Variability of molecular assays Quality-control samples were collected and analyzed for cyanobacterial genes. Fourteen field concurrent replicates were collected at essentially the same time to represent between-bottle variability. Six of those replicates were also processed as split replicates, in which two aliquots were filtered from each bottle to represent within-bottle variability. Laboratory replicates were extracted filters that were analyzed twice by qPCR or qRT-PCR. The absolute value log10 differences (AVLD) were calculated for all replicate pairs and the Kruskal-Wallis with Dunn’s test was used to identify differences among groups of variability (A’s and B’s in the graphs). Anabaena mcyE DNA and RNA assay results are not presented because these genes were not detected in any samples. Lab variability was small and statistically lower than within- or between-bottle variability for the five DNA assays but not for the two RNA assays. Sampling variability was small as compared to filtering, processing, and matrix variability.
USING MOLECULAR ASSAYS AND OTHER VARIABLES TO PREDICT HARMFUL CYANOBACTERIA BLOOMS IN FRESHWATER LAKES
D.S. Francy1, E. A. Stelzer1, C.D. Ecker1, J.L. Graham2, K.A. Loftin2, P. Struffolino3, Brady, A.M.G.1, and R.N. Bushon*1 1USGS Ohio Water Science Center, Columbus, OH; 2USGS Kansas Water Science Center, Lawrence, KS; 3University of Toledo, Oregon, OH
Conclusions The results of this study identified the water-quality and environmental variables and the cyanobacterial gene assay results that were statistically related to microcystin concentrations at four very different Ohio recreational sites. The significant variables are promising for use in site-specific predictive models, both to provide real-time swimming advisories to the public (daily predictions) or to provide advanced warning of the potential for a cyanoHAB (long-term predictions). In order to develop accurate models to predict toxin concentrations at freshwater lake sites, data need to be collected more frequently and for consecutive days in future studies.
Harmful cyanobacterial blooms (cyanoHABs) and associated toxins, such as microcystin, have been identified at Ohio Lake Erie and inland lake waters. Predicting when and where a bloom may occur is important to protect the public that uses and consumes a water resource; however, because of the many factors affecting toxin production, predictions are complicated and likely site specific. Monitoring for cyanobacteria using molecular assays such qPCR (for DNA) or qRT-PCR (for RNA) and for chlorophyll or phycocyanin using optical sensors may provide an early warning system for cyanoHABs. Samples were collected at Ohio recreational sites during May–November, 2013-14. Field crews measured physical parameters at the time of sampling. Composite samples were preserved and analyzed for dissolved and total nutrients, cyanotoxins, phytoplankton abundance, and cyanobacterial genes by qPCR or qRT-PCR. To assess sampling and analytical variability for the molecular assays, concurrent field replicates were collected and qPCR lab replicates were processed. Lab variability was small and statistically lower than within- or between-bottle variability for the five DNA assays, but not for the two RNA assays. Sampling variability was small as compared to filtering, processing, and matrix variability. The molecular assay results that were most significantly correlated to microcystin concentrations were Planktothrix mcyE DNA genes at a boater-swim site at a shallow canal-lake (Buckeye Lake) and Microcystis mcyE DNA genes at a beach at a water-supply reservoir (Harsha Lake) and at a Lake Erie beach (Maumee Bay State Park). Microsystis mcyE RNA transcripts were significantly correlated to microcystin concentrations at only the Lake Erie beach. Phycocyanin, nutrient concentrations, pH, lake-level change, and turbidity were among the variables significantly related to microcystin concentrations. Although results from this study showed that the use of molecular assays, sensor measurements, and other variables to provide an early warning of cyanoHABs is promising, data need to be collected more frequently and for consecutive days in order to develop accurate models to predict toxin concentrations at freshwater lake sites.
Abstract
Surrogate Recovery Control Competitive Internal Positive Control (CIPC)
OW
ML
Met
hod
Dat
a In
terp
reta
tion
Des
crip
tion
Sampling sites and frequency: Monthly at 8 beach sites during 2013 and weekly at 5 recreational sites at 3 lakes during 2014.
Sampling procedures: Several subsample locations within the swimming area were composited using sterilized glass bottles. Physical parameters were measured at each subsample location using a water-quality sonde. Samples were processed and preserved for laboratory analyses on site.
Analyses. Toxins (by ELISA), nutrients, phytoplankton, and cyanobacterial genes. Gene assays by qPCR or qRT-PCR included: (1) general cyanobacteria, (2) genus-specific for Microcystis and Anabaena, (3) genus-specific mcyE DNA toxin genes for Microcystis, Anabaena, and Planktothrix, and (4) genus-specific mcyE RNA transcripts for the 3 genera.
Environmental data were compiled from other sources. •Easily- or continuously-measured water-quality and environmental data were used for daily predictions that did not require a site visit. •Daily variables and results from samples collected and analyzed in a laboratory were used for long-term predictions.
