Exposure-based Chemical Priority Setting in the 21st Century


SEPA
United States
Environmental Protection
Agency
March 2,2020
Exposure-based	C
Setting	in the
John Wamb
Center for	Computatiooal
Office of Research
U.S. Environmental Protection Agency
The views expressed in this presentation are those of the author
and do not necessarily reflect the views or policies of the U.S. EPA
https://orcid.Org/0000-0002-4024-534X
Progress for o Stronger Future

-------
• Research conducted by a combination of Federal
scientists (including uniformed members of the
Public Health Service); contract researchers; and
postdoctoral, graduate student, and post-
baccalaureate trainees
EHiH Office of Research and Development
vvEPA	US EPA Office of Research and Development
United States
Environmental Protection
Agency
•The Office of Research and Development (ORD) is the scientific research arm of EPA
•562 peer-reviewed journal articles in 2018
• Research is conducted by ORD's four national centers, and three
offices organized to address:
• Public health and en v. assessment; comp. tox. and exposure;
env. measurement and modeling; and en v. solutions and
emergency response.
•13 facilities across the United States
ORD Facility in
Research Triangle Park, NC

-------
vvEPA
United States
Environmental Protection
Agency
Chemical Regulation in the United States
Park et al. (2012): At least 3221 chemical
signatures in pooled human blood samples, many
appear to be exogenous
A tapestry of laws covers the chemicals people
are exposed to in the United States (Breyer, 2009)
Chemical safety testing is primarily for food
additives, pharmaceuticals, and pesticide active
ingredients (NRC, 2007)
• Different levels depending on category
GIVE A DOG A PHONE
Technology for our furry friends
NewScientist
wnnr	dwar** l xm
We've made
150,000 new chemicals
*
We touch them,
we wear them, we eat them
But which ones should
we worry about?
SPECIAL REPORT, page 3-1
ms*N vasn CwaMM
THE GOOD FIGHT CHAMBER OF SEC RETS IS IT ALIVE?
Mcit viol nxc	The gfMt«**VCT find	Artlfid^wornitwld
ijaUo wtuoui	of early human bones	be first cfcgiua .animal
3 of 67
Office of Research arid Development
November 29, 2014

-------
vvEPA
United States
Environmental Protection
Agency
Chemical Regulation in the United States
Most other chemicals, ranging from industrial waste
to dyes to packing materials, are covered by the Toxic
Substances Control Act (TSCA)
Thousands of chemicals on the market were
"grandfathered" in without assessment
Judson et al. (2009), Egeghy et al. (2012), Wetmore et al. (2015)
GAO
March 2013
United States Government Accountability Office
Report to Congressional Requesters
TOXIC SUBSTANCES
EPA Has Increased
Efforts to Assess and
Control Chemicals but
Could Strengthen Its
Approach
"Tens of thousands	of chemical
Environmental	Protection	Age
use in the United States,	wit
chemicals	listed
U.S. Government Accountability
GAO
Accountability • Integrity • Reliability
GAO-13-249
4 of 67
Office of Research arid Development
March, 2013

-------
oEPA
United States
Environmental Protection
Agency
NRC (1983)
I hree Components for Chemical Risk
Dose-Response
(Toxicokinetics
/Toxicodynamics)
The National Academy of Sciences, Engineering and Medicine (1983)
| office of Research and Development outlined three components for determining chemical risk.

-------
TSCA 2.0
A New Era in
Chemical Risk Management
After decades of dysfunction, the Toxic Substances Control Act has been
overhauled with provisions that promise better protection against potentially
harmful chemicals. © Dmytro Grankin/Alamy; Alexander Aldatov/Alamy
I #: eb I H—

vvEPA
United States
Environmental Protection
Agency
Toxic Substances
Control Act (TSCA)
TSCA was updated in June, 2016 to allow more rapid
evaluation of chemicals (Frank R. Lautenberg
Chemical Safety for the 21st Century Act)
New approach methodologies (NAMs) are being
considered to inform prioritization of chemicals for
testing and evaluation (Kavlock et al., 2018)
EPA has released a "A Working Approach for
Identifying Potential Candidate Chemicals for
Prioritization" (September, 2018)
Schmidt, C. W. (2016). TSCA 2.0: A new era in
chemical risk management", Environmental
Health Perspectives, A182-A186.
6 of 67
Office of Research and Development

-------
vvEPA
United States
Environmental Protection
Agency
New Approach Methodologies (NAMs)
There are roughly 10,000 TSCA-relevant
chemicals in commerce
• Traditional methods are too
resource-intensive to address all of
these
NAMs include:
• High throughput screening (ToxCast)
• High throughput exposure estimates
(ExpoCast)
• High throughput toxicokinetics
(HTTK)
Chemical
Research in
Toxicology
& Cite This: Chem. Res. Toxicol. 2018, 31, 287-290
pubs.acs.orgA
Accelerating the Pace of Chemical Risk Assessment
J. ___ J. J. J. c
Robert J. Kavlock, Tina Bahadori, Tara S. B art on-Ma clar en, Maureen R Gwinn, Mike Rasenberg,
and Russell S. Thomas*'"
ABSTRACT: Changes in chemical regulations worldwide have
increased tlie demand for new data on chemical safety. New
approach methodologies (NAMs) are defined broadly here as
including in silico approaches and in chemico and in vitro assays,
as well as the inclusion of information from the exposure of
chemicals in the context of hazard [European Chemicals
Agency, "New Approach Methodologies in Regulatory Science",
2016]. NAMs for toxicity testing, including alternatives to
animal testing approaches, have shown promise to provide a
large amount of data to fill information gaps in both hazard
and exposure. In order to increase experience with the new
data and to advance the applications of NAM data to evaluate
the safety of data-poor chemicals, demonstration case studies
Accelerating the Pace of Chemical Risk Assessment
APCRA«r.»
TSCA Proof of concept: Examine ~200 chemicals with ToxCast, ExpoCast and HTTK
• HTTK was rate Iimiter on number of chemicals
• "A Proof-of-Concept Case Study Integrating Publicly Available Information to Screen Candidates for
Chemical Prioritization under TSCA"
7 of 67
Office of Research and Development

-------
oEPA
United States
Environmental Protection
Agency
Replacing Animal Testing with NAMs
Administrator of the EPA: "To aggressively pursue a
reduction in animal testing, I am directing leadership and
staff in the Office of Chemical Safety and Pollution
Prevention and the Office of Research and Development
[ORD] to prioritize ... the reduction of animal testing while
ensuring protection of human health and the
environment."
"These new approach methods (NAMs), include any
technologies, methodologies, approaches or combinations
thereof that can be used to provide information on
chemical hazard and potential human exposure that can
avoid or significantly reduce the use of testing on animals"
• NAMs for filling information gaps for decision-making
• integrating data steams into chemical risk assessment
• making the information publicly available
~| Office of Research and Development
(sbJ
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON. D C 20460
September 10, 2019
THE ADMINISTRATOR
MKMOKAMH M
FROM:
TO:
Andrew R. Wheeler
Administrator
Associate Deputy Administrator
General Counsel
Assistant Administrators
Inspector General
Chief Financial Officer
Chief of Staff
Associate Administrators
Regional Administrators
During my March 2019 all-hands address. I reiterated the U.S. Knvironmcntal Protection
Agency s commitment to move away from animal testing. We are already making significant
efforts to reduce, replace and refine our animal testing requirements under both statutory and
strategic directives, l-or example, the Toxic Substances Control Acl, amended June 22, 2016, by
the Frank R. Uutcnbcrg Chemical Safely for the 21" Century Act. requires the EPA to reduce
reliance on animal testing. Also, Objective 3.3 of the FY 20IX-2022 U.S. EPA Strategic Plan
outlines a commitment to turlher reduce the reliance on animal testing within live years. More
than 200,000 laboratory animals have been saved in recent years as a result of these collective
efforts.
Scientific advancements exist today that allow us to better predict potential hazards for risk
assessment purposes w ithout the use of traditional methods that rely on animal testing. These new
approach methods (NAMs), include any technologies, methodologies, approaches or combinations
thereof that can be used to provide information on chemical hazard and potential human exposure
that can avoid or significantly reduce the use of testing on animals. The benefits of NAMs are
extensive, not only allowing us to decrease animals used while potentially evaluating more
chemicals across a broader range ol potential biological effects, but in a shorter timeframe with
fewer resources while often achieving equal or greater biological predictiviiy than current animal
models.

-------
SEPA
Chemical Risk
United States
Environmental Protection
Agency
•	The U.S. National Research Council (1983) identified
chemical risk as a function of both inherent hazard
and exposure
•	Therefore, high throughput risk prioritization needs:
1.	High throughput hazard characterization
(Dix et al., 2007, Collins et al., 2008)
2.	High throughput exposure forecasts
(Wambaugh et al., 2013, 2014)
3.	High throughput toxicokinetics (i.e., dose-
response relationship) linking hazard and
exposure
(Wetmore et al., 2012, 2015)
| Office of Research and Development
= Hazard x Exposure
mg/kg BW/day
A
Potential Hazard
from in vitro with
Reverse
Toxicokinetics
Potential
Exposure Rate
Lower
Risk
Medium Higher
Risk	Risk

-------
vvEPA
United States
Environmental Protection
Agency
TOXICITY TESTING IN THE 21 ST CENTURY
A VISION AND A STRATEGY
NRC (2007)
High- hroughput Risk Prioritization
High-Throughput
Risk
Dose-Response
(Toxicokinetics
/Toxicodynamics)
Prioritization
High throughput screening
(HTS) for in vitro bioactivity
potentially allows
characterization of thousands
of chemicals for which no
other testing has occurred
10 of 67
Office of Research and Development
To perform high throughput risk prioritization, we need all three components

-------
oEPA
United States
Environmental Protection
Agency
High-throughput Screening
Hertzberg and Pope (2000):
• "New technologies in high-throughput screening have significantly increased throughput and reduced
assay volumes..."
Kaewkhaw et al. (2016)
"...new fluorescence
methods, detection
platforms and liquid-
handling technologies.'
Typically assess many
chemicals with a signal
readout (e.g., green
fluorescent protein).
Positive
Control
Titration of
Potential Hits
| Office of Research arid Development

-------
oEPA
United States
Environmental Protection
Agency
High-Throughput B ioactivity
Screening Projects
(S
A

