaque forming units (pfu) ml−1 in 1/4 Ringer’s buffer was used for microscopy. At least 10 different photographs were taken. The micrograph presents typical ϕD3 phage particle. Bar marker represents 100 nm [5]

Chemotaxonomic data

To better characterize bacteriophage ϕD3, we performed in addition to the genome characterization also SDS-PAGE and MS analysis of its structural proteins [8]. Protein bands were excised from the gels with a sterile scalpel and used for mass spectrometry analysis performed at the Mass Spectrometry Laboratory, Institute of Biochemistry and Biophysics, Polish Academy of Sciences in Warsaw, Poland. In order to predict the molecular functions of the unknown structural proteins obtained from SDS-PAGE and MS analysis we used GeneSillico Protein Structure Prediction Meta-server containing known three-dimensional (3D) protein structures [9] and PSI-BLAST accessed via NCBI website [10]. The computational protein predictions with the highest scores were considered as the most valid [9, 10]. This direct and bioinformatic approach led to the experimental identification of 10 structural proteins of ϕD3. From these, the function of 7 proteins could be assigned directly based on sequence similarities with the other known phage proteins (Fig. 2). The most abundant protein was major capsid protein gp23. Three proteins present in the ϕD3 proteome were characterized by MS as unknown structural proteins for which no function could be inferred based on homology with amino acid sequences present in the current databases. These proteins were analyzed by comparing their sequences with protein sequences deposited in the GeneSillico protein 3D structure database. We were then able to assign functions to all unknown proteins using this approach.
Fig. 2

SDS-PAGE and MS analysis of ϕD3 structural proteins. For SDS-PAGE electrophoresis ca. 109 pfu ml−1 were mixed with Laemmli buffer and frozen in liquid nitrogen for 1-2 min. following the boiling at 95 °C for 5 min. The phage proteins were separated in 12 % acrylamide SDS-PAGE gel for ca. 19 h t 50 V at 22 °C. The bands were stained with PageBlue Coomasie Blue (Thermo Scientific) according to protocol provided by the manufacturer. For MS analysis of phage structural proteins, protein bands obtained from SDS-PAGE were excised from gel with a sterile scalpel and sent to the mass spectrometry analysis to Mass Spectrometry Laboratory, Institute of Biochemistry and Biophysics, Polish Academy of Science in Warsaw, Poland. Possible molecular functions of the unknown structural proteins were elucidated using Gene Sillico Protein Structure Prediction Meta-server [9]

Genome sequencing information

Genome project history

A number of recent studies have shown that bacteriophages play a substantial role in global ecosystems and have a direct bearing on the ecology and evolution of their hosts. The ϕD3 genome is the third (after LimeStone1 and ϕD5) complete genome of lytic bacteriophage virulent to plant pathogenic Dickeya spp. available to the scientific community. Genome sequencing and analysis provide a better possibility to deduce phage infections in host cells and phage interaction with a variable environment. This genome project was deposited in NCBI Genbank as Bioproject PRJNA242299 under the title: “Bacteriophages of Pectobacterium spp. and Dickeya spp. Genome sequencing”. A summary of the project information is shown in Table 2.

Growth conditions and genomic DNA preparation

D. solani IPO2222 (type strain for D. solani ), grown on tryptone soya agar (Oxoid) and/or in tryptone soya broth (Oxoid), was used in all experiments as a ϕD3 host. Bacteriophage ϕD3 was isolated as described previously [5] from Dickeya spp.-free garden soil which may indicate that the phage can infect also different soil-borne bacteria as additional hosts. Purification and concentration of phage particles followed the previous protocols and included: DNase I and RNase A treatments, CsCl gradient ultracentrifugation and dialysis to remove CsCl from phage concentrated samples [7]. Purified phage particles were resuspended in 500 μl of 5 mM MgSO4 or in 1/4 Ringer’s buffer (Merck) and stored at 4 °C in the dark. The ϕD3 genomic DNA was purified using CTAB method as described in [11].

Genome sequencing and assembly

The genome was sequenced using the Illumina next generation technology at Baseclear, The Netherlands, following the manufacturer’s instructions (Illumina). The sequencing library yielded ca. 270 Mb clean data reads after sets of rigorous filtrations against bacterial host genomic DNA ( D. solani strain IPO2222, Genbank accession: AONU00000000). De novo assembly of the ϕD3 genome from the resulting raw reads was performed using CLC Genomic Workbench 7.5 (CLC bio) as described earlier [12] which provided >1500 x coverage of the genome.

