The document discusses how the EPA's CompTox Chemicals Dashboard can be used to support mass spectrometry analyses for structure identification. The Dashboard contains data on over 800,000 chemicals including properties, lists, and links to other resources. It allows searching by formula, structure, and mass to find related chemicals. Candidate structures can be ranked using metadata. Predicted mass spectra from over 800,000 structures may also be accessible. The Dashboard integrates data to help identify unknown chemicals detected by mass spectrometry.

1. Structure identification approaches using
the EPA CompTox Chemicals Dashboard
to support mass spectrometry analyses
Antony Williams, Charles Lowe, Alex Chao, Elin Ulrich and Jon Sobus
Center for Computational Toxicology and Exposure, U.S. Environmental Protection Agency, RTP, NC
August 2021
ACS Fall Meeting, Atlanta
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. 1
EPA’s CompTox Chemicals Dashboard
A publicly accessible website delivering:
- ~883,000 chemicals with related property data
- Experimental and predicted physicochemical property data
- Integration to “biological assay data” for 1000’s 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
- Downloadable Open Data for reuse and repurposing

3. A single app integrating…
2
2
SEARCH
TOX DATA
BIOACTIVITY
SIMILARITY
READ-ACROSS
PUBMED
BATCH SEARCH

4. Detailed Chemical Pages
3

5. Sources of Exposure to Chemicals
4

6. Physicochemical properties and
environmental fate and transport
5

7. Link farm to public resources
6

8. Mass Spec Links
7

9. NIST WebBook
https://webbook.nist.gov/chemistry/
8

10. MassBank of North America
https://mona.fiehnlab.ucdavis.edu
9

11. Chemical lists
10

12. >300 Chemical Lists (and growing)
11

13. “Volatilome” Human Breath
12

14. “Volatilome” Saliva
13

15. Disinfection By-Products
14

16. Tire Crumb Rubber (298)
15

17. Hydraulic Fracturing (1640)
16

18. PFAS Lists
17

19. Related Searches
to Support Mass
Spectrometry
18

20. Find me “related structures”
Formula-Based Search
19

21. Select Chemicals of Interest
20

22. Find me “related structures”
Based on Structure Similarity
21

23. Find me “related structures”
Based on Structure Similarity
22

24. Find me “related structures”
Structure Similarity – sort on mass
23

25. Mass & Formula
Searching
24

26. Advanced Searches
Mass Search
25

27. Advanced Searches
Mass Search
26

28. MS-Ready Structures for
Formula Search
27

29. “MS-Ready Structures”
https://doi.org/10.1186/s13321-018-0299-2
28

30. 29

31. MS-Ready Mappings
• EXACT Formula: C10H16N2O8: 3 Hits
30

32. MS-Ready Mappings
• Same Input Formula: C10H16N2O8
• MS Ready Formula Search: 125 Chemicals
31

33. MS-Ready Mappings
• 125 chemicals returned in total
– 8 of the 125 are single component chemicals
– 3 of the 8 are isotope-labeled
– 3 are neutral compounds and 2 are charged
32

34. MS-Ready Mappings
33

35. MS-Ready Mappings Set
34

36. Candidate
ranking
35

37. Data Source Ranking of
“known unknowns”
36
• Mass and/or formula is for an
unknown chemical but contained
within a reference database
• Most likely candidate chemicals
have the most associated data
sources, most associated lit.
articles or both
C14H22N2O3
266.16304
Chemical
Reference
Database
Sorted candidate
structures

38. Is a bigger database better?
37
• ChemSpider was 26 million chemicals then
• Much BIGGER today
• Is bigger better??

39. Using Metadata for Ranking
• Use available metadata to rank candidates
– Associated data sources
• Associated lists in the underlying database
• Associated data sources in PubChem
• Specific types (e.g. water, surfactants, pesticides etc.)
– Number of associated literature articles (Pubmed)
– Chemicals in the environment – the number of
products/categories containing the chemical is a very
important source of data
38

40. Comparing Search Performance
39
• Dashboard content was 720k chemicals
• Only 3% of ChemSpider size
• What was the comparison in performance?

