EPA’s National Center for Computational Toxicology is developing automated workflows for curating large databases within the DSSTox project, and providing accurate linkages of data to chemical structures, exposure and hazard information. The data are made available via the EPA’s CompTox Chemicals Dashboard (https://comptox.epa.gov/dashboard), a publicly accessible website providing access to data for ~880,000 chemical substances, the majority of these represented as chemical structures. The web application delivers a wide array of computed and measured physicochemical properties, in vitro high-throughput screening data and in vivo toxicity data, as well as integrated chemical linkages to a growing list of literature, toxicology, and analytical chemistry websites. In addition, several specific search types are in development to directly support the mass spectrometry non-targeted screening community, enabling cohesive workflows to support data generation for the detection and assessment of environmental exposures to chemicals contained within the database. The application provides access to segregated lists of chemicals that are of specific interest to relevant stakeholders, including, for example, scientists interested in Per- & Polyfluoroalkyl Substances (PFAS). Added lists include those sourced from the European Union as well as developed in-house and now containing thousands of chemicals. A procured testing library of hundreds of PFAS chemicals annotated into chemical categories has been integrated into the dashboard with a number of resulting benefits: a searchable database of chemical properties, with hazard and exposure predictions, and links to the open literature. This presentation will provide an overview of the dashboard, the developing library of PFAS chemicals and associated categorization, and new physicochemical property and environmental fate and transport QSAR prediction models developed for these chemicals. The application of the dashboard to support mass spectrometry non-targeted analysis studies for the identification of PFAS chemicals will also be reviewed. This abstract does not necessarily represent the views or policies of the U.S. Environmental Protection Agency.

1. PFAS Chemistry: Range, Complexity,
Groupings, and the CompTox
Chemicals Dashboard
Antony Williams
Center for Computational Toxicology and Exposure, U.S.-EPA, RTP, NC
September 2020
The Science of PFAS, Virtual Conference
http://www.orcid.org/0000-0002-2668-4821
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

2. CompTox Chemicals Dashboard
• A publicly accessible website delivering access:
– ~882,000 chemicals with related property data
– Experimental and predicted physicochemical property data
– Experimental Human and Ecological hazard data
– Integration to “biological assay data” for 1000s of chemicals
– Information regarding consumer products containing chemicals
– Links to other agency websites and public data resources
– “Literature” searches for chemicals using public resources
– “Batch searching” for thousands of chemicals
– Over 10,000 of the chemicals are classed as PFAS Chemicals
1

3. CompTox Chemicals Dashboard
https://comptox.epa.gov/dashboard
2
• Searching – CAS Numbers, systematic names
and synonyms, structures (as InChIs)

4. Substring search “perfluoro”
3
• Substring search ~2000 chemicals

5. Substring search “perfluoro”
“Explicit” structures
4

6. Substring search “perfluoro”
“Markush” representations
5

7. Substring search “perfluoro”
“UVCB” chemicals
6
• Unknown or Variable Composition, Complex
Reaction Products and Biological Materials

8. 1 of ~882,000 Chemical Pages
7

9. Physicochemical properties
8

10. Experimental Data
9

11. Detailed QSAR Prediction Reports
10

12. Training sets enhanced QSAR
11
• Comparison of COSMOtherm, EPI Suite
ACD/Labs, TEST and OPERA
• OPERA best performance: Vapor Pressure,
Solubility, Octanol-water partitioning, Octanol-
Air partitioning, Soil-Adsorption coefficient

13. Hazard Data – Human and Eco
12

14. What is PFOS Called?
Synonyms, CASRNs and more
13

15. Substance Relationship Mappings
• Similar compounds - based on structure
“fingerprints”
• Structure mappings - between parent and
salts, multicomponent chemicals, isotopomers
• Related substances – monomer to polymer,
parent to transformation products
14

16. Are there Similar Compounds?
128 chemicals >0.8 match factor
15

17. Relationships in the data
16

18. 9 salt forms of PFOS (and the ion)
17

19. CHEMICAL
LISTS OF PFAS
18

20. A List of Lists of Chemicals
https://comptox.epa.gov/dashboard/chemical_lists
19

21. Chemical Lists
• Assembled chemical lists give access to
curated data
– Names and synonyms
– Physicochemical/Fate and Transport data
– Toxicity data
– Relationships in the data
– Regulatory lists
20

