At the 2026 American Society for Mass Spectrometry 74th Annual Meeting on Mass Spectrometry and Allied Topics held in San Diego, CA, the week of 31 May, the Mass Spectrometry Data Center of the National Institute of Standards and Technology (United States Department of Commerce) announced the 2026 release of the NIST\EPA\NIH Electron Ionization (EI) Mass Spectral Library, the NIST 26 Tandem Library of Product-ion Mass Spectra, and Version 4.0 of the NIST Mass Spectral Search Program at a breakfast meeting attended by nearly 100 people. This was the 13th release of the EI Library and the 7th release of the Tandem Library. Both showed substantial increase in the number of compounds represented (10% to 382K compounds for the EI Library and 33% to 69K compounds for the Tandem Library, Figure 1). With the ability to process entire GC/MS data files (nominal and accurate m/z values) and DDA product-ion tandem LC‑MS/MS data files, along with other enhancement such as the improvements to the Hybrid-17 search, MS Search exhibited its single greatest advancement since first introduced in 1988.
Figure 1. From slide presented during announcement showing the growth of the Tandem Library (Courtesy of NIST Presentation)
In the 1970s the National Institute of Standards and Technology (known as the National Bureau of Standard (NBS) until 1988) began the distribution of the National Institutes of Health’s (NIH) mass spectral library of ~12,500 electron ionization (EI) spectra to mass spectrometry manufacturers. NBS provided an annual rental arrangement to these manufactures with no direct distribution to end-users. This library was incorporated into the manufacturers’ dedicated minicomputer data system for matching against acquired spectra.
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
In 1975 the EI library that had been curated by the United States Environmental Protection Agency (EPA) was merged with the NIH library resulting in a distribution of ~25,000 spectra which was then copyrighted under NIST’s Standard Reference Data Act and distributed in the same way as the original NIH library by NBS. In 1987 EPA\NIH passed the responsibility of the combined library to NBS.[1] This was followed by creation of the Mass Spectrometry Data Center (MSDC) and Stephen E. Stein was appointed as its director. In 1989 the mass spectral library was renamed the NIST\EPA\NIH Mass Spectral Library. At that time, the distribution of the library was changed from the annual rental of a tape which could be used to create unlimited copies for end-users by instrument manufacturers in their proprietary formats to a sale of licenses to individual copies that could only be used on a single computer accessible by no more than one person at a time. Manufacturers were then responsible for the reporting of the sales of licenses.
The minicomputer-based data systems provided by most instrument manufacturers had the limitation of only being able to retrieve a library spectrum when a measured spectrum was submitted, and the library spectra were condensed due to limits on disk storage space. It was not possible to search the library by data other than m/z-intensity pairs of the measured spectra against the condensed spectra. One of the first tasks undertaken by Steve Stein was to develop the Mass Spectral (MS) Search Program using MS DOS and the newly introduced IBM PC microcomputer. At this time, December 1987, NBS began sales direct to end-users as well as providing the data to manufacturers on a per copy basis. The library had grown to about 44,000 EI spectra (only one spectrum per compound). The software and data were provided by NIST to end-users on 5¼ in, 360K floppy disk.[2] Each spectrum had a complete list of all peaks (m/z values and relative intensities, which included isotope peaks), the first 25 characters of a chemical name, in many cases a list of synonyms, molecular formula (elemental composition), molecular weight (nominal mass[3]), CAS registry numbers, and the source of the spectrum (the contributor). A quality Index (QI[4]) was assigned to the individual spectrum. The review of version 1.02 stated, “Regular updates of the database (what the library was called at that time) will consist of additional spectra as well as replacing spectra with higher quality data are planned.” The software was written in C and was as flexible as the mass spectral search systems that were currently available on commercial timeshare systems. The library could be searched by ID number (an internal numbering system code found on each spectrum), CAS registry number, chemical name (limited to the first 25 characters) formula (elemental composition, the appropriate atomic symbol must be used for the elements), molecular weight (nominal mass), abundance of up to 10 major peaks along with their m/z value (Any Peaks Search), and a user-defined search consisting of any of the specific search criteria (Sequential Search).