We attempt to estimate points of departure in vitro using
high throughput screening (HTS)
Tox21: Examining >8,000 chemicals using ~50 assays
intended to identify interactions with biological pathways
(Schmidt, 2009)
ToxCast: For a subset (>2000) of Tox21 chemicals ran
>1100 additional assays (Kavlock et ol., 2012)
Most assays conducted in dose-response format (identify
50% activity concentration - AC50 - and efficacy if data
described by a Hill function, Filer et ol., 2016)
All data are public: http://comptox.epa.gov/dashboard/
| Office of Research and Development
\

I To
Nfc c — ^ NTP
National Institute of
Environmental Health Sciences
NoOcrtol Tbweobgy ftfagrow
CD
i/)
C
O
Q_
KTi
-------
oEPA
United States
Environmental Protection
Agency
1000
100
I
4
«
u
E
o
u
e
c
0.01
0.001
The Margin Between Exposure and Hazard
1000
100 3
u
e
o
u
n
£
U1
ra
10
0.1
0.01
0.001
o
E
3

5
Triclosan MBP MEHP PFOA 2,4-D
(90/615) (8/615) (35/615) (24/615) (10/615)
Range of bioactive concentrations
across ToxCast assays
^ Estimated or measured
average concentrations
associated with the LOAEL
in animal studies
O NOAEL in animal studies
£ Humans with chronic
exposure reference values
(solid circles)
X Volunteers using products
containing the chemical
+ Bio-monitored occupational
populations
A General populations
The five chemicals (as of 2011) with plasma biomonitoring AND ToxCast data... what do we do about the other 1000's?
| Office of Research arid Development
Ay I ward and Hays (2011)
-------
vvEPA
United States
Environmental Protection
Agency
The Nation*! Atadmia of
SCIENCES • ENGINEERING ¦ MEOCINE
USING
21 ST CENTURY
SCIENCE
TO IMPROVE
RISK-RELATED
EVALUATIONS
Most Chemicals Lack Data on Exposure and
/^Toxicokinetics
"Translation of high-throughput data into risk-
based rankings is an important application of
exposure data for chemical priority-setting.
Recent advances in high-throughput
toxicity assessment, notably the ToxCast
and Tox21 programs... and in high-
throughput computational exposure
NASEM (2017)
High-Throughput
Risk
Prioritization
assessment [ExpoCast] have enabled
first-tier risk-based rankings of
chemicals on the basis of margins
of exposure" - National Academies
of Sciences,
Engineering, and
Medicine (NASEM)
14 of 67
Office of Research and Development
In order to perform risk-based ranking we need data on hazard,
toxicokinetics, and exposure...
-------
vvEPA
United States
Environmental Protection
* gency
Chemical Prioritization NAMs
OJ
i_
=3
u~t
O
Q.
X
10
"D
CD
U
T3
CD
oT *****"
CL. fD
"O
o \ 10-
CD >
m cQ
Q
c
E
| ^ 10"
*5
cr
~o
CD
+-»
03
to
JL
P
IS i

iu^r
High throughput in vitro
screening can estimate doses
needed to cause bioactivity
(e.g., Wetmore et al.; 2015)
Exposure intake rates can
be inferred from
biomarkers
(e.g., Ring et al., 2018)
Chemicals Monitored by CDC NHANES
mg/kg
Potential
Hazard from
in vitro with
Reverse
Toxicokinetic
s
Potential
Exposure
Rate
3W/day
15 of 67
Office of Research and Development
Ring et al. (2017)
Lower Medium Higher
Risk Risk Risk
-------
oEPA
United States
Environmental Protection
Agency
In Vitro - In Vivo Extrapolation (IVIVE)
IVIVE is the use of In vitro experimental data to predict phenomena in vivo
• IVIVE-PK/TK (Pharmacokinetics/Toxicokinetics):
• Fate of molecules/chemicals in body
• Considers absorption, distribution, metabolism, excretion (ADME)
• Uses empirical PK and physiologically-based (PBPK) modeling
IVIVE-PD/TD (Pharmacodynamics/Toxicodynamics):
• Effect of molecules/chemicals at biological
target in vivo
• Assay design/selection important
• Perturbation as adverse/therapeutic effect,
reversible/ irreversible effeccts
Both contribute to in vivo effect prediction
| Office of Research and Development
Normalization of dose
PBPK models
NRC (1998)
Humans: in vivo
Testable predictions
Extrapolation
using PD and
PBPK models
Comparative testing
-------
vvEPA
United States
Environmental Protection
Agency
High Throughput Toxicokinetics (HTTK)
In vitrotoxicokinetic data + generic toxicokinetic model
= high(er) throughput toxicokinetics
V ^ V
¦=>

¦
j:V".
I
8

0
w

l 2 ¦=> !•! ¦=>

*al
CL
metab

Primary
Compartment

Gut Lumen
—>

CL
vabs
httk
17 of 67
Office of Research arid Development
-------
vvEPA
United States
Environmental Protection
Agency
Building Confidence inTK Models
To evaluate a chemical-specific TK model for ''chemical x" you can
compare the predictions to in vivo measured data
• Can estimate bias
• Can estimate uncertainty
• Can consider using model to extrapolate to other situations
(dose, route, physiology) where you don't have data
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
O)
>
1—
O)
LO
O
X
X
.X
X
X
X
X
Chemical
Specific
Model
Predicted Concentrations
18 of 67
Office of Research and Development
Cohen Hubal et al. (2018)
-------
vvEPA
United States
Environmental Protection
Agency
Building Confidence inTK Models
To evaluate a chemical-specific TK model for ''chemical x" you can
compare the predictions to in vivo measured data
• Can estimate bias
• Can estimate uncertainty
• Can consider using model to extrapolate to other situations
(dose, route, physiology) where you don't have data
However, we do not typically have TK data
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
O)
>
1—
O)
LO
_Q
O
X
X
4
X
X
X
X
Chemical
Specific
Model
Predicted Concentrations
19 of 67
Office of Research and Development
Cohen Hubal et al. (2018)
-------
vvEPA
United States
Environmental Protection
Agency
Building Confidence inTK Models
To evaluate a chemical-specific TK model for ''chemical x" you can
compare the predictions to in vivo measured data
• Can estimate bias
• Can estimate uncertainty
• Can consider using model to extrapolate to other situations
(dose, route, physiology) where you don't have data
However, we do not typically have TK data
We can parameterize a generic TK model, and evaluate that
model for as many chemicals as we do have data
• We do expect larger uncertainty, but also greater confidence in
model implementation
• Estimate bias and uncertainty, and try to correlate with
chemical-specific properties
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
CD
>
1—
O)
LO
_Q
O
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
O)
>
1—
O)
LO
O
X
X
X
X
X
X
Chemical
Specific
Model
Predicted Concentrations
x

x
x
x
x
x
Generic
Model
Predicted Concentrations
20 of 67
Office of Research and Development
Cohen Hubal et al. (2018)
-------
oEPA
United States
Environmental Protection
Agency
Building Confidence inTK Models
• To evaluate a chemical-specific TK model for ''chemical x" you can
compare the predictions to in vivo measured data
• Can estimate bias
• Can estimate uncertainty
• Can consider using model to extrapolate to other situations
(dose, route, physiology) where you don't have data
• However, we do not typically have TK data
• We can parameterize a generic TK model, and evaluate that
model for as many chemicals as we do have data
• We do expect larger uncertainty, but also greater confidence in
model implementation
• Estimate bias and uncertainty, and try to correlate with
chemical-specific properties
• Can consider using model to extrapolate to other situations
(chemicals without in vivo data)
| Office of Research and Development
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
CD
>
1—
O)
LO
_Q
O
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
O)
>
1—
O)
LO
_Q
O
X
X
X
X
X
X
Chemical
Specific
Model
Predicted Concentrations
x
y
x
x /

X
X
X
Generic
Model
Predicted Concentrations
Cohen Hubal et al. (2018)
-------
oEPA
United States
Environmental Protection
Agency
Building Confidence inTK Models
• To evaluate a chemical-specific TK model for ''chemical x" you can
compare the predictions to in vivo measured data
• Can estimate bias
• Can estimate uncertainty
• Can consider using model to extrapolate to other situations
(dose, route, physiology) where you don't have data
• However, we do not typically have TK data
• We can parameterize a generic TK model, and evaluate that
model for as many chemicals as we do have data
• We do expect larger uncertainty, but also greater confidence in
model implementation
• Estimate bias and uncertainty, and try to correlate with
chemical-specific properties
• Can consider using model to extrapolate to other situations
(chemicals without in vivo data)
| Office of Research and Development
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
CD
>
1—
O)
LO
_Q
O
LO
c
o
'+->
03
i—
+->
C
o>
u
c
o
u
~o
O)
>
1—
O)
LO
_Q
O
X
X
X
X
X
X
Chemical
Specific
Model
Predicted Concentrations
x
y
x
x /
A y
/
Z
Z z
X
X
X
Generic
Model
Predicted Concentrations
Cohen Hubal et al. (2018)
-------
vvEPA
United States
Environmental Protection
Agency
O)
The HTTK model estimates chemical
clearance from the body by two
processes: '
* 1
• hepatic metabolism (liver) -
• passive glomerular filtration (j
(kidney) "O
0
as
This appears to work better for £
pharmaceuticals than other "55
chemicals: ®
• ToxCast chemicals are §
overestimated ^
C
Non-pharmaceuticals may be
subject to extrahepatic metabolism
and/or active transport
23 of 67
Office of Research and Development
Evaluation Example
Pharmaceuticals
Other Chemicals
Pharm : MSE = 2.44, R2 = 0.19
Other :MSE = 2.93, R2 = 0.5
iiilni^ i nl.J L-LliuJ mlJ uH ntlnJ luH i
10~3 10~1 10 103
In vitro predicted CLtot (mg/LTh)
Wambaugh et al. (2018)
-------
oEPA
United States
Environmental Protection
Agency
ToxicokineticTriage: When DoesTKlviVE
Through comparison to in vivo data, a cross-
validated (random forest) predictor of success or
failure of HTTK has been constructed
All chemicals can be placed into one of seven
confidence categories
• Added categories for chemicals that do not
reach steady-state or for which plasma binding
assay fails
Plurality of chemicals end up in the ''on the order"
bin (within a factor of 3.2x) which is consistent
with Wang (2010)
| Office of Research and Development
150'
03
O
1100
_c
0
I—
1
M—
0
1
(D
_o
E
^ 50-
\<3^
O'

,-
&
O

.<&-
,>
140

Wambaugh et al. (2015)

id?

