Get documents about "Slide 1 - Iowa Geological Survey - University of Iowa
Document Sample
The University of Iowa
Center for Global and
Regional Environmental
Research
Seed Grant 2002-2003
Principal Investigators:
Mary Skopec, IDNR Geological Survey
Nancy Hall, University Hygienic Laboratory
Karen Owens, University Hygienic Laboratory
Bacteria Source Tracking in the
Upper Iowa River Watershed
• What we know about bacteria in Iowa
streams?
• What is bacteria source tracking?
• Why the Upper Iowa Watershed?
• What is E. coli bacteria?
• What is DNA Ribotyping?
• Results
• Discussion
What do we know about
bacteria in streams?
• Bacteria levels are highly
variable
• Rainfall affects bacteria
levels
• Bacteria levels vary
seasonally
• Sources of bacteria likely
vary seasonally
• Many potential sources of
bacteria in a watershed
Sources of fecal
material in water
• Leaking sewage lagoons
• Malfunctioning septic systems
• Sewage treatment plant discharges
• Dirty diapers
• Boating or swimming fecal “accidents”
• Overflowing manure lagoons
• Manure spills
• Runoff from fields after manure application
• Storm water runoff from lands with wildlife or pet
droppings
• Fecal material expelled by animals standing in the
water
• Swimmers
What is bacteria source
tracking?
• Using bacterial method to
determine sources of fecal bacteria
in the environment
• Different methods available
– Genotypic methods
– Phenotypic methods
Why is bacteria source
tracking needed in Iowa?
• Better target BMPs for watershed
projects – to reduce a bacteria
problem, need to know where it’s
coming from
• Improve remediation efforts at state
beaches
Why the Upper Iowa
watershed?
• Site of an active
watershed group
Why the Upper Iowa
watershed (cont.)?
• Elevated
bacteria
identified as
water quality
concern
Why the Upper Iowa
watershed (cont.)?
• Interest locally
in identifying
bacteria sources
Upper Iowa Watershed
• 1,005 mi2
watershed
– NE Iowa and SE
Minnesota
• Great recreational
value
Upper Iowa River Watershed
N
Coldwater Creek (#9) W E
Ê
Ú
Ú Ú
Ê Ê
Ú
Ê S
Silver Creek near Cresco Silver Creek near Waukon (#27)
(#8 and 801)
Ú
Ê Sampling Points
Upper Iowa River Watershed
Rivers
Sub Watersheds
20 0 20 40 Miles
Sample Collection
• Water Samples
– Weekly samples
from each
watershed
• Fecal Samples
– Collected from
known sources
Water E. coli Isolates
• 50 E. coli samples
– Isolated from: #
• Silver Creek 27 - 12
• Silver Creek 8 - 13
• Silver Creek 801 - 14
• Cold Water Creek - 11
Fecal Sample Collection
• Most samples were
collected in spring of
2003
Cattle 103 isolates
Human 55 isolates
Geese 29 isolates
Swine 26 isolates
Deer 36 isolates
Sheep 6 isolates
Raccoon 4 isolates
Escherichia coli (E.coli)
• Common inhabitant of human and animal
intestines
• Predominant fecal coliform bacterium
• Indicator of fecal pollution
• Presence indicates disease-producing
organisms may be present
• Presence does not determine source
E. coli Methods - Isolation
Ribotyping of E.coli
• Automatically
generates genetic
fingerprint
• Useful epi tool for
tracking outbreaks
• Useful tool for ident.