Genome annotation

The ϕD3 genome was mapped and annotated using available bacteriophage genomic sequences deposited in GenBank. Structural and functional annotations for the ϕD3 genome were obtained from the Annotation Service Automatic Pipeline (Institute for Genome Science, School of Medicine, University of Maryland, USA) and confirmed using RAST set to auto settings. Additional analysis of the gene predictions and annotations was supplemented using Manatee accessed via the website of IGS, University of Maryland, USA. The lifestyle of ϕD3 (temperate [lysogenic] or lytic) was predicted using PHACTS [13]. To find potential genes acquired by ϕD3 coding for toxins and allergens, the genome sequence was analyzed bioinformatic analysis using Virulence Finder 1.2 and VirulentPred.

Genome properties

Tables 1, 2, 3 and 4 summarize the properties and statistics of the ϕD3 genome. The capsid of ϕD3 contains circular double-stranded DNA genome of 152 308 bp, with an average GC content of 49.3%. The complete genome possesses 191 open reading frames (190 ORFs with the average gene length calculated to be 730 nucleotides) and one tRNA-Met (tRNA-methionine) ORF. A total of 105 ORFs (54.9%) have assigned function, whereas 45.1% (86 ORFs) are conserved hypothetical ORFs for which no homology with known genes was found in the NCBI database. Forty one ORFs (21.5%) were unclassified with no assigned role category (Fig. 3a). The lifestyle of ϕD3 predicted from PHACTS indicated that it is a lytic bacteriophage. The ϕD3 genome does not contain any genes coding for (known) toxins, allergens and other virulence factors as tested by VirulenceFinfer 1.2 and VirulencePred. Likewise, a search in BLAST did not reveal the presence of toxins, allergens, integrases and/or antibiotic resistance genes in the genome of ϕD3. The compete genome sequence of ϕD3 was deposited at DDBJ/EMBL/Genbank under accession number KM209228.
Table 1

Classification and general features of Dickeya spp. bacteriophage ϕD3

MIGS ID

Property

Term

Evidence codea

 

Classification

Domain: Viruses, dsDNA viruses, no RNA viruses

TAS [5]

  

Phylum: unassigned

TAS [5]

  

Class: unassigned

TAS [5]

  

Order: Caudovirales

TAS [5]

  

Family: Myoviridae

TAS [5]

  

Genus: unassigned

TAS [5]

  

Species: unassigned

TAS [5]

 

Gram stain

Not applicable

TAS [5]

 

Particle shape

Icosahedral

IDA

 

Motility

Not applicable

TAS [5]

 

Sporulation

Not applicable

TAS [5]

 

Temperature range

Not applicable

TAS [5]

 

Optimum temperature

Not applicable

TAS [5]

 

pH range; Optimum

Not applicable

TAS [5]

 

Carbon source

Not applicable

TAS [5]

MIGS-6

Habitat

Soil

IDA

MIGS-6.3

Salinity

Not applicable

TAS [5]

MIGS-22

Oxygen requirement

Not applicable

TAS [5]

MIGS-15

Biotic relationship

Obligate intracellular parasite of Dickeya spp.

IDA

MIGS-14

Pathogenicity

Lytic virus of Dickeya spp.

IDA

MIGS-4

Geographic location

Poland / Kujawsko-Pomorskie (Kuyavian-Pomeranian Province)

IDA

MIGS-5

Sample collection

February 18, 2013

IDA

MIGS-4.1

Latitude

53.68 N

IDA

MIGS-4.2

Longitude

18.09 E

IDA

MIGS-4.3

Depth

20 cm

IDA

MIGS-4.4

Altitude

118 m

IDA

aEvidence codes - IDA inferred from direct assay, TAS traceable author statement (i.e., a direct report exists in the literature), NAS non-traceable author statement (i.e., not directly observed for the living, isolated sample, but based on a generally accepted property for the species, or anecdotal evidence). These evidence codes are from the Gene Ontology project [20]

Table 2

Project information

MIGS ID

Property

Term

MIGS-31

Finishing quality

Complete

MIGS-28

Libraries used

One paired-end library

MIGS-29

Sequencing platforms

Illumina

MIGS-31.2

Fold coverage

1753×

MIGS-30

Assemblers

CLC Genomics Workbench, version 7.0.3

MIGS-32

Gene calling method

RAST version 4.0, IGS Annotation Service (Manatee)