J00
&

f-
&

Triage Category
-------
vvEPA
United States
Environmental Protection
* gency
Chemical Prioritization NAMs
OJ
i_
=3
u~t
O
Q.
X
10
"D
CD
U
T3
CD
oT *****"
CL. fD
"O
o \ 10-
CD >
m cQ
Q
c
E
| ^ 10"
*5
cr
~o
CD
+-»
03
to
JL
P
IS i

^
5 ilor
AJacdfar
Aco'achlor
Prop* (para
Ciciotochc
s
Ring et ol. (2017)
-------
oEPA
United States
Environmental Protection
Agency
Life-stage and Demographic Variation in K
• Ring et ol. (2017) made demographic-
specific predictions of change in plasma
concentrations for a 1 mg/kg bw/day
exposure
v~>
rt3
U
CU
u
co

-0.5
0.5
Change in Toxicokinetics
Relative to Total Population
| Office of Research and Development
&

^
5
-------
oEPA
United States
Environmental Protection
Agency
• Can calculate
margin between
bioactivity and
exposure for
specific
populations
Life-stage and Demographic Variation in Risk
PHnrifv
mg/kg BW/day
A
Potential Hazard
from in vitro with
Reverse
Toxicokinetics
Potential Exposure
from ExpoCast
-0.5
0.5
Change in Risk Relative to
Total Population
| Office of Research and Development
Lower Medium Risk Higher
Risk Risk
v~>
rt3
U
CU
u
co
5
-------
oEPA
United States
Environmental Protection
Agency
Open Source Tools and Data for HTTK
https://CRAN.R-proiect.orq/packaqe=httk
~
X
a ~ 0 0 %
<|f CRAM - Package httk X +
G cran.r-project.org/web/packages/httk/index.html
::: Apps Absence Request J Travei Request For... jg REMD-HTTK (& Confluence Q Bitbucket CompTox Dashboard A EHP © Change Password
httk: High-Throughput Toxicokinetics
Functions and data tables for simulation and statistical analysis of chemical toxicokinetics ("TK") as in Pearce et al. (2017) . Chemical-specific in vitro data have been obtained from relatively hisli
throughput experiments. Both physiologically-based ("PBTK") and empirical (e.g.. one compartment) "TK" models can be parameterized for several hundred chemicals and multiple species. These models are solved efficiently,
often using compiled (C-based) code. A Monte Carlo sampler is included for simulating biological variability (Ring et al., 2017 'l and measurement limitations. Calibrated methods are included
for predicting tissue:plasma partition coefficients and volume of distribution (Pearce et al., 2017
BugReports: httos: github.com USEPA CocinTox-ExcoCast-httk
License: GPL-3
URL: https:. www.epa.gov chemical-research, rapid-chemical-exposure-and-dose-researcli
NeedsCompilation: yes
Citation: httk citation info
Materials: NEWS
CRAN checks: httk results
Downloads:
R package "httk
i/
downloads 989/month
Reference manual: httk.pdf
Vignettes: Frank et al. i'2018~): Creating IVFVE Figure fFig. 61
Honda et al. (2019'i: Updated Armrtage et al. i'2014'} Model
Linakis et al. ("Submitted ): Analysis and Figure Generation
Pearce et al. <2017i: Creating Partition Coefficient Evaluation Plots
Ring et al. i'2017'i: Generating subuopulations
R -tier pt al C?0\ 7V F.valiiatincr HTTK models for tnhnfMnilafinns
Open source, transparent, and peer-reviewed
tools and data for high throughput
toxicokinetics (httk)
Available publicly for free statistical software R
Allows in vitro-in vivo extrapolation (IVIVE) and
physiologically-based toxicokinetics (PBTK)
Human-specific data for 944 chemicals and rat-
specific data for 171 chemicals
Described in Pearce et al. (2017)
-------
vvEPA
United States
Environmental Protection
Agency
High throughput screening (Dix et al.,
2006, Collins et a I., 2008) + in vitro-in
vivo extrapolation (IVIVE, Wetmore et
al., 2012, 2015) can predict a
(mg/kg bw/day) that might be
adverse
Risk = Hazard x Exposure
High-Throughput
Risk
Prioritization
30 of 67
Office of Research and Development
NRC (1983)
-------
vvEPA
United States
Environmental Protection
Agency
High throughput screening (Dix et al.,
2006, Collins et a I., 2008) + in vitro-in
vivo extrapolation (IVIVE, Wetmore et
al., 2012, 2015) can predict a
(mg/kg bw/day) that might be
adverse
Risk = Hazard x Exposure
High-Throughput
Risk
Prioritization
Need methods to forecast exposure for
thousands of chemicals
(Wetmore et al., 2015)
High throughput models exist to
make predictions of exposure via
specific, important pathways such
as residential product use and diet
31 of 67
Office of Research and Development
NRC (1983)
-------
vvEPA
United States
Environmental Protection
Agency
Understanding Exposure is a Systems
| Problem
11 ¦¦¦¦!
USE and RELEASE
Consumer
Products and
Durable Goods
Other Industry
Direct Use
(e.g., surface cleaner) (e.g.,
MEDIA
TARGET
Chemical Manufacturing and Processing
Environmental
Release
Outdoor Air, Soil, Surface
and Ground Water
32 of 67
Office of Research and Development
Figure from Kristin Isaacs
-------
vvEPA
United States
Environmental Protection
Agency
Exposure event is often unobservable
I
11 ¦¦¦¦!
USE and RELEASE
Consumer
Products and
Durable Goods
Other Industry
Chemical Manufacturing and Processing
Environmental
Release
Direct Us
(e.g., surface cleaned) (e.g.,
MEDIA
EXPOSURE
(MEDIA + TARGET)
TARGET
Can try to predict exposure by characterizing pathway
Some pathways have much higher average exposures: In home "Near field" sources significant (Wallace, et oi, 1987)
33 of 67
Office of Research and Development
Figure from Kristin Isaacs
-------
v>EPA
mg/kg BW/day
t \
United States
Environmental Protection
Agency
Consumer
Products arid
Durable Goods
Exposure Pathways
Chemical Manufacturing a
High Throughput
Screening +
Toxicokinetics
High
Throughput
Exposure
Rate
and Processing

Direct Use Residential Use
g. .flooring) Occupational
Indoor Air, Dust,
Surfaces
Near-Field Near-Field
Consumer
Ecological
Populations
Target Populations
Lower
Risk
Medium Higher
Risk Risk
Space of
Chemicals
Apply calibration and estimated
uncertaintyto other chemicals
Evaluate Model Performance
and Refine Models
34 of 67
Office of Research arid Development
NAMs for Exposure Science
Available online at www.sciencedirect.com
ScienceDirect
Current Opinion in
Toxicology
F! SFVTF.R
New approach methodologies for exposure science
John F. Wambaugh', Jane C. Bare , Courtney C. Carignan ,
Kathie L. Dionisio , Robin E. Dodson', Olivier Jolliet6,
Xiaoyu Liu , David E. Meyer2, Seth R. Newton4,
Katherine A. Phillips1, Paul S. Price , Caroline L. Ring8,
Hyeong-Moo Shin ', Jon R. Sobus\ Tamara Tal
Elin M. Ulrich;, Daniel A. Vallero , Barbara A. Wetmore and
Kristin K. Isaacs
%
Cimfktm
Abstract
Chemical risk assessment relies on knowledge of hazard, the
dose-response relationship, and exposure to characterize
potential risks to public health and the environment A chemical
with minimal toxicity might pose a risk if exposures are exten-
sive, repeated, and/or occurring during critical windows across
the human life span. Exposure assessment involves under-
standing human activity, and this activity is confounded by
interindividual variability that is both biological and behavioral.
Exposures further vary between the general population and
susceptible or occupational^ exposed populations. Recent
computational exposure efforts have tackled these problems
through the creation of new tools and predictive models. These
tools include machine learning to draw inferences from existing
data and computer-enhanced screening analyses to generate
new data. Mathematical models Drovide frameworks descnbina
9 Department of Earth and Environmental Sciences. University of
Texas, Arlington, TX 76019, USA
10 National Health and Environmental Effects Research Laboratory,
Office ot Research and Development, United States Environmental
Protection Agency, Research Triangle Park, NC 27711, USA
Corresponding author: Wambaugh, John F. (Wambaugfi.john@epa
gov)
Current Opinion in Toxicology 2019, 15:76-92
This review comes from a themed issue on Risk Assessment in
Toxicology
Edited by Anne Marie Vinggaard and Richard Judson
Available online 31 July 2019
For a complete overview see the issue and the Editoriai
https J/doi.org/10.1016/j.cotox.2019.07.001
-------
New Approach Methodologies for Exposure Science

Makes Use of

Exposure NAM Class
Description
Traditional Approach
Measurement
Toxicokinetics
Models
Descriptors
Evaluation
Machine
Learning
Measurements
New techniques including screening analyses
capable of detecting hundreds of chemicals
present in a sample
Targeted (chemical-specific) analyses

•
•
•

•
Toxicokinetics
High throughput methods using in vitro data to
generate chemical-specific models
Analyses based on in vivo animal studies
•
-

•

•
HTE Models
Models capable of making predictions for
thousands of chemicals
Models requiring detailed, chemical- and
scenario-specific information
•
•
-
•

Chemical Descriptors
Informatic approaches for organizing chemical
information in a machine-readable format
Tools targeted at single chemical analyses by
humans

•
Evaluation
Statistical approaches that use the data from
many chemicals to estimate the uncertainty in
a prediction for a new chemical
Comparison of model predictions to data on a
per chemical basis
•
•
•
•

•
Machine Learning
Computer algorithms to identify patterns
Manual Inspection of the Data
•
•