human and non-
human pollution
UHL’s Riboprinter
Ribotyping Process
• Purification
• Identification
• Harvesting
Ribotyping Process
E. coli cells are lysed
…releasing DNA
Ribotyping Process
DNA is cut into fragments
using special restriction
enzymes
Fragments separated by
size through electrophoresis
Ribotyping Process
• Fragment pattern is
transferred to
membrane, mixed
with DNA probe and
chemiluminescent
chemicals to produce
a visible band
pattern
Data Normalization
“Sample lanes”
Algorithms
RiboPrint® patterns
for 8 lanes of data
Each unique pattern is assigned a unique designation
Different RiboPrint Patterns for 4 Hog and 4 Cow Isolates
Ribotyping - Source Tracking
First - Building the Libraries
• Known E.coli riboprint patterns from
different species from the Upper Iowa
• Import these patterns into BioNumerics
• Patterns grouped into various libraries
• Perform band matching analysis
• Statistics:
– Cluster verification (Jackknife test)
– Discriminate analysis
Analysis by BioNumerics Software
(Applied Math, Belgium)
• Integrated software package
• Relational database with analysis and
clustering modules
• UHL currently has software (PulseNet)
• Library development
• Database sharing capabilities
Band Assignment and Quantification
Cluster Verification for 5 Groups
Swine Cattle Deer Geese Humans
Swine 63.64 6.76 0.00 12.50 0.00
Cattle 27.27 81.08 34.62 31.25 11.11
Deer 0.00 4.05 65.38 0.00 0.00
Geese 9.09 4.05 0.00 50.00 3.70
Humans 0.00 4.05 0.00 6.25 85.19
ARCC = 69% p < 0.001, <0.001, 15.843, 74.75
Cluster Verification for 3 Groups (CAH)
Animals Cattle Humans
Animals 76.56 14.86 3.70
Cattle 21.88 81.08 11.11
Humans 1.56 4.05 85.19
ARCC= 81% p < 0.001, 0.023
Cluster Verification for 2 Groups
Animals Humans
Animals 97.10 14.81
Humans 2.90 85.19
ARCC = 91% p<0.001
Recent Published Ribotyping Studies
Investigator Species Correct Database
Classification size
Tseng, 2001 Human 98%(ave.94%) 160
Animal (3) 92%
Carson, 2001 Human 95%(ave.97%) 287
Animal (7) 99%
8 sources 74%
grouped
Skopec/Hall Human 85% (ave.91%) 173
2003 Animal (4) 97%
5 sources 69%
grouped
Ribotyping - Source Tracking
Second – Unknown Identification
• Compare unknown E.coli patterns with
known E.coli pattern groups to
determine probable source
• Statistics:
– Curve-based Pearson Correlation
– Calculation of Quality quotient
Ribotyping - Source Tracking
Second – Unknown Identification
Criteria for Good Identification:
Similarity coefficient: >90%
(linear relationship between 2 entries)
Quality quotient/factor:
High probability is A or B
(how well it fits in the group, taking into
consideration the internal spread)
Interpretation Guidelines
• Interpretation based solely on the match
– Not quantitative - # identifications not
proportional to source contribution
– Sampling bias/small sample size/both in
E.coli and sample number
(As total # of E.coli in sample , the probability of
identifying all waste sources )
27 Silver Creek
Cow Human Animal Unknow n
3
# of Identifications
2
1
0
Fall Winter Spring Summer
Season
E.coli Results: 510 ave. 27 770 ave.
per season (3 samples, (1 sample, (3 samples,
5 isolates -1%) 2 isolates-7%) 5 isolates-<1%)
8 Silver Creek
Cow Human Animal Unknow n
3
# of Identifications
2
1
0
Fall Winter Spring Summer
Season
E.coli Results: 450 ave. 50 ave. 1000 ave. 9200
per season (3 samples, (1 sample, (3 samples) (1 sample)
4 isolates-<1%) 2 isolates-4%)
801 Silver Creek
Cow Human Animal Unknow n
# of Identifications
3
2
1
0
Fall Winter Spring Summer
Season
E.coli Results: 9100 ave. 100 400 ave. 750
per season (2 samples) (1 sample, (2 samples) (1 sample)
2 isolates-2%)
9 Cold Water Creek
Cow Human Animal Unknow n
# of Identifications
3
2
1
0
Fall Winter Spring Summer
Season
E.coli Results: 27 ave. 4600 ave. 150
per season (3 samples, (2 samples) (1 sample)
6 isolates-22%)
Observations
• Sources identified
matched watershed
surveys
– Cattle in streams
– Failing private septic
systems
Photos by Pat Kambesis, W. Kentucky University
In Summary
• There’s still no magic bullet
• All source tracking methods have their
strengths and limitations
• All source tracking methods need better
quantitative criteria
• These methods continue to evolve and
look very promising to differentiate
human and animal pollution sources
In Summary (continued)
• “Toolbox” approach advocated
– watershed evaluation
– key monitoring parameters
– Snap-shot sampling (HOT SPOTS)
– strategic monitoring sites
– Use >1 tracking tools (e.g. microbial &
chemical)
Future Studies
• Lake Darling
– Implementing “Toolbox” approach
– Smaller watershed
– Limited number of sources
– Collection over longer period of time
• Will Upper Iowa database be valid in
another geographic area a year later?
Presentation and Report
available at:
ftp://ftp.igsb.uiowa.edu/pub/Download/EOBrien/SourceTracking/
or
http://wqm.igsb.uiowa.edu/publications/presentations/presentations.htm
