Plenary: Advanced Molecular Detection in Environmental ... · MO10 O139. 623-39. 1587 O12. VL426...
Transcript of Plenary: Advanced Molecular Detection in Environmental ... · MO10 O139. 623-39. 1587 O12. VL426...
APHL Annual MeetingRhode Island Convention Center
June 12, 2017
Plenary:Advanced Molecular
Detection in Environmental
Matrices
Rita R. Colwell, Ph.D., D.Sc.Distinguished University Professor
University of Maryland, College Parkand
Johns Hopkins University Bloomberg School of Public Health
Water-related diseases
Amoebiasis
Arsenic
Diarrhoeal disease,Including cholera
Dracunuliasis (guinea worm)
Fluorosis
Giardiasis
Hepatitis A
Intestinal helminths
Malaria
Schistosomiasis
Trachoma
Typhoid
48,000,000
28-35m exposed to drinking water with elevated levels
1.5 billion
> 5000
26 million (China)
500,000
1,500,00
133,000,000
396,000,000
160,000,000
500,000,000
500,000
110,000
1,800,000
-
-
Low
-
9400
1,300,000
> 10,000
-
25,000
Cases per year Deaths per year
Background on Cholera: A Global Disease
• Acute water-related diarrheal disease
• Seventh pandemic started in 1960s
• Occurs in more than 50 countries affecting approximately 7 million people
• Bengal Delta is known as “native homeland” of cholera outbreaks
• Since cholera bacteria • exist naturally in aquatic habitats• evidence of new biotypes emerging, it is highly
unlikely that cholera will be eradicated but clearly can be controlled by provision of safe drinking water.
G. Constantin de Magny
Model for the Transmission of Vibrio Cholerae from the Environment to Humans
Work begins in Bangladesh in 1975
Dan Zimble, ESRI Inc.
West Virginia UniversityCivil and Environmental Engineering
Chronology of cholera and satellites
Colwell (1996)Colwell (1996)
1996: Colwell’s coastal vibrio hypotheses
2001: Lobitz’s chlorophyll-SST
Several attempts to link satellite data with cholera
Classification of coastal and Inland cholera
Cholera and SST in the Indian Ocean
0 0.6+
R2
Six-month SST lead: R2 = 0.72
Lobitz et al., 2000, PNAS Vol. 97, No. 4 pp. 1438-1443
Vezzulli et al. Proc. Natl. Acad. Sci. USA 2016 113 (34) E5062-E5071 doi:10.1073/pnas.1609157113
Vezzulli et al. Proc. Natl. Acad. Sci. USA 2016 113 (34) E5062-E5071 doi:10.1073/pnas.1609157113
Antarpreet Jutla, Elizabeth Whitcombe, Nur Hasan, Bradd Haley, Ali Akanda, Anwar Huq, Munir Alam, R. Bradley Sack and Rita Colwell. 2013. Environmental factors influencing epidemic cholera. Amer J Trop Med Hyg 89(3) 597-607
Could we have predicted the Haiti Cholera outbreak?
Recent cholera outbreak in Haiti indicated the disease remains a global threat.
Framework for developing cholera prediction models in cholera endemic (ER) and non-endemic regions (NER)
The sharp contrast in mortality rates between ER and NER exists not because we do not know how to treat cholera patients, but because of a persistent “knowledge barrier” between ER and NER.
We propose a pragmatic and adaptive framework which hypothesizes that convergence of three enabling situations - Inception, Environmental Conditions, and Transmission - are necessary for a cholera outbreak to become an epidemic.
050
100150200250300350400450
Jan
Feb
Mar
Apr
May Jun
Jul
Aug
Sep Oct
Nov
Dec
Pre
cipi
tatio
n (m
m/m
onth
)
Average 2010
Air temperature in Haiti in 2010 and average of air temperature data for last 50 years.
Monthly rainfall in Haiti in 2010 compared with historical rainfall data.