MS Search had the ability to compare a user-measured spectrum against a mass spectral library from its inception. Stein wrote an enhanced version of the INCOS Library Search, which, to this day, provides the most accurate search of a curated mass spectral library.
MS Search with a library of 49,500 spectra was first released to the public in 1988 (DOS v.2.0). Structures were added to the DOS version of MS Search with v.3.0, reported to have been released in 1990. At the 1990 Pittsburgh Conference (PITTCON), in New York City, PerkinElmer Corporation introduced the Q-MASS 910 GC-MS with the TurboChrom data system. One of the key features of this data system was its incorporation of the NIST MS Search Program v.3.0 with the NIST\EPA\NIH EI Library, which allowed the Hit List from a library search to be displayed as individual structures. Varian, Inc., who had just begun offering a quadrupole ion trap GC/MS based on the Finnigan quadrupole ion trap (QIT) technology, became aware of the MS DOS MS Search Program v.3.0 and began offering it with their GC/MS within the week.
In 1993, the policy of having only a single spectrum for each compound was changed to include replicates of compounds that have significance. The number of compounds has continued to grow. Figure 2 shows the growth since 1978. Prior to 1990 all spectra in the library were contributed from outside of NIST (NBS). Beginning in 1990, NIST started to measure spectra using a transmission quadrupole GC-MS. As of the NIST 26 release of the 431,280 spectra, 153,184 of these spectra have been measured at NIST.
Figure 2. Growth of what began as the NIH EI Mass Spectral Library (12,500 spectra) distributed by NBS and is now the NIST\EPA\NIH EI Mass Spectra Library (431,280 spectra of 382,185 compounds). Courtesy of NIST Presentation
In 1995, NIST introduced a single-screen Microsoft Windows version of the MS Search Program, replacing the MS DOS version, thus resetting the version number to 1. This was replaced in 2002 with the current multiple tab Microsoft® Windows® version (MS Search v.2.0). With each subsequent release, there have been improvements made to MS Search. MS Search v.4.0 has changed to the new style of Windows’ programs, which replaces tabbed multi functioned dialog boxes (right) with an item-selection version (left) shown in Figure 3.
Figure 3. Left: New Style Dialog; Right: Previous Style Dialog
Another enhancement to MS Search v.4.0 was the addition of a floating Structure window. Sometimes when structures are displayed with bar-graph spectra, visual distortion can occur due to the screen size and/or graphics resolution of the monitor. These distortions can lead to misinterpretation of the structure. This has been eliminated with this new feature. The floating structure is displayed by selecting the Show Structure item from the View menu displayed on the Main Menu Bar. Whatever is selected in the current tab view’s Hit List will have its structure displayed. This is especially valuable on crowded displayed structures such as those which exist with the Lib Search Chromatogram tab views (Figure 4).
Figure 4. Lib Search tab view with floating window. Notice how much clearer the structure is in the window as compared to the one overlaid on the Hit List bargraph spectrum.
In 1988, NIST tested the automated Quality Index (QI) and found that only 50% of the time it selected the better of two replicate spectra. As of the 1998 release, the QI was no longer included with the library’s spectra. By that time NIST had instituted the NIST Number. This is a unique number assigned to each spectrum that is added to the NIST Archive. This number has a one-time use. If the spectrum for a given compound is replaced in the distributed library, the replacement spectrum will have a new NIST Number. The replaced spectrum will remain in the NIST Archive and can be made available to end-users on request.