41. SAME dataset for comparison
40

42. How did performance compare?
41
For the same 162 chemicals,
Dashboard outperforms
ChemSpider

43. How did performance compare?
42

44. Will the correct Microcystin LR Stand Up?
ChemSpider Skeleton Search
43

45. Comparing ChemSpider Structures
44

46. Comparing ChemSpider Structures
45

47. Other Searches
46

48. Batch
Searches
47

49. List of Opioids – Presence in Lists?
48

50. Batch Search Names
49
Excel
Download

51. Batch Search in specific lists
50

52. Opioids and Metabolites (160)
51

53. Batch Searching
• We work with thousands of masses/formulae!
• Typical questions
– What is the list of chemicals for the formula CxHyOz
– What is the list of chemicals for a mass +/- error
– Can I get chemical lists in Excel files? In SDF files?
– Can I include properties in the download file?
52

54. Batch Searching Formula/Mass
53

55. Searching batches using MS-Ready
Formula (or mass) searching
54

56. Benefits of
Open Data
55

57. API services and Open Data
• Available API and web services
• Open Data available for download
56

58. Web Services
https://actorws.epa.gov/actorws/
• Dozens of web services to provide access
to data
• Data in UI, JSON and XML format
57

59. Example: InChIKey to DTXCIDs
58
https://actorws.epa.gov/actorws/dsstox/v02/msready?identifier
=UVOFGKIRTCCNKG-UHFFFAOYSA-N

60. MassBank mapping to Dashboard
59

61. NORMAN Suspect List Exchange
https://www.norman-network.com/?q=node/236
60

62. Integration to MetFrag in place
https://jcheminf.biomedcentral.com/articles/10.1186/s13321-018-0299-2
61

63. Work in
Progress
62

64. Prototype Work in Progress
• CFM-ID
– Viewing and Downloading pre-predicted spectra
– Search spectra against the database
• Structure/substructure/similarity search
• The EPA NTA WebApp
• Access to API and web services
63

65. Predicted Mass Spectra
http://cfmid.wishartlab.com/
• MS/MS spectra prediction for ESI+, ESI-, and EI
• Predictions generated and stored for >800,000
structures, to be accessible via Dashboard
64

66. • Predictions generated and stored for >800,000 structures
• Python code to score experimental vs predicted spectra
• Cosine dot product match score calculation
August 26, 2019
Nontargeted screening of wastewater for water reuse using mass spectrometry Current Advances in Water
Analysis
65
CFM-ID Predicted Library Available

67. Search Expt. vs. Predicted Spectra

68. Search Expt. vs. Predicted Spectra

69. Spectral Viewer Comparison
68

70. CASMI 2012-2017 revisited
• Application of metadata candidate ranking
and CFM-ID to all five years of CASMI data
69

71. 70

72. Published: Alex Chao et al
71

73. NTA data input
files
Parameters for
NTA clean-up /
database
searching
NTA
WebApp:
Input Page
Jeff
Minucci
NTA WebApp Development

74. Visualizations of Data – Cluster 1
Chemical
Candidate
Information

75. Patterns of Cluster Results
Cluster 1: Primarily drugs
(e.g. diphenhydramine,
labetalol, clindamycin)

76. Cluster 2: Likely xenobiotics
(ethanolamides)
Patterns of Cluster Results

77. Cluster 7: Mostly human
metabolism chemicals (acyl
carnitines, amino acids, peptides)
Patterns of Cluster Results

78. Prototype Development
77

79. Method Amenability Prediction
Charlie Lowe
Why?
• Chromatography-mass
spectrometry can be LC or GC
• Which phase is more appropriate
for which chemicals?

80. Ongoing Work
• Data sources to date
• Massbank of North America
• 9,275 chemicals for non-derivatized GC
• 846 chemicals for derivatized GC
• 816 chemicals for APCI+
• 454 chemicals for APCI-
• 4,907 chemicals for ESI+
• 3,430 chemicals for ESI-
• EPA Non-targeted Analysis Collaborative Trial (ENTACT)
• 886 chemicals for non-derivatized GC
• 44 chemicals for derivatized GC
• 774 chemicals for APCI+
• 431 chemicals for APCI-
• 1,113 chemicals for ESI+
• 648 chemicals for ESI-

81. Conclusion
• Dashboard access to data for ~883,000 chemicals
MS-Ready data facilitates structure identification
• Related metadata facilitates candidate ranking
80
• Relationship mappings and
chemical lists of great utility
• Curation and mutual
sharing of chemical lists is
important (e.g. NORMAN)

82. Acknowledgements
• NCCT IT development team
• Tommy Cathey, ACTOR Web Services
• Nancy Baker, Abstract Sifter
• Todd Martin & Valery Tkachenko,
WebTEST
• Kathie Dionisio & Kristin Isaacs, CPDat
• Thanks to Emma Schymanski, University
of Luxembourg, for coordinating all efforts
with the NORMAN Network for curation of
lists on the Suspect Exchange

83. Contact
Antony Williams
US EPA Office of Research and Development
National Center for Computational Toxicology
EMAIL: Williams.Antony@epa.gov
ORCID: https://orcid.org/0000-0002-2668-4821
82