 

Locus Tag

HQ80

 

Genbank ID

KM209228

 

GenBank Date of Release

16.07.2016 (earlier upon publication)

 

GOLD ID

GP0111934

 

BIOPROJECT

PRJNA242299

MIGS-13

Source Material Identifier

NCNRC002.D3

 

Project relevance

Biological effects in soil and plant environments

Table 3

Genome statistics

Attribute

Value

% of Total

Genome size (bp)

152,308

100.0

DNA coding (bp)

138,905

91.1

DNA G + C (bp)

75,088

49.3

DNA scaffolds

1

100.0

Total genes

191

100.0

Protein coding genes

190

99.5

RNA genes

1

0.5

Pseudo genes

0

0.0

Genes in internal clusters

0

0.0

Genes with function prediction

105

54.9

Genes assigned to COGs

64

33.5

Genes with signal peptides

0

0.0

Genes with transmembrane helices

0

0.0

Table 4

Number of genes associated with general COG functional categories

Code

Value

% age

Description

J

0

0.00

Translation, ribosomal structure and biogenesis

A

1

0.53

RNA processing and modification

K

4

2.11

Transcription

L

9

4.74

Replication, recombination and repair

B

0

0.00

Chromatin structure and dynamics

D

6

3.16

Cell cycle control, Cell division, chromosome partitioning

V

0

0.00

Defense mechanisms

T

0

0.00

Signal transduction mechanisms

M

0

0.00

Cell wall/membrane biogenesis

N

0

0.00

Cell motility

U

0

0.00

Intracellular trafficking and secretion

O

0

0.00

Posttranslational modification, protein turnover, chaperones

C

1

0.53

Energy production and conversion

G

0

0.00

Carbohydrate transport and metabolism

E

0

0.00

Amino acid transport and metabolism

F

2

1.05

Nucleotide transport and metabolism

H

0

0.00

Coenzyme transport and metabolism

I

0

0.00

Lipid transport and metabolism

P

0

0.00

Inorganic ion transport and metabolism

Q

0

0.00

Secondary metabolites biosynthesis, transport and catabolism

R

41

21.6

General function prediction only

S

10

5.3

Function unknown

-

139

60.98

Not in COGs

The total is based on the total number of protein coding genes in the genome

Fig. 3

Phage ϕD3 genome (a) and phylogenetic analysis (b). a The genome of bacteriophage ϕD3 (152,308 bp). Structural and functional annotations were obtained from IGS Annotation Service (http://ae.igs.umaryland.edu/cgi/index.cgi) and from RAST (http://rast.nmpdr.org/). ORFs coding for proteins involved in DNA metabolism, transcription and translation are marked in red, ORFs coding for proteins involved in phage particle assembly are marked in blue and ORFs coding for enzymes are marked in green. Arrows indicate the direction of transcription and translation. The ORFs coding for hypothetical proteins are not shown on the map. The figure was generated using a genome visualization tool – SnapGene ver. 2.6.2. b Maximum likelihood tree based on the aligned consensus nucleotide sequences (600 bp. long each) of gp20 genes of bacteriophages closely related to Dickeya sp. phage ϕD3. Enterobacteria phage T4 was used as an outgroup. Phylogenetic studies were performed using Phylip package. Bootstrap values (per 1000 replicates) are shown at branch points. The bar indicates the number of substitutions per sequence position

Comparisons with other genomes of Dickeya spp. bacteriophages and bacteriophage T4

Multiple genome alignment was performed using Mauve [14] and comparative genomics analysis was done using EDGAR [15]. A pairwise comparison of the complete four genome sequence of ϕD3, ϕD5 [7], LimeStone1 [6] and Enterobacteriaceae bacteriophage T4 revealed that ϕD3, ϕD5 and LimeStone1 share considerable genetic similarity which may suggest their common origin (Fig. 4). This is unexpected considering the fact that LimeStone1 was isolated in Belgium and ϕD3 and ϕD5 were isolated in different regions in Poland. The core (common) genome of ϕD3, ϕD5 and LimeStone1 consists of 178 genes, whereas only 7, 13 and 6 genes are specific for phages ϕD3, ϕD5 and LimeStone1, respectively (Fig. 4).
Fig. 4

Core genome of ϕD3, ϕD5 and LimeStone1 bacteriophages

Interestingly, the majority of the genes found in ϕD3 do not have homologs in T4 (one of the best described and characterized Myoviridae bacteriophages) and only two genes are present in both phages viz. (i) phage recombination protein and (ii) phage endoribonuclease translational repressor of early genes.