•

-
Prioritization
Integration of exposure and other NAMs to
identify chemicals for follow-up study
Expert decision making
•
•
•
•
•
•
35 of 67
| Office of Research and Development Wambaugh et al. (2019)
-------
vvEPA What Do We Know About Exposure?
United States
Environmental Protection
Agency
Biomonitoring Data
Centers for Disease Control and Prevention (CDC) National Health and Nutrition Examination Survey
(NHANES) provides an important tool for monitoring public health
Large, ongoing CDC survey of US population: demographic, body measures, medical exam,
biomonitoring (health and exposure),...
Designed to be representative of US population according to census data
Data sets publicly available (http://www.cdc.gov/nchs/nhanes.htm)
Includes measurements of:
• Body weight
• Height
• Chemical analysis of blood and urine
iinanes
National Health and Nutrition Examination Survey
| Office of Research and Development
-------
oEPA
What Do We Know About Exposure?
Exposure Models
United States
Environmental Protection
Agency
• Human chemical exposures can be coarsely grouped into ''near field" sources that are close to the
exposed individual (consumer or occupational exposures) 'far-field' scenarios wherein individuals are
exposed to chemicals that were released or used far away (ambient exposure) (Arnot et al., 2006).
• A model captures knowledge and a hypothesis of how the world works (MacLeod et al., 2010)
• EPA's EXPOsure toolBOX (EPA ExpoBox) is a toolbox created to assist individuals from within
government, industry, academia, and the general public with assessing exposure
• Includes many, many models
https://www.epa.gov/expobox
''Now it would be very remarkable if any system existing in the real world could be exactly represented by
any simple model. However, cunningly chosen parsimonious models often do provide remarkably useful
approximations... The only question of interest is 'Is the model illuminating and useful?'" George Box
37 of 67
Office of Research and Development
-------
oEPA
United States
Environmental Protection
Agency
Predictive
Modeling
USE and RELEASE
Models to Predict Exposure
I
11 ¦¦¦¦!
Consumer
Products and
Durable Goods
Other Industry
Direct Use
(e.g., surface cleaner) (e.g.,
MEDIA
EXPOSURE
(MEDIA + TARGET)
TARGET
Chemical Manufacturing and Processing
Environmental
Release
Outdoor Air, Soil, Surface
and Ground Water
We can try to predict exposure by describing the process leading to exposure
| Office of Research and Development
Figure from Kristin Isaacs
-------
oEPA
United States
Environmental Protection
Agency
Monitoring Data
USE and RELEASE
Consumer
Products and
Durable Goods
Chemical Manufacturing and Processing
Direct Us
(e.g., surface cleaned) (e.g.,
MEDIA
EXPOSURE
(MEDIA + TARGET)
TARGET
MONITORING DATA
We can also infer
exposure from monitoring data
Biomarkers
of Exposure
Sampling
Biomarkers
of Exposure
| Office of Research and Development
Figure from Kristin Isaacs
-------
oEPA
United States
Environmental Protection
Agency
Models to Infer Exposure
USE and RELEASE
Consumer
Products and
Durable Goods
Chemical Manufacturing and Processing
Direct Us^ Residential Use
(e.g., surface cleaner) (e.g.,
MEDIA
EXPOSURE
(MEDIA + TARGET)
TARGET
MONITORING DATA
Inference
("Reverse Modeling")
Biomarkers
of Exposure
Sampling
Biomarkers
of Exposure
| Office of Research and Development
Figure from Kristin Isaacs
-------
oEPA
United States
Environmental Protection
Agency
Predictive
Modeling
USE and RELEASE
Evaluating Models with Monitoring Data
Consumer
Products and
Durable Goods
Chemical Manufacturing and Processing
Direct Us^ Residential Use
(e.g., surface cleaner) (e.g.,
MEDIA
EXPOSURE
(MEDIA + TARGET)
TARGET
MONITORING DATA
Inference
("Reverse Modeling")
Biomarkers
of Exposure
Sampling
Biomarkers
of Exposure
| Office of Research and Development
Figure from Kristin Isaacs
-------
vvEPA
Evaluation NAMs: The SEEM Framework
United States
Environmental Protection
Agency
r
Space of
Chemicals
r
Chemicals
with
Monitoring
Data
Dataset 1
Dataset 2
Apply calibration and estimated uncertainty to
other chemicals
¦M
m
cc
0
JS£
fO
V!
Exposure
Ml inference
hi
CD
Model 1
Model 2
Estimate ¦
Uncertainty
Calibrate
models
•
\
Different
Chemicals
J <
Available Exposure Predictors
| Office of Research arid Development
3
Evaluate Model Performance
and Refine Models
FEMN
AEMN
CEMN
We use Bayesian methods to incorporate multiple models into consensus predictions for
1000s of chemicals within the Systematic Empirical Evaluation of Models (SEEM)
(Wambaugh et al., 2013, 2014; Ring et al., 2018)
Hurricane path
prediction is an
example of
integrating
multiple models
-------
vvEPA
United States
Environmental Protection
Agency
SEEM is a Linear Regression
Multiple regression models:
Log(Parent Exposure) = a + m * log(Model Prediction) + b* Near Field + e

fr~N(0, o2)
Residual error,
unexplained by
the regression
model
Weighted HTE Model Predictions
43 of 67
Office of Research and Development
-------
vvEPA
United States
Environmental Protection
Agency
SEEM is a Linear Regression
Multiple regression models:
Log(Parent Exposure) = a + m * log(Model Prediction) + b* Near Field + e

0
in
O
Q.
X
LU
"O
&_
&_
0

Weighted HTE Model Predictions
Not all models have predictions
for all chemicals
• We can run SHEDS-HT
(Isaacs et alv 2014) for
~2500 chemicals
What do we do for the rest?
• Assign the average value?
• Zero?
44 of 67
Office of Research and Development
-------
vvEPA
United States
Environmental Protection
Age
SEEM Analysis of NHANES Data
I. (2014)
./