Methods
Variability of replicates analyzed for general cyanobacteria, Microcystis and Anabaena by qPCR
Results and ConclusionsOther variables―correlations to
microcystin at Harsha Main beach
Variability of replicates analyzed for mcyE toxin genes and transcripts for Microcystis and Planktothrix
Molecular assays―correlations to microcystin at three sites
Spearman’s correlations to microcystin concentrations(ND, not determined because there were <5 detections;
results in bold were significant at p<0.05)
qPCR or qRT-PCR assayBuckeye LakeOnion Island
n=10
Harsha Mainn=17
Maumee Bayn=24
Cyanobacteria, general 0.71 0.57 0.49Microcystis, general 0.12 0.90 0.73Anabaena, general -0.16 0.90 0.41
Microcystis mcyE DNA -0.30 0.92 0.82Planktothrix mcyE DNA 0.73 0.09 -0.09Microcystis mcyE RNA ND ND 0.51Planktothrix mcyE RNA 0.51 ND ND
The suite of molecular assay results that were significantly correlated to microcystin concentrations was site specific. These assays can be used for long-term predictions. At Harsha Main, Planktothrix dominated the
cyanobacterial community in May and late October, Anabaena and Microcystis were substantial portions from late May through August, and non-microcystin producing taxa were dominant from late-July through early-October. Statistically-significant correlations (p<0.05) were found for 7 variables for daily predictions and 13 variables for long-term predictions.
Highest Spearman’s correlation to microcystin (n=17)(A = average of two or more values)
Daily predictions rho p
Phycocyanin, turbidity, and pH 0.73 ‒ 0.93 <0.0001
Algae category 0.72 0.0010Secchi depth -0.69 0.0022
Chlorophyll and water temperature 0.59 ‒0.63 A0.0095
Long-term predictions
Microcystis mcyE DNA, Microcystis or Anabaena qPCR 0.90 ‒ 0.92 <0.0001
Cyanobacteria BV; Microcystis, Anabaena, abundance or BV 0.63 ‒ 0.90 <0.0001
Cyanobacterial abundance, cyanobacteria qPCR 0.58 ‒ 0.63 A0.0116
Ammonia, nitrate + nitrite, orthophosphate -0.48 ‒ -0.53 A0.0400
Continuously-measured variables at Harsha Main beach
Continuous water-quality data from a USEPA-operated sonde, located 1 mile north of Harsha Main, can provide data for daily predictions. Statistically significant correlations (p<0.05) to microcystin concentrations were found for many manipulated variables.
Most significant manipulation for each variable rho p
Phycocyanin, 7-day average 0.98 <0.0001
Dissolved oxygen, 14-day average 0.88 <0.0001
pH, 7-day average 0.83 <0.0001
Temperature, instantaneous 10 a.m. 0.73 0.0031
Chlorophyll, 24-hour average 0.53 0.0358
Specific conductance, 3-day average -0.20 0.4473 0 5 10 15 20Phycocyanin, 7-day average, in RFU
10-1
100
101
102
Mic
rocy
stin
, in
µg/L
rho= 0.98, p<0.0001
5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0Planktothrix mcyE, in log copies/100 mL
0
10
20
30
40
50
60
70
Mic
rocy
stin
, in
µg/L
DNARNA
E.
Buckeye Onion Islandrho = 0.51, 0.73
E.
3 4 5 6 7 8Microcystis mcyE DNA, in log copies/100 mL
10-1
100
101
102
Mic
rocy
stin
, in
µg/L
*Open circles indicate < values of Microcystis mcyE DNA.
Harsha Mainrho = 0.92
F.
0 2 4 6 8 10Microcystis mcyE DNA, in log copies/100 mL
10-1
100
101
102
103
Mic
rocy
stin
, in
µg/L
*Open circles indicate < values of Microcystis mcyE DNA.
Maumee Bayrho = 0.82
0.0
0.5
1.0
1.5
2.0
AV
LD, c
opie
s/10
0 m
L
Cyanobacteria Microcystis Anabaena
35 11 28 34 11 28 23 11 22
A B B A B B A B B
0.0
0.5
1.0
1.5
2.0
AV
LD, c
opie
s/10
0 m
L
Micro mcyE DNA Plankto mcyE DNA Micro mcyE RNA Plankto mcyE RNA
33 11 28 30 11 26 15 6 14 10 3 8
A B B A B B A A A A A A
C.
0 1 2 3 4 5 6 7Chlorophyll, in RFU
10-1
100
101
102
Mic
rocy
stin
, in
µg/L
rho= 0.63, p=0.0065
Micro mcyE DNA Micro mcyE RNAPlank mcyE DNA Plank mcyE RNAMicrocystis AnabaenaCyanobacteria
Microcystis mcyE DNA, in log copies/100 mL Microcystis mcyE DNA, in log copies/100 mL
Planktothrix mcyE DNA, in log copies/100 mL
*Open circles indicate < values of Microcystis mcyE DNA*Open circles indicate < values of Microcystis mcyE DNA