Jutla, A.S., Whitcombe, E, Hasan, H., Haley, B., Akanda, A., Huq, A., Alam, M., Sack, B., Colwell, R.2013. Environmental factors influencing epidemic cholera. American Journal of Tropical Medicineand Hygiene, 89(3):597-607.
Epidemic cholera model
Source: The Institute for Genomic Research
Vibrio cholerae
Sequenced and published in 2000
Chromosome I (2,961,149 bp, 2,742 ORFs)
RC9 O1 Ctx+ Kenya 2740-80 O1 ElTor Env USNCTC8457 O1 ElTor 1910B33 O1 ElTor MzbMJ-1236 O1 hybrid biotype MO10 O139623-391587 O12VL426 albensisO395 O1 ClassicalAM-19226 O39 TMA21 non-O1 BrazilMAK757 O1 ElTor 1937 MZO-2 O14 MZO-3 O37 RC385 O135 Csp BayV51 O141 USV52 O37 Sudan
Chromosome II (1,072,315 bp, 1,093 ORFs)
Missing ORFs in V. cholerae strains (Reference: N16961; cutoff = 70% DNA similarity)
RC9 O1 Ctx+ Kenya 2740-80 O1 ElTor Env USNCTC8457 O1 ElTor 1910B33 O1 ElTor MzbMJ-1236 O1 hybrid biotype MO10 O139623-391587 O12VL426 albensisO395 O1 ClassicalAM-19226 O39 TMA21 non-O1 BrazilMAK757 O1 ElTor 1937 MZO-2 O14 MZO-3 O37 RC385 O135 Csp BayV51 O141 USV52 O37 Sudan
Small Chromosome Large Chromosome
Chun et al. Proc. Natl. Acad. Sci. USA 2009 106(36):15442-15447 doi:10.1073/pnas.0907787106
0.00002
HC61A1
HC81A1
HC41A1HC7A1
HC80A1
HC46A1HC17A2HC22A1
HC32A1HC68A1
HCUF01HC48A1HC20A2
2010-EL17862010-EL1792HC38A1HC49A2
HC21A1HFU022010-EL1798HC40A1HC70A1
HC42A1HC57A2HC47A1
CP1038
CP1048CP1050
CIRS 101
CP1042
CP1040CP1041
MO10MJ-1236
B33CP1032
CP1033RC9
INDRE 91/1N16961
HC43A1HC23A1HC06A1
(c)7th pandemic (c)
HC-2 (b)
NCTC8457BX330286
M66-2MAK7572740-80V52O395RC27
LMA3984-412129(1)
MZO-3AM-19226
HE-25TMA21
623-391587MZO-2
HE-40HE-46HE-39
HE-48HC-44C1HC-46B1HC-43B1HC-41B1HE-45
AmazoniaTM11079-80
V51CT5369-93
RC385VL426
HE-09HE-16
0.002
(a) (b)HC-51A1
HC-02C1HC-1A2HC-2A1HC-36A1HC-50A1HC-55B2HC-55C2HC-57A1HC-59A1HC-59B1HC-60A1HC-61A2HC-78A1
HC-56A1HC-52A1
HC-55A1
0.0000005
HC-1
Haiti, 2010
Bangladesh, 2010
Zimbabwe, 2009
Bangladesh, 2002
Thailand, 2010
Zambia, 2004
Mexico, 1991Mexico, 2000
Phyl
ogen
omic
sThe Haitian V. cholerae O1 strains clustered with other 7th pandemic V. cholerae strains in a single monophyletic clade
Haitian strains branching separately from South Asian V. cholerae strains
10 Haitian strains (red) form a cluster cloud, distinct and yet, distant, from CP genomes (concurrent epidemic isolates form different parts of the world) (blue) and others (green).
Interestingly, one reference strain CP 1038 (from Zambia) genome falls into the Haitian cluster.