R2 =
0.5
1e-08 1e-05 1e-02
Predicted Parental Exposure (mg / kg body weight / day)
R2 ~ 0.5 indicates that we can predict
50% of the chemical to chemical
variability in median NHANES
exposure rates
Same five predictors work for all
NHANES demographic groups
analyzed - stratified by age, sex, and
body-mass index:
• Industrial and Consumer use
• Pesticide Inert
• Pesticide Active
• Industrial but no Consumer
use
• Production Volume
45 of 67
Office of Research and Development
-------
vi-EPA
United States
Environmental Protection
Agency
_ 2-
C
CD
"o
Q3
O
O
c 0-
o
CO
CO
8M -
O)
0
DC
Wambaugh et al. (2014)
I I l < I
46 of 67
Office of Research arid Development
Heuristics of Exposure
— Total
— Female
— Male
— ReproAgeFemale
— 6-11_years
— 12-19_years
— 20-65_years
— 66+years
— BMI_LE_30
BMI GT 30
R2 ~ 0.5 indicates that we can predict
50% of the chemical to chemical
variability in median NHANES
exposure rates
Same five predictors work for all
NHANES demographic groups
analyzed - stratified by age, sex, and
body-mass index:
• Industrial and Consumer use
• Pesticide Inert
• Pesticide Active
• Industrial but no Consumer
use
• Production Volume
-------
vvEPA
United States
Environmental Protection
Agency
Correlation is Not Causation
Wambaugh et al. (2014) found that "pesticide inerts"
had higher than average levels in biomonitoring data,
while "pesticide actives" had lower than average
In World War II, there Royal Air Force (UK) wanted to
armor planes against anti-aircraft fire
• Initial proposal was to place armor wherever
bullet holes were most common
• Mathematician Abraham Wald pointed out that
they were looking at the planes that had returned
• See Drum, Kevin (2010) "The Counterintuitive
World"
Pesticide inerts have many other uses, but there are
more stringent reporting requirements for pesticides
• Exposure is occuring by other pathways
47 of 67
Office of Research and Development
-------
vi-EPA
United States
Environmental Protection
Agency
Oil the Solvability of the Six Degrees
of Kevin Bacon Game
A Faster Graph Diameter
and Radius Computation Method
Mkbde Borawu'. Picrluigi Crcsrcmi2, Mirhrl Habit)3,
Walter Kurtm*, Andrea Marino*", ami Frank Takes'
1 IMT Institute u( Advauted Stwlkn, Uhi«. Italy
1 Dipartuwnto «b Stait«uii e InformaiK a. I nivif.ila ili Rnnai, Italy
' LIAFA. t'MH 7IW ('NRS t I'aiverMitt- Paris Diderot • 1'nrin 7. Franco
4 Uxhi Institute iif Adviuicail Computer Sot-no-.
I>-»den Univvnrity. Tlit' Nrtliprlaiiiln
I>i(Mituiniili> t(-«d ol |'| BFSos (independent of tbt- value uf i\ |). and (bun
luiving running tune 0(|£T|). Apart from etlick-ucy. compared to other
similar method*, the one (iropoa-d in this paper has three other advan-
tages It u mote robust (even in th«- worst rm-x. the uumlM-r of BFSe*
perfemned » not wry high), it i» able to simultaneously cum put® radius
and diameter lhalvmg the total runuing time whenever both values are
neuied). and a works both on directed and undirected graphs with very
few m«difiratk«ia As an applk-ation example. we uml- cair new algorithm
in order to deteruiuie the solvability over time of the "six degree* of
Kevin Baron" game.
1 Introduction
The six degrees of separation game is a trivia game which has been inspired by
the well-known social experiment of Stanley Milgram [II]. which was in turn
a continuation of the empirical study of the structure of social networks by
Michael Gurevich [7]. Indeed, the notion of six degrees of separation has been
formulated for the first time by Frigyes Karinthy in 1929. who conjectured that
any two individuals can be connected through at most five acquaintances. This
conjecture has somehow l>een experimentally verified by Milgram ;uid extremely
popularized by a theater play of .John Guare. successively adapteil to the cinema
by Fred Scbepisi. The corresponding game refers to a social network, such as the
¦ ¦ ; II ! '• M 'I'¦ I - - \ >v.h
48 Of 67
Office of Research arid Development
ll he Six Degrees of Kevin Bacon
Kevin Baron and Graph Theory
KEVIN BACON
AND GRAPH THEORY
Brian Hopkins
RKSS Department of Mathematics, Saint Peter's College, Jersey
City NJ 07306 USA bbopkinsfspc adu
The interconnected world of actors and movies is a familiar,
rich example foe graph theory Yhi» paper gives the history of the
"Kevin Bacon Game" and makes extensive use of a Web site to analyse
the underlying graph The main content » the classroom development
of the weighted average to determine the beat choice of "center" for
the graph The arUcie conclude* with additional student activities
and lume responses to the material
Cinema, finite mathematics, graph theory, popular culture,
six degrees of separatum, weighted averages
1 INTRODUCTION
h theory is the mathematics of connections. It has wide applications to
interconnected systems transportation networks, epidemiology, ami
to name just a few But »e teach graph theory with pictures
of dots and lines There is one Urge system that is easy to work
thanks to a Web site run by the University of Virginia, Department
Science The Oracle of Bacon at Virginia <6| uses the Internet
Database (3). which documents almost all of cinematic history This is
tool for illustrating complete subgraphs, connected components, arid
between vertices There is also a nice application of weighted
I have used this material in freshman finite mathematics classes
major courses that cover graph theory, students always
enthusiastically
-------
vi-EPA
United States
Environmental Protection
Agency
Kevin Bacon
. RON HOVARD •* H v vN
APOLLOjl3
IMAGINE ENTtRTAINMENTwKdiBRI^'cRAZOiiionoio 'AP0U013* KATHUEN QUINUN ":|AM{$ HORNffi
SSSBTA RYACK -"SSAEDMC LA'AULI PORTER MICHAEL BOSTKK JSMt Hilt DAN HANIEY
-SSMICHAEL COHENBLITH .."KKKAN CUNOEV.w SSKKTODD HAU0WHL -»SS|M tOVEU.
(H^WaSS1 R!f JEFFREY KLUCGR """5WIU.IAM8ROYIES, |8. *Al REINER! "" i'8RlAN CRAZB!
mm -."Ronhowaro coming soon a unmrsal picture
49 of 67
| Office of Research and Development
-------
Kevin Bacon
United States
Environmental Protection
Agency
1990
*>EPA
OOD MEN
ihhmmwqiwmtipws' ffflW»-ra»M-irom-w
-------
oEPA
Michael B. Jordan
United States
Environmental Protection
Agency
51 of 67
Office of Research arid Development
AWKWARD
MOMENT
WHEN YOU REALIZE GETTING SOME
MEANS WANTING MORE
-------
vi-EPA
United States
Environmental Protection
Agency
Frances McDormand
Best Actress Winner 2018
Expendables
Willis &
Sylvester Stallone
Creed!
Stallone & Jordan
Connectedness to Michael B.Jordan
Hail Caesar
McDormand &
ChanningTatum
Gl Joe: Retaliation
Tatum & Bruce Willis
Office of Research and Development
-------
oEPA
United States
Environmental
Agency
| Office of Research arid Development
Superman
with Gene Hackman
Marlon Brando
Best Actor 1954 and 1972
Died 2004
The Royal Tenenbaums
Hackman & Gwyneth Paltrow
Protection
Connectedness to Michael B.Jordan
Avengers:
Infinity War
Paltrow &
Chadwick
Boseman
Black Panther
Boseman & Jordan
-------
vvEPA
United States
Environmental Protection
Agency
letters to nature
Watts and Strogatz (1998)
typically slower than -lions"'I might differ significantly from
what is assumed by current modelling efforts27. The expected
equation-of-state differences among small bodies (ice versus rock,
for instance) presents another dimension of study; having recently
adapted our code for massively parallel architectures (K. M. Olson
and E.A, manuscript in preparation}, we are now ready to perform a
more comprehensive analysis.
The exploratory simulations presented here suggest that when a
young, non-porous asteroid (if such exist) suffers extensive impact
damage. live resulting fracture pattern largely defines the asteroid's
response to future impacts. The stochastic nature of collisions
implies that small asteroid interiors may be as diverse as their
shapes and spin states. Detailed numerical simulations of impacts,
using accurate shape models and theologies, could shed light on
how asteroid collisional response depends on internal configuration
and shape, and hence on how planetesimals evolve. Detailed
simulations are also required before one can predict the quantitative
effects of nuclear explosions on Earth-crossing comets and
asteroids, cither for hazard mitigation9 through disruption and
deflection, or for resource exploitation29. Such predictions would
require detailed reconnaissance concerning die composition and
internal structure of the targeted object.
Collective dynamics of
'small-world' networks
Duncan J. Watts- & Steven H. Strogatz
Department of Theoretical and Applied Mechanics. Kimball Hall.
Cornell University, Ithaca, Sew York 14S53, USA
Networks of coupled dynamical systems have been used to model
biological oscillators1'*, losephson junction arrays'", excitable
media', neural networks""1, spatial games", genetic control
networks11 and many other self-organizing systems. Ordinarily,
the connection topology is assumed to be either completely
regular or completely random. But many biological, technological
and social networks lie somewhere between these two extremes.
Here we explore simple models of networks thai can be tuned
through this middle ground; regular networks 'rewired' to intro-
duce increasing amounts of disorder. We find that these systems
can be highly clustered, like regular lattices, yet have small
characteristic path lengths, like random graphs. We call (hem
'small-world' networks, by analogy with the small-world
phenomenon1344 ( popularly known as six degrees of separation1').
The neural network of the worm Caenorhabditis elegans, the
power grid of the western United States, and the collaboration
graph of film actors are shown to be small-world networks.
Models of dynamical systems with small-world coupling display
enhanced signal-propagation speed, computational power, and
synchronizability. In particular, infectious diseases spread more
easily in small-world networks than in regular lattices.
To interpolate between regular and random networks, we con-
sider the following random rewiring procedure (Fig. 1). Starting
from a ring lattice with h vertices and k edges per vertex, we rewire
each edge at random with probability p. This construction allows us
to 'tune' the graph between regularity (p = 0) and disorder (p = 1),
and thereby to probe the intermediate region 0 < p < 1, about
which little is known.
We quantify' the structural properties of these graphs by their
characteristic path length Hp) and clustering coefficient C(p). as
defined in Fig. 2 legend. Here Up) measures the typical separation
between two vertices in the graph (a global property), whereas C(p)
measures the cliquishness of a typical neighbourhood (a local
property). The networks of interest to us have many vertices
with sparse connections, but not so sparse that the graph is in
danger of becoming disconnected. Specifically, we require
n3S» k»ln(n) » 1, where k ln(n) guarantees thai a random
graph will be connected". In this regime, we find that
L-n/lk 3> 1 and C~3/4 as p—*0, while I =*» L .„ — ln(h)/lntt>
andf'^C , ~kln eg 1 asp—• I. Thus the regular lattice at p = 0
is a highly clustered, large world where L grows linearly with it,
whereas the random network at p = 1 is a poorly clustered, small
world where L grows only logarithmically with n. These limiting
cases might lead one to suspect that large Cis always associated with
large L, and small Cwith small L.
On the contrary. Fig. 2 reveals that there is a broad interval of p
over which Hp) is almost as small as luadom yet C(p) >
These small-world networks result from the immediate drop in Hp)
caused by the introduction of a few long-range edges. Such 'short
cuts' connect vertices that would otherwise be much farther apart
than For small p, each short cut lias a highly nonlinear effect
on L, contracting the distance not just between the pair of vertices
that it connects, but between their immediate neighbourhoods,
neighbourhoods of neighbo urhoods and so on. By contrast, an edge
54 of 67
Office of Research arid Development
Small World
Networks
Travers and
Milgram (1977):
296 arbitrary
individuals in
Nebraska and
Boston were
asked to give a
letter to an
acquaintance
most likely to
help it reach a
target person in
Massachusetts.
64 reached the
target person,
average number
of intermediaries
was 5.2
Collins and Chow (1998)
It's a small
world
The concept of Six Degrees of Separation has been formalized in
so-called 'small-world networks'. The principles involved could be of use
in settings as diverse as improving networks of cellular phones and
understanding the spread of infections.
James J. Collins and Carson C. Chow
A few years ago, on American campus-
es, it was popular to play Six Degrees
of Kevin Bacon. In this game, partici-
pants attempt to link the actor Kevin Bacon
to any other actor through as few common
films and co-stars as possible. Links are
formed directly between Bacon and another
actor if they appeared in the same film
or indirectly through a chain of co-stars in
different films (Fig. 1).
In the world of mathematics, a similar
amusement involves assessing one's Erd&s
number, which measures the number of
links needed to connect one to the prolific
mathematician Paul Erd&s through jointly
authored papers. For example, individuals
have an Erd&s number of 1 if they co-
authored a paper with Erdfts. If one of their
co-authors wrote a paper with Erdds, then
they have an Erdos number of 2, and so fortlt
It has been pointed out1 that Dan Kleitman
has a combined Erd&s/Bacon number of 3
because he wrote a paper with Erdits and
appeared in Good Will Hurtling with Minnie
Driver, who appeared with Bacon in Sleepers.
These games are related to the popular
concept of Six Degrees of Separation1, which
is based on the notion that everyone in the
world is connected to everyone else through
a chain of at most six mutual acquaintances.
If two people have one mutual acquaintance,
then they have one degree of separation. The
estimate ofsix degrees of separation, which is
related to the small-world phenomenon'-4,
arises from pioneering empirical work by
Milgram' and can be understood heuristi-
cally from a somewhat unrealistic assump-
tion of random connectivity. That is, if each
person knows about one hundred individu-
als. and given that there are about a billion
people on die Earth, then seven connections
or six degrees of separation are enough to
linkeveryone together.
On page 440 of this issue', Watts and
Strogatz formalize this idea in what they
call small-world networks. They demonstrate
through numerical simulations that a net-
work need not be very random to get this
small-world effect They consider a connect-
ed network with nodes and links. In the
friendship analogy, each node represents a
person and each link represents a single con-
nection to an acquaintance. They then define
NATURElVOtMSl* IUNE1998
two measures. The first is a characteristic
path length. This is the smallest number of
links it lakes to connect one node to another,
averaged over all pairs of nodes in the net-
work. The second measure is the clustering
coefficient. This measures the amount of
cliquishness of the network, diat is, die
fraction of neighbouring nodes dial are also
connected to one another. For example, in an
all-to-all connected network, the clustering
coefficient is one
An example of a large-world network is
one that is regularly and locally connected
like a crystalline Lattice. Such a network is
higlily clustered and the characteristic path
length is large, scaling with the typical linear
dimension of the network. On the other
hand, a completely random network is
poorly clustered and the characteristic path
news and views
length is short, scaling logarithmically with
the sizeofthe network.
What Watts and Strogatz' do is to shift
gradually from a regular network to a ran-
dom network by increasing the probability
of making random connections from 0 to 1
(see Fig. I, page 441). They then measure the
characteristic path length and the amount of
clustering of the network as a function of the
amount of randomness. They find that path
length and clustering depend differently on
the amount of randomness in the network.
The characteristic path length drops quickly,
whereas the amount of clustering drops
rather slowly, litis leads to a small-world
network in which the amount of clustering is
high and the characteristic path length is
short. So a small world can exist even when
the cliquishness is imperceptibly different
from that of a large world.
The explanation for this effect is that it
only takes a few short cuts between cliques to
turn a large world into a small world. In the
friendship analogy, it only takes a small num-
ber of well-connected people to make a world
small. The interesting and surprising thing is
that it is impossible to determine whether or
not you live in a small world or a large world
from local information alone. The average
person (node) is not direcdy associated with
die key people (the clique-linkers).
Small-world connectivity lias con-
sequences that could be good or bad,
figure 1 Three degrees. Because Kevin Bacon has appeared in many films most actors have low Bacon
numbers and the game Six Degrees of Kevin Bacon has declined in popularity. It is possible to centre
the game around a newer star such as Leonardo DiO.aprio. These film stills, running clockwise, show
that in this case there are at most three degrees of separation between DiCaprio and Helena
Bonham-Carter. through Kale Winsfet I Titanic. Columbia TriStar; Sense and Sensibility. Columbia
TriStar), Emma Thompson (Sense and Sensibility, Much Ado About Nothing, Entertainment Films)
and Kenneth Branagh (Much Ado About Nothing; Frankenstein', Columbia TriStar). Short cuts
between cliques couid be created in this game through some of DiCaprio's well-connected co-stars
such as Sharon Stone (The Quick and the Dead; TriStar; not shown).
0 PihHh Ltd tana 409
-------
Complex is Not the Same as Random
Environmental Protection
Agency
Regular Small-world Random
55 of 67
p~ 0 >• p~ 1
Increasing randomness
Watts and Strogatz (1998)
| Office of Research arid Development
-------
vvEPA
United States
Environmental Protection
Agency
Knowledge of Exposure Pathways Limits
High "hroughput Exposure Models
"In particular, the
assumption that 100%
of [quantity emitted,
applied, or ingested] is
being applied to each
individual use scenario
is a very conservative
assumption for many
compound / use
scenario pairs."
This is an open access amide published under an ACS AuthorChoice License. which pe-rmits
copying and redistribution of the article or any adaptations for non-commercial pur poses.
Article
pub5_aC5vOrg/est
Risk-Based High-Throughput Chemical Screening and Prioritization
using Exposure Models and in Vitro Bioactivity Assays
Hyeong-Moo Shin,*' Alexi Emstoff,*'^ Jon A. Arnot,"_L* Barbara A Wetmore,V Susan A Csiszar,s
Peter Fantke," Xianming Zhang,0 Thomas E, McKone,^'^ Olivier Jolliet/ and Deborah H. Bennett'
56 of 67
Office of Research arid Development
-------
oEPA
United States
Environmental Protection
Agency
Chemical Use Identi les Relevant Pathways
>2000 chemicals with Material Safety Data Sheets
(MSDS) in CPCPdb (Goldsmith etol., 2014)
CO
ro
u
0)
u
IS)