Bacteriophage capsid assembly protein (gp20) was used for phylogenetic analysis as previously described [16, 17]. Nucleotide sequences of gp20 proteins of LimeStone1 (Genbank accession: NC019925), bacteriophage ϕD5 (KJ716335), Shigella phage phiSboM-AG3 (NC013693), Salmonella phage SKML-39 (NC019910), Klebsiella phage 0507-KN2-1 (NC022343), Salmonella phage vB SalM SJ3 (NC024122), Escherichia coli phage PhaxI (NC0194521), E. coli phage vB_EcoM_CBA120 (NC016570), Salmonella phage SFP10( NC016073), Salmonella phage PhiSH19 (NC019530), Salmonella phage Maynard (NC022768), Salmonella phage Marshall (NC022772), E. coli phage ECML-4 (NC025446), Salmonella phage vB SalM SJ2 (NC023856), Salmonella phage Vi01 (NC015296) were obtained from GenBank. ClustalX was used to align nucleotide sequences and to manually correct aligned sequences prior to further analyses. Phylogeny studies were performed with the use of the Phylip program [18] and Molecular Evolutionary Genetic Analysis (MEGA6) software [19]. Dendrograms were created using the Maximum likelihood method followed by calculating the p-distance matrix for aligned gp20 nucleotide sequences (length of gp20 nucleotide sequences: 600 bp, nucleotide substitution model: K80 Kimura) with the bootstrap support fixed to 1000 re-samplings. To root the tree, a gp20 nucleotide sequence from Enterobacteriaceae bacteriophage T4 derived from its complete genome (NC000866) was used.

As expected, ϕD3 showed the highest similarity to the other described Dickeya spp. bacteriophages (LimeStone1 and ϕD5). On the basis of the gp20 phylogenetic analysis, ϕD3 was also closely related to Shigella phage phiSboM-AG3 and Salmonella phage SKML-39. The largest phylogenetic distance was observed between ϕD3 and Enterobacteriaceae phage T4 (Fig. 3b).

Conclusions

As far we know, the ϕD3 is the third bacteriophage able to infect (and kill) several species of Dickeya that has been genetically characterized in depth and is also the second Dickeya spp. lytic bacteriophage isolated in Poland. We expect that the availability of an additional Dickeya spp. specific bacteriophage would improve our understanding of bacteriophage – bacteria interactions and gives an insight on conservation and evolution of Dickeya spp. lytic bacteriophages as well as improve our knowledge on Dickeya spp. ecological fitness in complex (soil, rhizosphere and phyllosphere) environments.

Declarations

Acknowledgements

This work was financially supported by the National Science Centre, Poland (Narodowe Centrum Nauki, Polska) via a postdoctoral research grant FUGA1 (DEC-2012/04/S/NZ9/00018) to Robert Czajkowski. The authors are grateful to Michel C. M. Perombelon (ex. SCRI, Dundee, Scotland) for helpful discussion and his editorial work on the manuscript.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

(1)
Department of Biotechnology, Intercollegiate Faculty of Biotechnology, University of Gdansk and Medical University of Gdansk
(2)
Netherlands Institute of Ecology (NIOO-KNAW)
(3)
Laboratory of Electron Microscopy, Faculty of Biology, University of Gdansk