NIST created a database of GC Methods and Retention Index (RI) from the scientific. literature. This database was incorporated into the NIST 05 release. Like the mass spectral library this database has grown with each release. The NIST 26 release of the GC Method\Retention Index database contains 527K values for 216K compounds, 188,198 of these have mass spectra in the NIST\EPA\NIH EI Library. NIST is measuring the RI value and recording the GC Method of each compound measured. If a clear RI value is not obtained for a measurement and the quality of the mass spectrum is evaluated as true by two mass spectrometrists, the spectrum is entered into the library without an experimental RI value. The NIST\EPA\NIH EI mass spectral Library contains some spectra (not measured by NIST) that were measured by introduction of the pure compound by direct insertion probe or heated batch inlet, as well as gas chromatography. Going forward, only spectra obtained by sample introduction using GC are being added. MS Search v.2.0d, provided with the NIST 05 release, also saw the inclusion of a chemical group-additive (incremental) approach to estimating n-alkane RI values based on the chemical structure. This provided RI data when experimental data was not available. This feature not only was applied to items in the NIST\EPA\NIH EI Library but also to items in user-libraries with associated structures.
An artificial intelligence (AI) method for predicting RI has also been developed[5]. This proved to more closely match measured values than the previous method. As of NIST 23, all the compounds represented in the NIST\EPA\NIH EI Library have AIRI values. New in MS Search v.4.0 is when the search of a measured spectrum is constrained by an RI value using a semi-standard non-polar (SSNP) column and no experimental measured value is available, the AIRI value is used.
An example of a GC Method as it appears in the Text Information window of the Lib Search, Other Search, Name Search, and Librarian tab views is seen on the right (Figure 5). The text for this window can be searched using the Fine selection from the Right Mouse Button (RMB) menu displayed in the Text Information window for the selected Hit List spectrum.
Figure 5. Example of a GC Method from the Experimental GC Method RI Database.
The NIST 05 release saw a significant advancement which was the introduction of search methods of product-ion mass spectra against a library of product-ion spectra that were obtained by MS/MS of precursor ions mainly formed by positive and negative ion electrospray (ES). The NIST Tandem Library introduced with MS/MS search algorithms did include a few spectra where the precursor ion was formed by APCI. The NIST 26 Tandem Library contains >3 million integer- and accurate-m/z spectra of 576K precursor ions formed from >65K compounds. Figure 6 shows the growth-by-release of the NIST Tandem Library. Figure 1 shows the diversity of compound types in the Tandem Library
Through the NIST 23 release, both the Tandem and the NIST\EPA\NIH EI Libraries were provided in a single distribution, The Tandem Library with MS Search, MS Interpreter, and other utilities could be purchased separately. As of the NIST 26 release, this is no longer the case and the two must be individually purchased. The two can be installed so that a single copy of the NIST MS Search Program can be used for both Tandem and EI libraries.
Figure 6. Yearly Growth of the Tandem Library Since Inception
In 1995, NIST introduced a single-screen Microsoft Windows version of the MS Search Program, replacing the MS DOS version. This was replaced in 2002 with the current multiple tab Windows version. With NIST 26, a Chromatogram tab view was added. This will be discussed in more detail later in this presentation.
One of the more significant improvements to MS Search was the introduction with the 2017 release of a new Hybrid search (Hybrid-17 search) for both the EI and Tandem data. The Hybrid-17 search[6], [7] uses the precursor ion for tandem spectra (molecular ions for EI spectra or the nominal mass of the compound) to calculate neutral losses and similarities of the query spectrum with those of the spectra in the searched libraries to create a spectrum that has the same attributes as a corresponding structure associated with the model compound. A delta mass is then reported which can be translated into number of atoms of multiple elements to be added or removed from the model in order to produce the hybrid structure. This technique is used in postulating the structure of an unknown, when there is no spectrum present for that compound in the searched libraries. Its effectiveness for tandem data is illustrated in a paper by Oliver Fiehn which appeared in Analytical Chemistry in 2019.[8]
In MS Search v.4.0, this delta mass is now translated into specific elements and the number of atoms of these elements. This greatly facilitates the identification of the compound when no mass spectrum is in one of the searched libraries and provides a proposed elemental composition for the hybridized compound. The Hit List column headers are dForm (the number of atoms of the most common elements added or subtracted from that of found library spectrum. pctRelForm is the relative frequency of occurrence in high scoring searches. predForm is the resulting elemental composition of the queried spectrum.
Figure 7. Results of a Hybrid Search of an unknown EI Spectrum.