< "i
Apparel
Auto and Tires
Baby
Beauty
Craft and Party
Electronics
Grocery
Health
Home
Home Improvement
Patio and Garden
Pets
Sports and Outdoors
T oys
Some pathways have
much higher average
exposures!
Near Field Occ
Ecological
| Office of Research arid Development
Near field sources have been known to be important at least since 1987 -
see Wallace, et al.
-------
oEPA
United States
Environmental Protection
Arz^n
ELSEVIER
Contents lists available ail ScienceDirect
Food and Chemical Toxicology
journal homepage: www.elsevier.com/locate/foodchemtox
How Can we Know Chemical Use?
Chemical Property NAMs
Development of a consumer product ingredient database for chemical ,•»
exposure screening and prioritization
M.-R. Goldsmith'*, CM. Ctulke1, R.D. Brooksb, T.R. Transuec, Y.M. Tan A. Frame'10, P.P. Egeghy ',
R. Edwards d, D.T. Chang8, R. Tornero-VelezK. Isaacs3, A. Wang3C, J. Johnsona, K. Holm3, M. Reich',
J. Mitchells, DA Valleroa. L Phillipsa, M. Phillipsa, J.F. Wambaugha. R.S. Judsona,
T.J. Buckley 3, C.C. Dary
Occurrence and
quantitative
chemical composition
58 of 67
| office of Research and Development slide from Kristin Isaacs httpSl//cOIYiptOX#0pcl.gOV/cl«lshbo«l.FCl
-------
vvEPA
United States
Environmental Protection
Agency
Goldsmith et al. (2014):
• ~20,000
product-
specific
Material
Safety Data
Sheets (MSDS)
curated
• "2,400
chemicals
Product-specific
uses determined
using web spider
to click through
categories (e.g.,
home goods, bath
soaps, baby) to
find each product
CPCPdb: Material Safety Data Sheets
Material Safety
> .. 1 Data Sheet
1 Product: * soap scum remover & disinfectant 3 /
Description: pale blue to blue-green liquid with herbal pine odor
Other Designations
Manufacturer
Emergency Telephone No.
: SOAP SCUM REMOVER

