1. 3/2/2014
Development of the Environmental Fate Simulator (EFS): A
tool for predicting the degradation pathways of organic
chemicals in groundwater aquifers
Process
Scientists:
Caroline Stevens
Said Hilal
Dalizza Colón
Jack Jones
Eric Weber
Ecosystems Research
Division
US Environmental
Protection
Athens, GA
Multi‐Media Modelers:
Gene Whelan
Justin Babendreier
Software
Engineers:
Kurt Wolfe
Rajbir Parmar
Mike Galvin
Mitch Pelton
(PNNL)
The EFS will be
publicly
available
1

2. 3/2/2014
What is the need for the
Environmental Fate Simulator?
The Problem:
Current tools available to EPA for
conducting exposure and health
(human and ecological) assessments
are not adequate:
• TSCA inventory :
− > 85,000 chemicals
− High quality data for < 2%
• New Chemicals (PMN Program):
− 20 to 30 new chemicals per
week
• FIFRA inventory:
− ~ 1,100 agrochemicals
− High quality pchem data for
nearly 100%
Our Response:
Development of the Environmental
Fate Simulator (EFS):
• High throughput computational
system for providing molecular
and environmental descriptors for
consumption by EF&T models
Requires:
Knowledge of the process science
controlling chemical fate and
transport
The ability to encode this
information into a readable format
Integration of existing
cheminformatics applications and
modeling software technologies
What is it that requires automation?
Exposure/Testing
Scenario:
Chemical Structure of
Parent Chemical
Reaction
Medium
The information required
to simulate this scenario:
What is needed to
automate this process:
• Knowledge of the process
science underlying
transformation pathways
Cheminformatics
applications for encoding
the process science
• Molecular descriptors
necessary for predicting
mobility and reaction rates
Access to physico‐chemical
calculators
Estimated Concentrations of • Environmental descriptors
necessary for predicting
the Parent Chemical and
reaction rates
Predicted Transformation
Products
• Parameritization of EF&T
models
Software for providing
access to data from
online databases
Software providing
seamless
parameritization of
EF&T models
2

3. 3/2/2014
The EFS represents the integration of the most robust process
science available with state‐of‐the‐art cheminformatics
application and modeling software technologies
Process
science
Java‐based
cheminformatics
applications
Modeling software technologies
developed through ERD‐Athens
Integrated Environmental
Modeling (IEM) Program
EFS
Cheminformatics:
the generation, storage, indexing and
search of information relating to chemical
structure and chemical processes
5
Example of an EFS Workflow
Chemical Editor (CE):
Provides options for
chemical entry
Reaction Pathway
Simulator (RPS):
Generates potential
transformation products
based on user‐specified
conditions
Structure‐based Database
(SBD):
populated with calculated
and measured physico‐
chemical properties of
parent and potential
transformation products
Earth Systems
Model: Data
Mining for
environmental
descriptors
Physicochemical
Properties Calculator
(PPC):
Molecular descriptors
for the parent chemical
and predicted
transformation products
Reaction Rate
Calculator:
Parameritization and
Execution of QSARs
and Algorithms
3

4. 3/2/2014
Tautomer Identification/distribution
4

5. 3/2/2014
MarvinSketch: Calculation of pKa values
The selection of the environmental
conditions will determine which reaction
libraries will be executed in the Reaction
Pathway Simulator
Reaction Libraries consisting of one‐
step reactions and reaction rules for
various transformation pathways:
Chemical Processes:
• Reduction
• Hydrolysis
• Photolysis
Biological Processes:
• Aerobic Biotransformation
• Anaerobic Biotransformation
5

6. 3/2/2014
UM‐Pathway Prediction System (UM‐PPS)
• Web‐based system for the prediction of microbial biotransformation
Database
(http://umbbd.ethz.ch)
Prediction System
(http://umbbd.ethz.ch/predict)
11
Encoding the Process Science
MarvinSketch: Translation of
chemical structures into a
readable code
SMILES String
O=N(=O)C1=CC=CC=C1
SMART Reaction String
O=N(=O)C1=CC=CC=C1>>NC1=CC=CC=C1
12
6

7. 3/2/2014
Development of
Reaction
Libraries based
on Chemical
Terms Language
Abiotic Reductions:
Data Sources:
• Peer‐reviewed
literature
• Registration data
submitted to EPA
Implementing the Reaction Libraries
Functional group transformation
based on execution of reaction
libraries
X
7

8. 3/2/2014
Encoding the Process Science
Product formation
based on the
execution of the
reduction library
Likelihood: Likely
Generation: 95%
Accumulation: 10%
15
Prototype EFS: Environmental Systems Model
Environmental Descriptor collection
for site‐specific assessments
8