Based on the results shown in the Hit List (right side of Figure 7) a structure is drawn and associated with the spectrum. The spectrum/structure pair is then sent to MS Interpreter (Figure 8) for confirmation.
Figure 8. MS Interpreter with acquired spectrum and proposed structure with Isotope Calculator overlaid.
Soon after NIST began the measuring of spectra (both mass and gas-phase infrared) in 1990, the presence in non-mass spectral databases was included for compounds. Initially the number of non-mass spectral databases was limited to the Commercially Available Fine Chemical Index (FINE), Toxic Substances Control Act Inventory (TSCA), Registry of Toxic Effects of Chemical Substances (RTECS), Environmental Monitoring Methods Index (EPA), U. S. Pharmacopoeia/U.S.A.N. (USPS), CRC Handbook of Data of Organic Compounds (HODOC), NIH-NCI Inventory File (NIH), European Index of Commercial Chemical Substances (EINECS), and NIST/EPA Gas Phase IR Database (IR). A new chapter was introduced after the release of NIST 14. This was a new way of selecting compounds whose mass spectra would be measured for both the EI and Tandem Libraries and a new process of evaluating the newly acquired data. The number of non-mass spectral database containing compounds with spectra in both libraries grew by a significant number. The listing of these other databases is much more significant in both NIST 23 and NIST 26 than it was in previous editions.[9]
What is Considered the Most Significant Advancement to NIST 26: The Chromatogram tab in MS Search
From its inception, the NIST MS Search Program has been considered as a “single spectrum” search tool. It was possible to submit a file containing multiple spectra extracted by some process from a data file that contained sequential spectra based on the retention time in a chromatogram.
The NIST 98 (1998) release of the NIST\EPA\NIH EI Library (first fully evaluated release) saw the introduction of a program separate from MS Search, AMDIS (Automated Mass spectral Deconvolution and Identification System). AMDIS is a program developed by the Mass Spectrometry Data Center under contract to the United States Army for verification used to the Chemical Weapons Convention (CWC), which took effect 29 April 1997[10]. AMDIS was/is the primary tool used by OPCW (Organization for the Prohibition of Chemical Weapons, headquarters in The Hague, Netherlands) for GC/MS data. Gas chromatography is a dynamic process with respect to time. This means the ions in the source are continually changing (both in m/z values and abundance). This can result in the loss of chromatogram fidelity and the inability to detect low concentration components in the presence of more abundant ones. AMDIS allows for the visualization of these types of situations. AMDIS is the basis of the Chromatogram tab for processing both integer and accurate m/z value GCMS data. The EI Input file section of the Chromatogram tab will allow the selection of integer or accurate m/z files and their processing with or without a retention index constraint. It also allows for the creation of an RI calibration file needed when the RI values are a part of the processing. At the present time RI values are associated with semi-standard non-polar columns (SSNP). Identifications are performed using the NIST 26 EI Libraries and/or third‑party EI libraries such as the Wiley Registry of Mass Spectral Data or crowdsourced libraries like the one distributed by MassBank of North America (MoNA), https://mona.fiehnlab.ucdavis.edu/, at University of California at Davis (Davis CA, U.S.A.).
AMDIS continues to be provided with the distribution of the NIST\EPA\NIH EI Library and MS Search v.4.0. AMDIS is limited to processing integer m/z value GC/MS data. If accurate m/z values are submitted to AMDIS, the values are rounded to integers.
Figure 9. A display of a processed integer GC/MS data file with the Structure window for the selected Hit (highlighted in the Hit List in lower half of the display) and enlarged circle colored green in the reconstructed total ion current (RTIC) chromatogram window. The Chromatogram has a separate Structure window to the right which is closed to make the Hit List appearance clearer.