For Medical Emergencies, call
Rocky Mountain Poison Center: 1-800-446-1014
For Transportation Emergencies, call:
Chemtrec: 1-800-424-9300
II Health Hazard Data
III Hazardous Ingredients
Eye irritant. Prolonged inhalation of vapors or mist may cause respiratory
irritation. There are no known medica! conditions aggravated by exposure
to this product
FtRST AID: EYE CONTACT: Immediately flush eyes with plenty of water
fnr 15 minutes. If irritation oersists. call a chvsician. INHALATION: If
hroathinn is affected breathe fresh air. SKIN CONTACT: Remove
contaminated clothing. Rush skin with water. If imt25'cn persists, call a
Dhvsician. IF SWALLOWED: Drink a glassful of water and immediately
call a physician.
Inoredient Concentration Worker Exoosure Limit
Tetrasodium ethylenediarr.ina < 10% none establish
tetra acetate (EDTA)
CAS #64-02-8
Glycol ether solvent < 8% none established
Cationic/nonionic surfactants < 5% none established
Trisodium nitrilotriacetate 0.14% none established
CAS #5064-31-3
This product contains trisodium nitrilotriacetate. IARC and NT? list
nitrilotriacetic acid (NTA) and its sodium salts as potential carcinogens.
IV Special Protection and Precautions
V Transportation and Regulatory Data
Do not get in eyes, on skin, or on clothing.
Avoid contact with food.
U.S. DOT Hazard Class: Not restricted
U.S. DOT Prooer Shiooina Name: Compound, cleaning, liquid
EPA CSRCLA/SARA TITLE III:
59 of 67
Office of Research and Development
-------
oEPA
United States
Environmental Protection
Arz^n
ELSEVIER
Contents lists available at ScienceDirect
Food and Chemical Toxicology
journal homepage: www.elsevier.com/locate/foodchemtox
How Can we Know Chemical Use?
Chemical Property NAMs
Development of a consumer product ingredient database for chemical
exposure screening and prioritization
M.-R. Goldsmith'*, CM. Ctulke1, R.D. Brooksb, T.R. Transuec, Y.M. Tan A. Frame'10, P.P. Egeghy ',
R. Edwards d, D.T. Chang8, R. Tornero-VelezK. Isaacs3, A. Wangac, J. Johnsona, K. Holm3, M. Reich',
J. Mitchells, DA Valleroa. L Phillipsa, M. Phillipsa, J.F. Wambaugha. R.S. Judsona,
T.J. Buckleya, C.C. Dary"
Occurrence and
quantitative
chemical composition
Broad "index" of chemical uses
ELSEVIER
Contents lists available at ScienceDirect
Toxicology Reports
journal homepage: www.elsevier.com/locate/toxrep
Exploring consumer exposure pathways and patterns of use ¦.
for chemicals in the environment
Kathie L Dionisioa, Alicia M. Frame11'1, Michael-Rock Goldsmitha-2,
John F. Wambaughb, Alan Liddell'-3, Tommy Catheyd, Doris Smith",
James Vailb, Alexi S. Ernstoff , Peter Fantkee, Olivier Jolliet',
60 of 67
| office of Research and Development slide from Kristin Isaacs httpSl//cOIYiptOX#0pcl.gOV/cl«lshbo«l.FCl
-------
oEPA
United States
Environmental Protection
Arz^n
ELSEVIER
Contents lists available at ScienceDirect
Food and Chemical Toxicology
journal homepage: www.elsevier.com/locate/foodchemtox
How Can we Know Chemical Use?
Chemical Property NAMs
Development of a consumer product ingredient database for chemical
exposure screening and prioritization
M.-R. Goldsmith'*, CM. Ctulke1, R.D. Brooksb, T.R. Transuec, Y.M. Tan A. Frame'10, P.P. Egeghy ',
R. Edwards d, D.T. Chang8, R. Tornero-VelezK. Isaacs3, A. Wangac, J. Johnsona, K. Holm3, M. Reich',
J. Mitchells, DA Valleroa. L Phillipsa, M. Phillipsa, J.F. Wambaugha. R.S. Judsona,
T.J. Buckleya, C.C. Dary"
Occurrence and
quantitative
chemical composition
Broad "index" of chemical uses
ELSEVIER
Contents lists available at ScienceDirect
Toxicology Reports
journal homepage: www.elsevier.com/locate/toxrep
Exploring consumer exposure pathways and patterns of use ¦.
for chemicals in the environment
Kathie L Dionisioa, Alicia M. Frame11'1, Michael-Rock Goldsmitha-2,
John F. Wambaughb, Alan Liddell'-3, Tommy Catheyd, Doris Smith",
James Vailb, Alexi S. Ernstoff , Peter Fantkee, Olivier Jolliet',
Journal of Exposure Science and Environmental Epidemiology (2018) 28, 216-222
0 2018 Nature America, Inc. past of Sponger Mature. AM rights reserved 1559-0631/18
w.nature.com ijes
ORIGINAL ARTICLE
Consumer product chemical weight fractions from
ingredient lists
Kristin K. Isaacs1, Katherine A. Phillips', Derya Biryol1J, Kathie L Dionisio1 and Paul S. Price5
61 of 67
office of Research and Development slide from Kristin Isaacs httpSj//comptOX.0p3..gOV/cl3.shbO3.l*cl
-------
oEPA
United States
Environmental Protection
Arz^n
ELSEVIER
Contents lists available at ScienceDirect
Food and Chemical Toxicology
journal homepage: www.elsevier.com/locate/foodchemtox
How Can we Know Chemical Use?
Chemical Property NAMs
Development of a consumer product ingredient database for chemical
exposure screening and prioritization
M.-R. Goldsmith'*, CM. Ctulke1, R.D. Brooksb, T.R. Transuec, Y.M. Tan A. Frame'10, P.P. Egeghy ',
R. Edwards d, D.T. Chang8, R. Tornero-VelezK. Isaacs3, A. Wangac, J. Johnsona, K. Holm3, M. Reich',
J. Mitchells, DA Valleroa, L Phillipsa, M. Phillipsa, J.F. Wambaugha. R.S. Judsona,
T.J. Buckleya, C.C. Dary"
Occurrence and
quantitative
chemical composition
ELSEVIER
Broad "index" of chemical uses
Contents lists available at ScienceDirect
Toxicology Reports
journal homepage: www.elsevier.com/locate/toxrep
Exploring consumer exposure pathways and patterns of use
for chemicals in the environment
Kathie L Dionisioa, Alicia M. Frame11'1, Michael-Rock Goldsmitha-2,
John F. Wambaughb, Alan Liddell'-3, Tommy Catheyd, Doris Smith"
James Vailb, Alexi S. Ernstoff , Peter Fantkee, Olivier Jolliet',
Journal of Exposure Science and Environmental Epidemiology (2018) 28, 216-222
0 2018 Nature America, Inc. past of Sponger Mature. AM rights reserved 1559-0631/18
www.nature.com/jes
ORIGINAL ARTICLE
Consumer product chemical weight fractions from
ingredient lists
Kristin K. Isaacs1, Katherine A. Phillips', Derya Biryol1-2, Kathie L Dionisio1 and Paul S. Price5
*
Clie ThUi Environ. SCr. Tethnoi. 2018. 52, 3125-3135
pubsjcs.org/est
Suspect Screening Analysis of Chemicals in Consumer Products
Katherine A Phillips.1 Alice Yau, Kristin A Favela, Kristin K Isaacs/ Andrew McEachran, "
Christopher Grulke," Ann M. Richard,' Antony J. Williams," Jon R. Sobus, Russell S. Thomas,"
and John F. Wambaugh*"
r Measurement of chemicals in
| Office of Research and Development Slide from Kristin Isaacs
consumer products
https://comptox.epa.gov/dashboard
-------
oEPA
United States
Environmental Protection
Arz^n
ELSEVIER
Contents lists available at ScienceDirect
Food and Chemical Toxicology
journal homepage: www.elsevier.com/locate/foodchemtox
How Can we Know Chemical Use?
Chemical Property NAMs
Development of a consumer product ingredient database for chemical
exposure screening and prioritization
M.-R. Goldsmith'*, CM. Grulke3, R.D. Brooksb, T.R. Transuec, Y.M. Tan A. Frame'10, P.P. Egeghy ',
R. Edwards d, D.T. Chang8, R. Tornero-VelezK. Isaacs3, A. Wangac, J. Johnsona, K. Holm3, M. Reich',
J. Mitchells, DA Valleroa, L Phillipsa, M. Phillipsa, J.F. Wambaugha. R.S. Judsona,
T.J. Buckleya, C.C. Dary"
Occurrence and
quantitative
chemical composition
Green Chemistry
ELSEVIER
Broad "index" of chemical uses
Contents lists available at ScienceDirect
Toxicology Reports
journal homepage: www.elsevier.com/locate/toxrep
Exploring consumer exposure pathways and patterns of use
for chemicals in the environment
Kathie L Dionisioa, Alicia M. Frame11'1, Michael-Rock Goldsmitha-2,
John F. Wambaughb, Alan Liddell'-3, Tommy Catheyd, Doris Smith"
James Vailb, Alexi S. Ernstoff , Peter Fantkee, Olivier Jolliet',
Journal of Exposure Science and Environmental Epidemiology (2018) 28, 216-222
0 2018 Nature America, Inc. past of Sponger Mature. AM rights reserved 1559-0631/18
www.nature.com/jes
ORIGINAL ARTICLE
Consumer product chemical weight fractions from
ingredient lists
Kristin K. Isaacs1, Katherine A. Phillips', Derya Biryol1-2, Kathie L Dionisio1 and Paul S. Price5
*
High-throughput screening of chemicals as
functional substitutes using structure-based
classification modelsf
Katherine A. Phillips, John F. Wambaugh. Christopher M. Grulke.
Kathie L. Dionisioc and Kristin K. Isaacsc
The roles that
chemicals serve in
products
Clie ThUi Environ. SCr. Tethnoi. 2018. 52, 3125-3135
Suspect Screening Analysis of Chemicals in Consumer Products
Katherine A Phillips.1 Alice Yau, Kristin A Favela, Kristin K Isaacs/ Andrew McEachran,
Christopher Grulke," Ann M. Richard,' Antony J. Williams," Jon R. Sobus, Russell S. Thomas,
and John F. Wambaugh*"
r Measurement of chemicals in
| Office of Research and Development Slide from Kristin Isaacs
consumer products
https://comptox.epa.gov/dashboard
-------
oEPA
United States
Environmental Protection
Arz^n
ELSEVIER
Contents lists available at ScienceDirect
Food and Chemical Toxicology
journal homepage: www.elsevier.com/locate/foodchemtox
How Can we Know Ci SCIENTIFIC DATA:
Chemical Proper
Development of a consumer product ingredient database for chemical
exposure screening and prioritization
M.-R. Goldsmith'*, CM. Grulke3, R.D. Brooksb, T.R. Transuec, Y.M. Tan A. Frame'10, P.P. Egeghy ',
R. Edwards d, D.T. Chang8, R. Tornero-VelezK. Isaacs3, A. Wangac, J. Johnsona, K. Holm3, M. Reich',
J. Mitchells, DA Valleroa, L Phillipsa, M. Phillipsa, J.F. Wambaugha. R.S. Judsona,
T.J. Buckleya, C.C. Dary"
Occurrence and
quantitative
chemical composition
ioiJo
Olllloi
www.
OPEN
s 2017
Accepted: 30 April 2018
Published: 10 July 2018
Green Chemistry
ELSEVIER
Broad "index" of chemical uses
Contents lists available at ScienceDirect
Toxicology Reports
journal homepage: www.elsevier.com/locate/toxrep
Exploring consumer exposure pathways and patterns of use
for chemicals in the environment
Kathie L Dionisio11, Alicia M. Frame11'1, Michael-Rock Goldsmitha-2,
John F. Wambaughb, Alan Liddell'-3, Tommy Catheyd, Doris Smith"
James Vailb, Alexi S. Ernstoff , Peter Fantkee, Olivier Jolliet',
Journal of Exposure Science and Environmental Epidemiology (2018) 28, 216-222
0 2018 Nature America, Inc. past of Sponger Mature. AM rights reserved 1559-0631/18
www.nature.com/jes
ORIGINAL ARTICLE
Consumer product chemical weight fractions from
ingredient lists
Kristin K. Isaacs1, Katherine A. Phillips', Derya Biryol1-2, Kathie L Dionisio1 and Paul S. Price5
*
Data Descriptor: The Chemical and
Products Database, a resource for
exposure-relevant data on
chemicals in consumer products
Kathie L. Dionisio1, Katherine Phillips1, Paul S. Price1, Christopher M. Grulke2,
Antony Williams2, Derya Biryol1'3, Tao Hong* & Kristin K. Isaacs1
tOF CHEMISTRY
High-throughput screening of chemicals as
functional substitutes using structure-based
classification modelsf
Katherine A. Phillips,**c John F. Wambaugh.b Christopher M. Grutke.b
Kathie L. Dionisioc and Kristin K. Isaacsc
The roles that
chemicals serve in
products
Clie ThUi Environ. SCr. Tethnoi. 2018. 52, 3125-3135
Suspect Screening Analysis of Chemicals in Consumer Products
Katherine A Phillips.1 Alice Yau, Kristin A Favela, Kristin K Isaacs/ Andrew McEachran,
Christopher Grulke," Ann M. Richard,' Antony J. Williams," Jon R. Sobus, Russell S. Thomas,
and John F. Wambaugh*"
r Measurement of chemicals in
| Office of Research and Development Slide from Kristin Isaacs
consumer products
https://comptox.epa.gov/dashboard
-------
Exposure NAM: Machine Learning to Fill Data Gaps
Environmental Protection EXAMPLE; Predicting Function Based on Structure
>=>
Prediction of
Of Potential
Alternatives from
Chemical Libraries
65 of 67
| Office of Research and Development
Machine Learning Based Classification Models
(Random Forest, Breiman, 2001)
Phillips et al. (2017)
additive
for liquid
system
additive
for rubber
adhesion
promoter
anti-
microbial
anti-
oxidant
chelator
colorant
crosslinker
film
forming
agent
flame
retardant
flavorant
heat
stabilizer
masking
agent
lubricating
agent
humectant
organic
pigment
oxidizer
perfumer
stabilizer
viscosity
controlling
agent
ubiquitous
absorber
Use Database (FUSE)
additive
buffer
emulsion
stabilizer
hair condi-
tioner
catalyst
foam
boosting
agent
oral care
photo-
initiator
plasticizer
pre-
servative
reducer
rheology
modifier
skin condi-
tioner
skin
protectant
soluble
dye
solvent
wetting
agent
white ner
Chemical Structure and
Property Descriptors
emollient
foamer
emulsifier
fragrance
surfactant
antistatic
agent
-------
oEPA
What is "High Throughput"?
United States
Environmental Protection
Agency
• Tox21: Testing one assay across 10,000 chemicals takes 1-2 days, but only 50 assays have been
developed so far that can run that fast
• ToxCast: ~1100 off-the-shelf (pharma) assay-endpoints tested for up to 4,000 chemicals over the past
decade, now developing new assays as well
HTS tox assays often use single readout, such as fluorescence, across many chemicals, measuring
concentration for toxicokinetics or exposure requires chemical-specific methods...
• ExpoCast: Ring et al. made in silico predictions for ~480,000 chemicals from structure, but based on
NHANES monitoring for ~120 chemicals
• Quantitative non-targeted analysis (NTA) may eventually provide greater evaluation data to
reduce uncertainty
• HTTK: In vitro data on 944 chemicals collected for humans, starting with Rotroff et al. (2010)
• Work continues to develop in silico tools, e.g. Sipes et al. (2016)
66 of 67
Office of Research and Development
Our work is not done
-------
oEPA
United States
Environmental Protection
Agency
Summary
A tapestry of laws covers the chemicals people are exposed to
in the United States (Breyer, 2009)
Many chemicals, ranging from industrial waste to dyes to
packing materials, are covered by the recently updated Toxic
Substances Control Act (TSCA) and administered by the EPA
New approach methodologies (NAMs) are being developed to
prioritize these existing and new chemicals for testing
All data are being made public:
• The CompTox Chemicals Dashboard (A search engine for
chemicals) http://comptox.epa.gov/
• R package "httk": https://CRAN.R-project.org/package=httk
| Office of Research and Development
mg/kg BW/day
Potential hazard
from in vitro
converted to dose
by HTTK
Potential
Exposure Rate
Lower
Risk
Medium
Risk
Higher
Risk
The views expressed in this presentation are those of the authors
and do not necessarily reflect the views or policies of the U.S. EPA
-------
ExpoCast reject
Exposure Forecasting)
Center for Computational Toxicology and Exposure
Linda Adams
Miyuk Breen*
Alex Chao*
Daniel Dawson*
Mike Devito
Kathie Dionisio
Christopher Eklund
Peter Egeghy
Marina Evans
Chris Grulke
Hongtai Huang*
Mike Hughes
Kristin Isaacs
Ashley Jackson*
Richard Judson
Jen Korol-Bexell*
Anna Kreutz*
Charles Lowe*
Katherine Phillips
Ann Richard
Risa Sayre*
Mark Sfeir*
Jane Ellen
Simmons
Marci Smeltz*
Jon Sobus
Mike Tornero-Velez
Rusty Thomas
Elin Ulrich
Dan Vallero
Barbara Wetmore
John Wambaugh
Antony Williams
3
TMjL n i