9. 3/2/2014
Environmental Descriptor collection through
the executions of Data for Environmental
Modeling (D4EM):
an open source software system consisting of a
library of utilities that can be used to access,
retrieve and process model data automatically
from sources on the internet
Access the necessary databases for the
collection of the required environmental
descriptors (e.g., pH, aqueous Fe(II) and (DOC))
Identifying Predominant Chemical
Reductants Anaerobic Aquifers and
Sediments
Flow Path
Aquifer
Intrusion of Dissolved Organic Matter
Primary Redox Reactions
Aerobic
Nitrate
Reducing
Corg
CO2
Corg
O2
H2 O
NO3-
Manganese
Reducing
HCO3- Corg
N2
MnO2
Iron
Reducing
HCO3-
Corg
HCO3-
Mn2+
Fe(OH)3
Fe2+
Sulfate
Reducing
Corg
HCO3-
SO42-
H2S
Methanogenic
Corg
HCO3
CH4
Working Hypothesis: The reactivity of chemical reductants in natural sediments will
vary as a function of redox zonation as described by the dominant terminal electron
accepting processes (TEAPs)
9

10. 3/2/2014
Formation of Potential Chemical Reductants
as a Function of Redox Zonation
Chemical Reductants
Mineral Formation
Complexation
Redox (DOM)
O
2
+
Fe2+ + HCO32-
C
O
O
2
Fe +
3
Methanogenic
Fe
Sulfate Reducing
Redox Zones
Fe Reducing
Fe
+
FeCO3 + H+
O
Green Rust Formation
O
e , H+
[Fe2+Fe3+(OH)8+ [Cl nH2O][Fe42+Fe23+(OH)12]2+ [SO4 nH2O]
Surface
O
-
Solution Phase
O
OH
[Fe42+Fe23+(OH)12]2+ [CO3 nH2O]
O
Fe2+
+
HS-
FeS + So
FeS +
+ H2S
FeS2
SH
O
OCSPP Harmonized* Test Guidelines
Series 835 ‐ Fate, Transport and
Transformation Test Guidelines
*Harmonized OPPT, OPP and OECD Test guidelines
Group A — Laboratory Transport Test Guidelines
OH
H+
OH
Environmental conditions
can also be entered by the
user through selection of the
appropriate test OECD test
guideline
835.1230 - Adsorption/Desorption (Batch Equilibrium) (November 2008)
835.1240 - Leaching Studies (November 2008)
835.1410 - Laboratory Volatility (November 2008)
Group B — Laboratory Abiotic Transformation Test Guidelines
835.2120 - Hydrolysis (November 2008)
835.2130 - Hydrolysis as a Function of pH and Temperature (January 1998)
835.2210 - Direct Photolysis Rate in Water by Sunlight (January 1998))
835.2240 - Photodegradation in Water (November 2008)
835.2410 - Photodegradation in Soil (November 2008)
835.Weber- Reduction
Group C — Laboratory Biological Transformation Test Guidelines
Group D —Transformation in Water and Soil Test Guidelines
835.4100 - Aerobic Soil Metabolism / 835.4200 – Anaerobic Soil Metabolism (October 2008)
835.4300 - Aerobic Aquatic Metabolism / 835.4400 – Anaerobic Aquatic Metabolism (October
2008)
Group E — Transformation Chemical-Specific Test Guidelines
835.5045 - Modified SCAS Test for Insoluble and Volatile Chemicals (January 1998)
835.5154 - Anaerobic Biodegradation in the Subsurface (January 1998)
835.5270 - Indirect Photolysis Screening Test: Sunlight Photolysis in Waters Containing
Dissolved Humic Substances (January 1998)
10

11. 3/2/2014
Prototype EFS: Physico‐Chemical Properties Calculator
The number of required calculated
data for a given physico‐chemical
property is based on its intended use
Physico‐Chemical Properties Calculator
3-nitro-5-oxo-1,4dihydro-1,2,4triazol-1-ide
Chemical Specific
Parameters
Abbrev
Units
Molecular Weight
Melting Point
MW
MP
g/mole
oC
Boiling Point
Water Solubility
Vapor Pressure
Molecular
diffusivity in water
Ionization constant
Henry’s Law
Constant
Octanol Water
Partition
Coefficient
Organic Carbon
Partition
Coefficient
Distribution
Coeffecient
(pH dependent
BP
WS
VP
oC
mg/L
torr
Measured
Goal:
•Provide complete
coverage
•Consensus approach
EPI Suite
– Fragment based
major species
at pH 7.5
Calculated
(EPI
Suite)
Calculated
(SPARC)
Calculated
(ChemAxon)
Calculated
(QSAR)
SPARC
– Mechanistic based
ChemAxon
– Atom based
cm2/sec
pKa
unitless
Atm
m3/mole
Kow
mL/g
Koc
mL/g
KD
Available
Not Available
Chemical Specific
mL/g
22
11

