![Page 1: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/1.jpg)
MS Reference Libraries for Forensics: Past, Present and Future
Forensics@NIST 2012
Steve Stein et al.
NIST MS Data Center
![Page 2: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/2.jpg)
Identification A Central Task in Forensics
• People
– DNA, Fingerprints, Features, …
• Objects
– Clothing, Weapon, …
• Chemicals
– Molecular Identity
![Page 3: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/3.jpg)
Outline
• Library Background
• Nature of the Data
• Identification by MS
• NIST Tools
• Tandem MS
• Future
![Page 4: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/4.jpg)
Library Background
![Page 5: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/5.jpg)
NIH/EPA Collection of Collections
Fales, Heller
Red Books 9-track Tape
300 Baud Modem
To NIST
PC-XT Version
Evaluated Library
AMDIS Peptides
High Resolution
To EPA
Cincinnati
Budde
Structures
Begin Manual Evaluation
Tandem MS
GC Retention
2000’s 1990’s 1980’s 1970’s
Evolution of the NIST MS Library
2010’s
![Page 6: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/6.jpg)
0K
50K
100K
150K
200K
250K
300K
'78 '80 '83 '86 '88 '90 '93 '98 '02 '05 '08 '11
Numbers of EI Spectra
Replicates
Compounds
![Page 7: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/7.jpg)
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
EI Libraries Distributed/Year
50 Distributors
![Page 8: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/8.jpg)
Data Sources
• In the Beginning: Library of Libraries + Literature • Contractor Labs • NIST Measurements • Contributors
– Industry, Academics, Organizations, Crime Labs, …
• New Spectra (ca. 10,000 / year) – Derivatives of common chemicals – Metabolites (human and plant) – Environmental/Security – Newly regulated compounds
![Page 9: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/9.jpg)
![Page 10: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/10.jpg)
“Evaluation”
• Initial Manual Evaluation
• Chief Evaluator: Mark as Best, Alternate, Reject
• Spectrum + Structure Computer Processing
• Add to Archive → Build Library
![Page 11: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/11.jpg)
Nature of the Data
![Page 12: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/12.jpg)
“The Decomposition of Hydrocarbons in the Positive Ray Tube”
H.R. Stewart & A.R. Olson, 1931
JACS, 53, 1326
![Page 13: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/13.jpg)
Mass Spectra are Reproducible
O’Neal et al. Anal. Chem.
1951
NIST 2012
A mass spectrum is a property of an ion
![Page 14: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/14.jpg)
Identification by MS
![Page 15: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/15.jpg)
Identification by Pattern Matching
• Mass spectra are molecular ‘properties’
– Reflect molecular structure
• Peaks are easily formed stable fragments
– May not be unique to compound
![Page 16: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/16.jpg)
Identification by GC/MS
Match EI Mass Spectra and Retention Time
Compute Score
http://en.wikipedia.org/wiki/File:Goldkey_logo_removed.jpg
But, Identification is Indirect and Depends on the Analyte
![Page 17: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/17.jpg)
Bayes Rule*
False Negative Potential
Prior Probability: Before Experiment
False Positive Potential
Analyte is Identified Correctly
Final Confidence
Starting Confidence
Change in Confidence
* Odds Version
P ( Score | ID) P ( ID )
P ( FP ) P ( Score | FP) X
Influence of Library Search
P ( ID | Score )
P ( FP | Score )
![Page 18: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/18.jpg)
Class Identification or False Positive?
![Page 19: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/19.jpg)
chemdata.nist.gov
![Page 20: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/20.jpg)
Traditional Library Search
Query Spectrum
Library Spectrum Hit
List
Score Histogram
2011 Version - 213K EI, 5K CID, 71K RI Compounds
Search List
![Page 21: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/21.jpg)
Substructure Analysis
Chemical Substructure Identification by Mass-Spectral Library Searching JASMS 6 (8) 644-655 (1995)
![Page 22: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/22.jpg)
MS Interpreter
![Page 23: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/23.jpg)
AMDIS
Automated Mass Spectral Deconvolution and Identification System Created for Chemical Weapons Treaty Verification: “Blinded CW Identification”
JASMS 1999 10 770-781
![Page 24: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/24.jpg)
peptide.nist.gov
NISTMSQC: Full Analysis of LC-MS/MS data Library/quality metrics
“Performance Metrics for Liquid Chromatography-Tandem Mass Spectrometry Systems in Proteomics Analyses”, Molecular & Cellular Proteomics, 9, 225, 2010
![Page 25: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/25.jpg)
Tandem MS
![Page 26: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/26.jpg)
# Precursor Ions
2005
2008
2011
2012
NIST Tandem Mass Spectral Library 2012
Fragmentation Type Precursor Ions
Ion Trap >10,000
Beam Collision Cell (QTOF, QQQ, HCD)
>8,000
Classes: Metabolites, Drugs, Sugars, Phospholipids, Peptides, Surfactants, etc.
Precursors: [M+H]+, [M+2H]2+, [M-H]-, [M+Na]+, [M+NH4]+, [Cat]+, [An]-, [p-H2O], [p-NH3], etc.
New Software Features: • Exact or isotopic precursor mass & fragment ions. • Formats: mzXML, mzData, mgf, msp, dta, pkl, JCAMP, …. • Compatible with NIST EI & Peptide Tandem Libraries. • New methods for finding targets in the presence of noise.
New Scoring: • Compounds with few dominant peaks. • Compensates for m/z tuning errors.
Compounds 7,020
Precursor Ions 15,517
Spectra 123,781
![Page 27: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/27.jpg)
Emerging MS Methods 3,362 DART CID Spectra, 757 Compounds
Robert L. Steiner, Virginia Crime Lab Chip Cody, JEOL
http://chemdata.nist.gov/
![Page 28: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/28.jpg)
Future
![Page 29: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/29.jpg)
Future Work
• Algorithms – Accurate ID confidence
• ‘Recurrent’ Spectrum Libraries – Combine with IDs for all mixture components
– Substance-based libraries
• SRM/D – Reference Materials + Reference Data
![Page 30: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/30.jpg)
Algorithms with Wallace, Kearsley, Allison @NIST
• ‘Dot Product’ Function is Best Measure of Spectrum Similarity
• Using Spectrum Similarity Only Ignores:
– Chemical/Spectrum Class
– ‘Prior Probability’
• Secondary Scoring is Promising
– Use spectrum/compound class to re-score
– Adjust for Sample/Method
• Target: Identification Probability with Error Limits
![Page 31: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/31.jpg)
Goal: Interpret All Spectra
• GC/MS: Begin with AMDIS – Chemdata.NIST.Gov
• LC-MS/MS: Begin with NISTMS QC – Peptide.NIST.Gov
• Classify Each Spectrum:
– Identified – ‘Recurrent’ spectrum – Unknown compound – Mixture – Noise/Background
![Page 32: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/32.jpg)
SRD
• Standard Reference Materials
– Analytes (Cholesterol, Vitamin D)
• Often in Matrix (urine, plasma, …)
– Substances (Plasma Metabolites, …)
• Identify and value-assign many components
• Standard Reference Data
– MS Library, Thermodynamic Data, Chemistry Webbook, …
• Standard Reference Materials + Data
– Substance + Data + Libraries
SRD + SRM
SRMD.NIST.GOV
![Page 33: MS Reference Libraries for Forensics: Past, Present and Future](https://reader031.vdocuments.net/reader031/viewer/2022022206/621385c418c42c776b439777/html5/thumbnails/33.jpg)
NIST MS Data Center