SwRi cypnotex
Center for
Environmental
Measurement
and Modeling
Hongwan Li
Xiaoyu Liu
Seth Newton
John Streicher*
Mark Strynar
Southwest Research Institute
an evoTec company
^Trainees
Collaborators
Arnot Research and Consulting
Jon Arnot
Johnny Westgate
Integrated Laboratory Systems
Kamel Mansouri
Xiaoqing Chang
National Toxicology Program
Steve Ferguson
Nisha Sipes
RamboOl
Harvey Clewell
Silent Spring Institute
Robin Dodsor
Simulations Plus
Michael Lawless
Southwest Research Institute
Alice Yau
Kristin Favela
Summit Toxicology
Lesa Ay I ward
Technical University of Denmark
PeterFantke
ToxStratejies
Caroline Ring
Unilever
Beate Nicol
Cecilie Rendal
Ian Sorrell
United States Air Force
HeatherPangburn
Matt Linakis
University of California, Davis
Deborah Bennett
University of Michigan
Olivier Jolliet
University of Texas, Arlington
Hyeong-Moo Shin
-------
References
oEPA
United States
Environmental Protection
Agency
Arnot, Jon A., et al. "Screening level risk assessment model for chemical fate and effects in
the environment." Environmental science & technology 40.7 (2006): 2316-2323.
Aylward, Lesa L., and Sean M. Hays. "Consideration of dosimetry in evaluation of ToxCast™
data." Journal of Applied Toxicology 31.8 (2011): 741-751.
Breyer, Stephen. Breaking the vicious circle: Toward effective risk regulation. Harvard
University Press, 2009
Cohen Hubal, EA, et al. "Advancing internal exposure and physiologically-based
toxicokinetic modeling for 21st-century risk assessments." Journal of exposure science &
environmental epidemiology (2018).
Collins FS, Gray GM, Bucher JR. Transforming environmental health protection. Science.
2008;319:906-907.
Dionisio, Kathie L., et al. "Exploring consumer exposure pathways and patterns of use for
chemicals in the environment." Toxicology reports 2 (2015): 228-237.
Dionisio, Kathie L., et al. "The Chemical and Products Database, a resource for exposure-
relevant data on chemicals in consumer products." Scientific data 5 (2018): 180125.
Dix David, et al. "The ToxCast program for prioritizing toxicity testing of environmental
chemicals." Toxicol Sci. 2007;95:5-12
Egeghy, P. P., et al. (2012). The exposure data landscape for manufactured chemicals.
Science of theTotal Environment, 414,159-166.
Filer, Dayne L., et al. "tcpl: the ToxCast pipeline for high-throughput screening data."
Bioinformatics 33.4 (2016): 618-620.
Goldsmith, M-R., et al. "Development of a consumer product ingredient database for
chemical exposure screening and prioritization." Food and chemical toxicology 65 (2014):
269-279.
Hertzberg, R. P., & Pope, A. J. (2000). High-throughput screening: new technology for the
21st century. Current opinion in chemical biology, 4(4), 445-451.
Isaacs, Kristin K., et al. "Consumer product chemical weight fractions from ingredient lists."
Journal of Exposure Science and Environmental Epidemiology 28.3 (2018): 216.
Jamei, et al. "The Simcyp® population-based ADME simulator." Expert opinion on drug
metabolism & toxicology 2009b;5:211-223
Judson, Richard, et al. "The toxicity data landscape for environmental chemicals."
Environmental health perspectives 117.5 (2008): 685-695.
Kaewkhaw, R., et al. (2016). Treatment paradigms for retinal and macular diseases using 3-
D retina cultures derived from human reporter pluripotent stem cell linestreatment design
using PSC-Derived 3-D retina cultures. Investigative ophthalmology & visual science, 57(5),
ORSFI1-ORSFI11.
Kavlock, Robert, et al. "Update on EPA's ToxCast program: providing high throughput
decision support tools for chemical risk management." Chemical research in toxicology
25.7(2012): 1287-1302.
Kavlock, R. J., et al. (2018). Accelerating the pace of chemical risk assessment. Chemical
research in toxicology, 31(5), 287-290
MacLeod, Matthew, et al. "The state of multimedia mass-balance modeling in
environmental science and decision-making." (2010): 8360-8364.
Mansouri, Kamel, et al. "OPERA models for predicting physicochemical properties and
environmental fate endpoints." Journal of cheminformatics 10.1 (2018): 10.
McNally, et al., "PopGen: a virtual human population generator." Toxicology 2014
National Research Council. (1983). Risk Assessment in the Federal Government: Managing
the Process Working Papers. National Academies Press.
National Research Council. (2007). Toxicity testing in the 21st century: a vision and a
strategy. National Academies Press.
National Research Council. Exposure Science in the 21st Century: a Vision and a Strategy.
National Academies Press, 2012.
Park, Youngja, H., et al. "High-performance metabolic profiling of plasma from seven
mammalian species for simultaneous environmental chemical surveillance and bioeffect
monitoring." Toxicology 295:47-55 (2012)
Pearce, Robert, et al. "httk: R Package for High-Throughput Toxicokinetics." Journal of
Statistical Software, 2017
Phillips, Katherine A., et al. "High-throughput screening of chemicals as functional
substitutes using structure-based classification models." Green Chemistry 19.4 (2017):
1063-1074.
Phillips, Katherine A., et al. "Suspect screening analysis of chemicals in consumer
products." Environmental science & technology 52.5 (2018): 3125-3135.
Ring, Caroline L., et al. "Identifying populations sensitive to environmental chemicals by
simulating toxicokinetic variability." Environment International 106 (2017): 105-118.
Ring, Caroline L., et al. "Consensus modeling of median chemical intake for the US
population based on predictions of exposure pathways." Environmental science &
technology 53.2 (2018): 719-732.
Rotroff, Daniel M., et al. "Incorporating human dosimetry and exposure into high-
throughput in vitro toxicity screening." Toxicological Sciences 117.2 (2010): 348-358
Schmidt, Charles W. "TOX 21: new dimensions of toxicity testing." Environmental health
perspectives 117.8 (2009): A348.
Shibata, Yoshihiro, et al. "Prediction of hepatic clearance and availability by cryopreserved
human hepatocytes: an application of serum incubation method." Drug Metabolism and
disposition 30.8 (2002): 892-896.
Shin, Hyeong-Moo, et al. "Risk-based high-throughput chemical screening and
prioritization using exposure models and in vitro bioactivity assays." Environmental science
& technology 49.11 (2015): 6760-6771.
Sipes, Nisha S., et al. "An intuitive approach for predicting potential human health risk with
theTox2110k library." Environmental science & technology 51.18 (2017): 10786-10796.
US Congress. "Frank R. Lautenberg Chemical Safety forthe 21st Century Act." (2016).
U.S. E.P.A. (2018) "A Working Approach for Identifying Potential Candidate Chemicals for
Prioritization." https://www.epa.gov/assessing-and-managing-chemicals-under-
tsca/identifving-existing-chemicals-prioritization-under-tsca
U.S. G.A.O.. "Toxic substances: EPA has increased efforts to assess and control chemicals
but could strengthen its approach." (2013).
Wallace, Lance A., et al. "The TEAM study: personal exposures to toxic substances in air,
drinking water, and breath of 400 residents of New Jersey, North Carolina, and North
Dakota." Environmental research 43.2 (1987): 290-307.
Wambaugh, John F., et al. "High-throughput models for exposure-based chemical
prioritization in the ExpoCast project." Environmental science & technology 47.15 (2013):
8479-848.
Wambaugh, John F., et al. "High Throughput Heuristics for Prioritizing Human Exposure to
Environmental Chemicals." Environmental science & technology (2014).
Wambaugh, John F., et al. "Toxicokinetic triage for environmental chemicals." Toxicological
Sciences 147.1 (2015): 55-67.
Wambaugh, John F., et al. "Evaluating in vitro-in vivo extrapolation of toxicokinetics."
Toxicological Sciences 163.1 (2018): 152-169.
Wambaugh, John F., et al. "Assessing Toxicokinetic Uncertainty and Variability in Risk
Prioritization" Toxicological Sciences (2019), in press
Wambaugh, John F., et al. "New Approach Methodologies for Exposure Science." Current
Opinion in Toxicology (2019).
Wang, Ying-Hong. "Confidence assessment of the Simcyp time-based approach and a static
mathematical model in predicting clinical drug-drug interactions for mechanism-based
CYP3A inhibitors." Drug Metabolism and Disposition 38.7 (2010): 1094-1104.
Waters, Nigel J., et al. "Validation of a rapid equilibrium dialysis approach forthe
measurement of plasma protein binding." Journal of pharmaceutical sciences 97.10(2008):
4586-4595.
Wetmore, Barbara A., et al. "Integration of dosimetry, exposure and high-throughput
screening data in chemical toxicity assessment." Tox. Sciences (2012)
Wetmore, Barbara A., et al. "Incorporating high-throughput exposure predictions with
dosimetry-adjusted in vitro bioactivity to inform chemical toxicity testing." Toxicological
Sciences 148.1 (2015): 121-136.
Office of Research and Development
-------