References

  1. Toth IK, van der Wolf JM, Saddler G, Lojkowska E, Hélias V, Pirhonen M, et al. Dickeya species: an emerging problem for potato production in Europe. Plant Pathol. 2011;60:385–99.View ArticleGoogle Scholar
  2. Mansfield J, Genin S, Magori S, Citovsky V, Sriariyanum M, Ronald M, et al. Top 10 plant pathogenic bacteria in molecular plant pathology. Mol Plant Pathol. 2012;13:614–29.View ArticlePubMedGoogle Scholar
  3. Czajkowski R, Pérombelon MCM, van Veen JA, van der Wolf JM. Control of blackleg and tuber soft rot of potato caused by Pectobacterium and Dickeya species: a review. Plant Pathol. 2011;60:999–1013.View ArticleGoogle Scholar
  4. Jones JB, Jackson LE, Balogh B, Obradovic A, Iriarte FB, Momol MT. Bacteriophages for plant disease control. Annu Rev Phytopathol. 2008;45:245–62.View ArticleGoogle Scholar
  5. Czajkowski R, Ozymko Z, Lojkowska E. Isolation and characterization of novel soilborne lytic bacteriophages infecting Dickeya spp. biovar 3 (‘D. solani’). Plant Pathol. 2014;63:758–72.View ArticleGoogle Scholar
  6. Adriaenssens EM, Van Vaerenbergh J, Vandenheuvel D, Dunon V, Ceyssens PJ, De Proft M, et al. T4-related bacteriophage LIMEstone isolates for the control of soft rot on potato caused by ‘Dickeya solani’. PLoS ONE. 2012;7(3), e33227.PubMed CentralView ArticlePubMedGoogle Scholar
  7. Czajkowski R, Ozymko Z, Zwirowski S, Lojkowska E. Complete genome sequence of a broad-host-range lytic Dickeya spp. bacteriophage ϕD5. Arch Virol. 2014;159:3153–5.PubMed CentralView ArticlePubMedGoogle Scholar
  8. Boulanger P. Purification of bacteriophages and SDS-PAGE analysis of phage structural Proteins from ghost particles. In: Bacteriophages. Methods in Molecular Biology™: Humana Press. Edited by Clokie MJ, Kropinski A, 2009: 227-38.Google Scholar
  9. Kurowski MA, Bujnicki JM. GeneSilico protein structure prediction meta-server. Nucl Acids Res. 2003;31:3305–7.PubMed CentralView ArticlePubMedGoogle Scholar
  10. Kawabata T, Arisaka F, Nishikawa K. Structural/functional assignment of unknown bacteriophage T4 proteins by iterative database searches. Gene. 2000;259:223–33.View ArticlePubMedGoogle Scholar
  11. Frampton RA, Taylor C, Holguín Moreno AV, Visnovsky SB, Petty NK, Pitman AR, et al. Identification of bacteriophages for biocontrol of the kiwifruit canker phytopathogen Pseudomonas syringae pv. actinidiae. Appl Environ Microbiol. 2014;80:2216–28.PubMed CentralView ArticlePubMedGoogle Scholar
  12. Sykilinda NN, Bondar AA, Gorshkova AS, Kurochkina LP, Kulikov EE, Shneider MM, et al. Complete genome sequence of the novel giant Pseudomonas phage PaBG. Genome Announc. 2014;2.Google Scholar
  13. McNair K, Bailey BA, Edwards RA. PHACTS, a computational approach to classifying the lifestyle of phages. Bioinformatics. 2012;28(5):614–8.PubMed CentralView ArticlePubMedGoogle Scholar
  14. Darling AE, Mau B, Perna NT. ProgressiveMauve: multiple genome alignment with gene gain, loss and rearrangement. PLoS ONE. 2010;5(6), e11147.PubMed CentralView ArticlePubMedGoogle Scholar
  15. Blom J, Albaum S, Doppmeier D, Puhler A, Vorholter F-J, Zakrzewski M, et al. EDGAR: a software framework for the comparative analysis of prokaryotic genomes. BMC Bioinformatics. 2009;10(1):154.PubMed CentralView ArticlePubMedGoogle Scholar
  16. Brewer TE, Stroupe ME, Jones KM. The genome, proteome and phylogenetic analysis of Sinorhizobium meliloti phage ϕM12, the founder of a new group of T4-superfamily phages. Virology. 2014;450–451:84–97.View ArticlePubMedGoogle Scholar
  17. Comeau AM, Krisch HM. The Capsid of the T4 phage superfamily: the evolution, diversity, and structure of some of the most prevalent proteins in the biosphere. Mol Biol Evol. 2008;25:1321–32.View ArticlePubMedGoogle Scholar
  18. Felsenstein J. PHYLIP (phylogeny inference package), version 3.5 c. Joseph Felsenstein.; 1993.Google Scholar
  19. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: molecular evolutionary genetics analysis version 6.0. Mol Biol Evol. 2013;30:2725–9.PubMed CentralView ArticlePubMedGoogle Scholar
  20. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. Nat Genet. 2000;25:25–9.PubMed CentralView ArticlePubMedGoogle Scholar

Copyright

© Czajkowski et al. 2015

Advertisement