22. The OECD List of PFAS
http://www.oecd.org/chemicalsafety/portal-perfluorinated-chemicals/
21

23. The OECD List of PFAS
http://www.oecd.org/chemicalsafety/portal-perfluorinated-chemicals/
22

24. Building a “Master PFAS List”
• The definition of PFAS can differ across
research groups and scientific publications.
• There is no consistent definition across the
scientific community
• We have an evolving approach for a list
– Substances with specific substructural elements as
chemical “structures”
23

25. PFAS Structure List (8163)
24

26. Building a “Master PFAS List”
• The definition of PFAS can differ across
research groups and scientific publications.
• There is no consistent definition across the
scientific community
• We have an evolving approach for a list
– Substances with specific substructural elements as
chemical “structures”
– Substances with specific “substrings” to represent
PFAS elements in UVCB chemicals
25

27. UVCB Chemicals
26

28. PFAS “UVCB Chemicals”
27

29. Example PFAS-UVCBs
28

30. 1-Propene, 1,1,2,3,3,3-hexafluoro-, polymer with 1,1-
difluoroethene, ethene, 1,1,2,2-tetrafluoroethene and
1,1,2-trifluoro-2-(trifluoromethoxy)ethene
29

31. “Markush” Chemical Categories
• PFOS is a linear perfluoroalkyl sulfonate
30

32. …and their Markush Children…
• The Markush form of linear perfluoroalkyl sulfonates has children…
31

33. PFAS Categories - evolves with
new classes as added
32

34. PFAS Categories in Development
(112 categories so far… and evolving)
33

35. BATCH
SEARCHING
34

36. Batch Search
• Search thousands of chemicals based on
CASRN, names and identifiers
• Harvest en masse the data available for
single chemicals – properties, tox data,
chemical relationships, category mappings,
presence in lists
35

37. Batch Searches
36

38. SUPPORTING
MASS
SPECTROMETRY
37

39. 38

40. Advanced Search
Supporting Target/Non-Target MS
39

41. 2 Chemicals match the formula
40

42. Advanced Search
Supporting Target/Non-Target MS
41

43. 23 Chemicals match the formula
42

44. Supporting future work
• 1. Mobility: A wide and dynamic distribution of short chain PFAS due to
their high polarity, persistency and volatility. (QSAR Predictions)
• 2. Substitution of regulated substances: The ban or restrictions of
individual molecules will lead to a replacement with substitutes of
similar concern. (Database content and Markush Enumeration)
• 3. Increase in structural diversity of existing PFAS molecules:
Introduction of e.g., hydrogens and chlorine atoms instead of fluorine,
as well as branching and cross-linking lead to a high versatility of
unknown target molecules. (Database content)
• 4. Unknown “Dark Matter”: The amount, identity, formation pathways,
and transformation dynamics of polymers and PFAS precursors are
largely unknown. (Working with agency analytical scientists and
collaborators to link and host data)
43Front Chem. 2018; 6: 103.

45. Conclusions
• CompTox Chemicals Dashboard supports PFAS
research at EPA in numerous ways
– Delivery of curated lists of PFAS chemicals (growing)
– Flexible search capabilities – support for Mass Spec
– Relationships in the data enrich navigation between chemicals
• Ongoing research efforts for PFAS chemicals
– Continue harvesting physicochemical & fate and transport data
– Classification approaches and Markush representations
– Expand available toxicity data and integration to systematic
review data as it becomes available
44

46. Acknowledgements
EPA-RTP
• Ann Richard, Chris Grulke and
Grace Patlewicz
• An enormous team of contributors
from CCTE, especially the IT
software development team
• Our curation team for their care
and focus on data quality

47. For More Information
• The research discussed in this presentation
is part of EPA’s overall efforts to rapidly
expand the scientific foundation for
understanding and managing risk from PFAS.
• For more information on EPA’s efforts to
address PFAS, please visit the following
websites
– EPA PFAS Action Plan - https://www.epa.gov/pfas/epas-
pfas-action-plan
– EPA PFAS Research - https://www.epa.gov/chemical-
research/research-and-polyfluoroalkyl-substances-pfas
46

48. Contact
Antony Williams
NCCT, US EPA Office of Research and Development,
Williams.Antony@epa.gov
ORCID: https://orcid.org/0000-0002-2668-4821
47
https://doi.org/10.1186/s13321-017-0247-6