Figure 9 is a modified view of the Chromatogram tab with the results of the processing of an integer GC/MS EI data file. The normally displayed Structure window (which appears on the right of the lower part of the display) is closed, and the floating Structure window (described earlier) is displayed. Just above where the stationary Structure window would appear is window divided into two vertical halves. The left of the display shows a butterfly plot of the deconvoluted component (top) and the library spectrum for the selected Hit. There are four separate displays for this spectrum window. Optionally, the structure of the Hit List library spectrum can be displayed with the plot. To the right of this plot window is a pair of split windows. The top split window is the Text Information for the deconvoluted component. The bottom is the Text Information showing the meta data of the Hit List spectrum.
The processing of accurate m/z GC/MS data is accomplished using an encoding process which converts the accurate m/z values to integers. These integer values are used by the AMDIS algorithm, which will only work with whole numbers. When the deconvoluted components are brought into the display, they appear as accurate m/z data. When an accurate m/z value GC/MS data file is processed using the Analyze EI Input File rather than the Analyze HiRes Input File the components do not score as high.
The Chromatogram tab allows the RTIC chromatogram to be zoomed. When zoomed, the Hit List only displays components that are displayed in the RTIC chromatogram window. Deconvoluted/Identified components can be transferred to other window of MS Search for further investigation such as the Hybrid-17 search.
The Chromatogram tab uses the XIC Analyzer that was originally developed for the use of protein analysis[11] in the processing of DDA (Data-Dependent Acquisition) LC-MS/MS data. Such data files form any manufacturer’s instrument, in the mzML format are processed. If your instrument does not allow the conversion to this format, use MSConvert (https://proteowizard.sourceforge.io/download.html).
Figure 10. Comparison of a Manual Workflow for LC-MS/MS Data and that of MS Search V.4.0 9 (Courtesy of NIST Presentation)
The data file can be processed with or without an implementation of the Hybrid-17 search. These chromatographic analyses can be time consuming; therefore these processes can be run in background while other processes are being carried in the other tabs of MS Search (only one chromatographic file at a time can be processed). When complete, the user is given the option to store the results or bring them into focus in the Chromatogram tab (Figure 11).
Figure 11. Dialog box displayed with background processing of Chromatographic data file is complete.
To illustrate the complexity of a data file that can be processed by MS Search, an orange juice extract was analyzed and the data processed.
Figure 12. Results of processing orange juice extract by DDA LC-MS/MS (Courtesy of NIST Breakfast Presentation).
The two vertical axis of the RTIC chromatogram window of the Chromatogram tab represent Abundance (left) and the TIC (right). Both can be presented as linear (default), log, or square root. Colors of the circles are specific to the component type as indicated in Figure 12. In all cases, the abscissa is a time scale, usually in minutes.
MS Search v.4.0 requires a 64-bit version of Microsoft® Windows®. Previous releases of MS Search were of either the NIST EI and Tandem Library with MS Search, MS Interpreter, and AMDIS or just the Tandem Library with MS Search and MS Interpreter. Beginning with NIST 26, the EI and Tandem Libraries are available only as individual distributions. The two can be installed on top of one another and used from the same platform on a single computer.
NIST MS libraries are available through authorized distributors. A list of these distributors can be found at: https://chemdata.nist.gov/dokuwiki/doku.php?id=chemdata:distributors. NIST offers three different products to end users: The EI Library with software, the Tandem (MS2) Library with software, and the GC Method/Retention Index with software, which is different from MS Search (RI). Not all distributors provide all three products. Three columns on the left side of the Authorized Distributors List show what products are offered by the individual distributor. Previous users of any prior NIST EI Library have upgrade pricing available. There is no upgrade pricing for the Tandem Library, even if obtained with an earlier version of EI Library. Each distributor sets their own price; therefore, it may be worth looking at several different distributors’ websites. Different distributors have different ways of distributing the NIST products. Some allow for direct secure FTP downloads.
NIST offers a free download of a Demo versions of MS Search v.4.0 if either the Tandem or EI library at https://chemdata.nist.gov/dokuwiki/doku.php?id=start. The EI version is provided with sample spectra, MS Interpreter, and AMDIS as well as companion GC Method/RI data for provided EI compound. The Tandem demo is provided with sample spectra and MS Interpreter. No Demo is available for the GC Method/RI Data and its software which allows for the searching of a range of RI value (not part of MS Search).