Principal component analysis:The three-dimensional PCA projection plots based on divergence of average nucleotide identity
Haitian Cluster Cloud
Microbiome based Diagnostics
Microbiome Analysis for Rapid Detection of Enteric Pathogens and Characterization of the Intestinal Microbiome in Health and Disease
COSM
OSI
D G
ENIU
S® A
NAL
YSIS
SER
VICE
Microbes Extracted DNA DNA Sequencing
GENEBOOK® Library Raw Sequence ReadsGENIUS®: Probabilistic Matching
Identified BacteriaGENEBOOK® Antibiotic Resistance &
Virulence Factor Library
TetR
CIPR
mecActxA
Pathogen detection for health and wellness;
food safety and probiotics biothreat and public health
surveillance; forensic and scientific
investigation
Applications:
RESE
ARCH
& T
ESTI
NG
LAB
: SE
QU
ENCI
NG
SER
VICE
CosmosID provides analysis for shotgun metagenomics
Patented Methodologies1. U.S. Patent No. 8,478,544, Issued: July 2, 2013. Direct Identification and Measurement of Relative Populations of
Microorganisms with Direct DNA Sequencing and Probabilistic Methods2. U.S. Patent No. 12/276,037, Issued Jul 8, 2014. A Method and System for Genome Identification3. U.S. Patent Application No. 13/836,139, Filed: March 15, 2013. Characterization of Biological Material in a Sample or
Isolate Using Unassembled Sequence Information, Probabilistic Methods and Trait-Specific Database Catalogs
Microbial ID using NGS Data
GenBook®: Expert-curated
genome databases
Highest performance
bioinformatics algorithms
Graphical User Interface
CosmosID Value Proposition
CosmosID: Next Generation Sequencing
Based Solution for Microbial Identification and Pathogen Characterization
Incredibly fast output – in minutes
(post sequencing)
Expert-curated proprietary genome databases (GenBook®)
High Sensitivity & High Specificity
No A priori assumptions necessary
Polymicrobialidentification from a single
test
Sub-species and Strain level identification,
including abundance
Credentials / On-going Collaborations
Client/Entity Engagement Overview
• 5 Year contract as in the in-house metagenomics solution after an in depth two year bake off
• NIST & CosmosID Metagenomics MVP Challenge v1.0
• BaseSpace Import Integration• #1 all time attended webinar through the end of 2016
• Partnering as CLC’s Metagenomics Solution
• Partnering as Thermofisher Cloud’s Metagenomics Solution
• Nominated to the White House “Cancer Moon Shot” by Vice President Joe Biden
• Working with Big and Small Pharma companies on discovery R&D, clinical trials and quality control
Diarrheal Disease Study: Infectious Disease in the Genomics Era
Diarrheal Disease Study
National Institute of Cholera and Enteric DiseaseKolkata, India
www.niced.org.ind
Total # of NICED samples: 74Indian Healthy Control (HC): 20Sick with Unknown Etiology (UE): 28Sick with Known Etiology (KE): 26Healthy Human Microbiome Project (HMP): 20
BacteriaVibrio choleraeVibrio parahaemolyticusVibrio fluvialisAeromonas spp.Campylobacter jejuniCampylobacter coliShigellaSalmonellaEscherichia coli
VirusesRotavirusAdenovirusNorovirusSapovirusAstrovirus
ParasitesGiardia lambliaCryptosporidium parvumEntamoeba histolyticaBlastocystis hominis
Enteric Pathogens Monitored By NICED(using non-metagenomic methods)
Diarrheal Disease Study
In collaboration with the National Institute of Cholera and Enteric Disease(NICED), Calcutta, India
γ- Proteobacteria – Firmicutes - Bacteroidetes
(HEALTHY)
DIA
RR
HEA
L PA
TIEN
TS
HEA
LTH
Y IN
DIV
IDU
ALS
Microbiomes of Diarrheal Subjects Compared to Healthy Subjects
Many Pathogens Readily Identified From Diseased Patients
Healthy Control
Human Microbiome Project
Diseased, Known Etiology
Diseased, Unknown Etiology
Individuals with AMR genes present in their microbiome
beta.lactamase
tetracycline
sulphonamide
rifampicin
quinolone
fosfomycin
nitroimidazole
phenicol
macrolide
trimethoprim
aminoglycoside
Unknown Etiology
Known Etiology
Healthy or Asymptomatic Control
0
4.8
9.6
15
Genes which match at > 50% coverageHMP samples had no genes present which matched at this level of coverage
• Microbial communities of healthy volunteers suggest that the microbiome of healthy humans of Indian descent is markedly different than those of Western European descent.