12. 3/2/2014
Calculation of P‐Chem Data Base Based on Consensus Approach
ChemAx
ChemAx
ChemAx
Braekevelt et al
AVERAGE (2003)
calculated calculated measured KLOP
log Kow log Kow log Kow log Kow
PHYS
log Kow
VG
log Kow
calculated measured
log Kow
SPARC
Name
PBDE‐28
PBDE‐47
PBDE‐66
PBDE‐85
PBDE‐99
PBDE‐100
PBDE‐138
PBDE‐153
PBDE‐154
PBDE‐183
PBDE‐209
6.46
7.14
7.22
7.96
7.92
7.95
8.74
8.71
8.73
9.52
12.01
4.638
Calculated log Kow
SSE =
EPIsuite
EPIsuite
5.88
6.77
6.77
7.66
7.66
7.66
8.55
8.55
8.55
9.44
12.11
‐‐‐‐
‐‐‐‐
‐‐‐‐
‐‐‐‐
6.84
‐‐‐‐
‐‐‐‐
‐‐‐‐
‐‐‐‐
‐‐‐‐
‐‐‐‐
5.97
6.76
6.76
7.54
7.54
7.54
8.32
8.32
8.32
9.10
11.45
2.706
1.297
5.51
6.25
6.25
6.98
6.98
6.98
7.71
7.71
7.71
8.44
10.64
0.915
10.0
9.5
9.0
8.5
8.0
7.5
7.0
6.5
6.0
5.5
5.0
5.85
6.64
6.64
7.43
7.43
7.43
8.23
8.23
8.23
9.02
11.39
0.923
5.94
6.71
6.73
7.51
7.51
7.51
8.31
8.30
8.31
9.10
11.52
• Structure
Searching
• Data
Analysis
5.94
6.81
7.37
7.32
7.24
7.90
7.82
8.27
1.237
Provide structure
SPARC
EPIsuite
ChemAxon KLOP
ChemAxon PHYS
ChemAxon VG
y = x
5.0
6.0
7.0
8.0
9.0
Measured log Kow
Calculation of P‐Chem Data Base Based on Consensus Approach
Compound class
KOWWIN SPARC
VG
KLOP
PHYS
ALOGP XLOGP2 XLOGP3‐AA
PBDEs
0.58
0.76
0.34
0.40
0.34
0.25
0.38
0.39
Phthalate esters
0.78
0.40
0.48
0.79
0.54
0.53
1.17
0.79
PCBs
0.76
0.87
0.57
0.72
0.71
0.73
0.77
0.65
Fused ring
structures
0.29
0.41
0.74
0.85
0.93
1.24
0.36
0.37
Others
0.31
0.86
1.51
0.87
0.61
1.19
1.32
1.09
ALL
0.58
0.74
0.94
0.75
0.64
0.90
0.96
0.78
Root mean square error (RMSE) for log Kow calculated by selected models
Results of Consensus Approach for poorly soluble chemicals
12

13. 3/2/2014
Reaction Rate Calculator:
Parameritization and Execution of
QSARs and Algorithms
Ability to populate and
execute QSARs for
calculating rate constants
2.15
2.98
5.71
DNAN
3.03
QSAR based on irreversible sorption of mono‐
substituted anilines in aerobic sediment
Correcting for environmental
conditions
Reaction Rate
Calculator:
Parameritization and
Execution of QSARs
and Algorithms
Temperature:
k = Ae
−
Ea
RT
where A is the frequency factor or pre‐
exponential factor and Ea is the activation
energy (Default value for Ea = 50 kJ/mol)
Sorption:
kapp =
k
(1 + ρ K d )
where k is the first‐order rate constant for
transformation in the aqueous phase, (Kd) is the
sorption coefficient and ρ is the solid‐to‐
solution ratio
Ionization :
1
⎛
K d ,app = ⎜
pH − pK a
⎝ 1 + 10
⎛ 10 pH − pKa
⎞
⎟ K d , HA + ⎜
pH − pK a
⎠
⎝ 1 + 10
⎞
⎟ K d , A−
⎠
where pKa is the negative of the logarithm
of the acid dissociation constant for the
chemical
13

14. 3/2/2014
Required Hallmarks of the EFS:
Vibrant
– Representing the most current process science and software
technologies available
Transparent
– Presentation of the meta data
High Throughput capability
– Relatively short run times
– Allows for operation in batch mode
Accessible
– Web‐based
Usable
– Reasonable run times
– User friendly
Flexible
– Customized for the user’s need
Quality Controlled
– Based on peer‐reviewed science
14