It is important to remember the single-computer/one user at a time restriction imposed by the NIST license. The NIST mass spectral libraries are copyrighted by the United States Department of Commerce.
An extremely valuable resource for MS Search v.4.0 and previous releases of the NIST EI and Tandem data has been developed by James Little (https://mzinterpretation.com/). This resource contains links to instructional videos and associated documents to aid in getting started with this very powerful program. It also includes valuable downloads.
The author (O. David Sparkman) declares that he is a paid contractor for the Mass Spectrometry Data Center of the National Institute of Standards and Technology, United Staes Department of Commerce responsible for end-user and distributor support, providing evaluation of the data and software, and preparation of certain written materials pertaining to the products described in this presentation.
References
[1] Stephen R. Heller “The History of the NIST\EPA\NIH Mass Spectral Data Base” Today’s Chemist at Work, 1999, 8(2), 45-46. 49-50. Copyright © 1999 by the American Chemical Society)
[2] Stephen R. Heller “Computer Software Reviews: NBS Mass Spectral Data. PC Version 1.02 (Database 1-A), J. Am. Chem. Soc., Vol. 110, No. 10, 1988, 3336 – 3337.
[3] Nominal mass is the mass of an elemental composition calculated based on the number of atoms of each element and the integer mass of the element’s most abundant isotope.
[4] Dillard, J. G.; Heller, S. R.; McLafferty, F. W.; Milne, G. W. A.; Venkataraghavan, R. “Critical evaluation of class II and class III electron impact mass spectra. Operating parameters and reporting mass spectra” Org. Mass Spectrometry. 1981, 16, 48-49.
[5] Geer, L. Y., Stein, S. E., Mallard, G., & Slotta, D. “AIRI: Predicting Retention Indices and Their Uncertainties Using Artificial Intelligence” J. Chem. Inf. Model, 2024, 64(3), 690−696 DOI: 10.1021/acs.jcim.3c01758 (Published online: January 17, 2024)
[6] Arun S. Moorthy, William E. Wallace, Anthony J. Kearsley, Dmitrii V. Tchekhovskoi, and Stephen E. Stein “Combining Fragment-Ion and Neutral-Loss Matching during Mass Spectral Library Searching: A New General Purpose Algorithm Applicable to Illicit Drug Identification” Anal. Chem. 2017, 89, 13261−13268
[7] Meghan C. Burke, Yuri A. Mirokhin, Dmitrii V. Tchekhovskoi, Sanford P. Markey, Jenny Heidbrink Thompson, Christopher Larkin, and Stephen E. Stein “The Hybrid Search: A Mass Spectral Library Search Method for Discovery of Modifications in Proteomics” J. Proteome Res. 2017, 16, 1924−1935.
[8] Ivana Blaženović, Tobias Kind, Michael R. Sa, Jian Ji, Arpana Vaniya, Benjamin Wancewicz, Bryan S. Roberts, Hrvoje Torbašinović, Tack Lee, Sajjan S. Mehta, Megan R. Showalter, Hosook Song, Jessica Kwok, Dieter Jahn, Jayoung Kim, and Oliver Fiehn “Structure Annotation of All Mass Spectra in Untargeted Metabolomics” Anal Chem 2019 91 (3), 2155-2162.
[9] O. David Sparkman “NIST 23: The Largest Increases in Compound Coverage for the Tandem and NIST/EPA/NIH EI Libraries Since NIST Became Curator” Separation Science Primer Learning for Analytical Scientists, Aug 3, 2023, https://www.sepscience.com/nist-23/.
[10] History of the Chemical Weapons Convention, https://www.opcw.org/about-us/history
[11] Guanghui Wang, Zheng Zhang, Yi Liu, Meghan C. Burke, Sergey L. Sheetlin, and Stephen E. Stein, “An XIC-Centric Strategy for Improved Identification and Quantification in Proteomic Data Analyses” J. Proteome Res. 23 (5), 1571-1582.