• Indian population may tolerate low number of pathogenic microorganisms that may indicate a “disease state” for Western European descent
• Multiple pathogens can readily be identified from disease patients
• Metadata revealed that patients who exhibited profound watery diarrhea contained in their microbiome pathogens primarily of the Escherichia coli complex, namely pathogenic E. coli and Shigella species.
• The microbial community of the Indian population encodes higher rates of antibiotic resistance genes when compared to healthy HMP samples
• Functional analysis of the Indian microbiome indicates predominance of carbohydrate metabolism genes
Summary
Infectious Disease in the Genomics Era
Necrotizing fasciitis
Neighbor-joining tree using 2,514 conserved full length predicted proteins
Polymicrobial Infection: Necrotizing Fasciitis
Relative distribution of four Aeromonas hydrophila strains NF1, 2, 3, and 4 into different metagenomic datasets derived from muscle, spleen and liver samples
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Mus
cle
1
Mus
cle
2
Mus
cle
4
Mus
cle
35
Sple
en 4
5
Sple
en 1
23
Live
r 1
Live
r 2
Live
r 3
Live
r 4
Live
r 5
Rel
ativ
e ab
unda
nce
of N
F st
rain
s
Aeromonas hydrophilaNF4
Aeromonas hydrophilaNF3
Aeromonas hydrophilaNF2
Aeromonas hydrophilaNF1
Strain level ID shows selective dissemination
Strain interaction in mixed infection
Dissemination to peripheral
organs
Dissemination to peripheral
organs
Dissemination to peripheral
organs
• 15 Klebsiella pneumoniae carbapenemase (blaKPC)-Aeromonas sp. and Enterobacteriaceae clinical isolates were collected between 2/2012-5/2013
• These were paired with 32 sink trap biofilm samples collected from the same hospital in 3/2014.
• In addition, the 15 isolates and 9 of the sink trap biofilms were sequenced.
• Metagenomic WGS data was analyzed by CosmosID.
• To determine if the patient isolates were present in any of the biofilm samples, their genomes were assembled and added to the ~31,000 genomes already in the CosmosID bacterial database.
• The biofilm WGS metagenomes were then analyzed using this supplemented database.
• Both the sink trap biofilm samples and the patient isolates were also profiled using the CosmosID antimicrobial resistance and virulence gene databases.
Collaboration with Amy J. Mathers, UVA
Hospital Biofilms: Source Tracking
• An edge connects an isolate (in red) to a sink trap biofilm (in blue), if the isolate is present in the biofilm.
• An edge connects 2 isolates if they were both found in the same sink trap sample.
• The size of the isolate vertices represents the average abundance in the biofilm samples.
Co-occurency graph of patient isolates and sink trap biofilms based on metagenomic analysis
Patient isolate (red)Sink trap biofilm (blue)
Patient Isolates Present in Sink Trap Biofilms
• Except for 3 genes, all genes found in common between the isolates and the biofilms are plasmid-associated. • There are 2 distinct clusters of biofilms and isolates (circled).
Accessory Gene Profiling for Biofilm Samples and Patient Isolates
Plasmid AssociatedNot Plasmid Associated
Antibiotic Resistance and Virulence Genes
The Microbiome as Therapy
Pre and Post-Fecal Microbiota Transplant (FMT) Microbial Shift
Pre-FMT Post-FMT
✪ Strain associated with disease
✪
37
Protection of Microbiome from Antibiotic Induced Change
Prevented overgrowth of organisms known for secondary infection
Orange County Water District Study
The influent, secondary treated municipal wastewater of the AWPF treatment train ispurified by a three-step process: microfiltration, reverse osmosis, and ultraviolet (UV) light with hydrogen peroxide. Initially the wastewater is screened at 4mm Sodium hypochlorite as disinfectant is added prior to microfiltration. Hydrogen peroxide (H2O2) is added before UV treatment. The decarbonators remove CO2 and raise the pH; addition of lime further stabilizes the purified water.
Metagenomics and Public Health
SodiumHypochlorite
influent
Filter Screens
Microfiltration (MF) process
Reverse Osmosis (RO)Process
H2O2
AWPF final product water
UltravioletIrradiation System
Decarbonators
Add’n Lime
Relative abundance and diversity of parasite DNA in MF-biofilm and Q1-water. The approximate relative abundance heat map was simplified, using the GENIUS bioinformatics algorithm and curated databases. The 99 relative abundance corresponds to sequences classified as Paramecium biaurelia strain v14, Thalassiosira, and Acanthamobea polyphaga based on observed frequency of DNA sequences identified. Parasite sequences were not found in the RO-biofilms.
Acanthamobeapolyphaga Thalassiosira
Paramecium biaureliastrain v14
Q1-water
MF-biofilm
RO-biofilm
99
30
9
2
0
RO-biofilm
Adeno-associated
MF-biofilm
Q1-water
Pseudomonas
Klebsiella
Salmonella
Aeromonas
Burkholderia
Escherichia
Yersinia
Enterobacteria
Vibrio
Mycobacterium
Shigella
Stx1 Converting
Erwinia
Bacteroides
Lactococcus
Enterobacteria
Streptococcus
bacteriophage
virus
N=15
N=10N=4
N=0
Viral DNA 1709
Stx2 Converting
Roseophage
Yersinia
Escherichia
Lactococcus prophage
Phage-cdtl
Viral DNA 236
Salmonella
Enterococcus
Viral DNA 1241
Virus and bacteriophage DNA sequences comparison demonstrate the presence of bacteriophages and virus DNA in the membrane filter (MF)-biofilm and in the influent water, Q1. Note, the absence of bacteriophages and DNA viruses in the reverse osmosis (RO)-biofilm. Presence and absences of sequences (partial or complete) related to bacteriophage and viruses in MF-biofilm were compared to the Q1 water.
Comparison of virus and bacteriophage DNA sequences
A Simple, Sustainable Method for
Reducing Cholera
A Simple Solution for Cholera Prevention: Sari Filtration
Full Study
0.000.20
0.400.60
0.801.00
1.201.40
Control Sari Nylon
Test Group
Cas
es o
f Cho
lera
Per
100
0 Po
pula
tionFull Study
Control NylonSari
Test Group
Cas
es o
f Cho
lera
Per
100
0 Po
pula
tion
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
“When one tugs at a single thing in nature, he finds it hitched to the rest of the universe.”
John Muir(1838-1914)
Collaborators and Colleagues
Anwar Huq, ProfessorUniversity of Maryland,College Park, MD
Dr. Nur HasanVice-President, Research and DevelopmentCosmosID, Inc.College Park, MD
Antarpreet Jutla, Assistant Professor, West Virginia UniversityMorgantown, WV
Dr. Seon Young Choi,
Bioinformatic Scientist, CosmosID Inc.College Park, MD
NICED:G. Balakrish NairThandavarayan Ramamurthy
University of Maryland:Shah M. RashedSeon Young ChoiAnwar Huq
CosmosID:Nur A. HasanPoorani SubramanianKelly MoffatHuai LiBrian FanelliManoj Dadlani
Orange County Water DistrictMenu Leddy
ConsultantJoseph Cotruvo
Amy J. Mathers, UVAAshok Chopra, University of Texas Medical BranchChris Grim, FDA
Genomics Team
Courtesy of GB Nair, NICED, Kolkata, India
Safe water is a global challenge