Searching Low-Resolution Spectra and Low-Resolution Libraries in NIST26

Importance of the topic

Low-resolution tandem mass spectra and corresponding low-resolution libraries are widely available across routine analytical laboratories. Their effective use extends MS/MS identification capability beyond high-resolution instruments (QTOF, Orbitrap) to single-quadrupole systems with in-source CID and triple-quadrupole collision-cell instruments. Recognising how to search, interpret, and import low-resolution spectra into the NIST26 environment therefore increases the practical reach of spectral library identification workflows and enables many labs to leverage large reference collections without requiring accurate-mass instrumentation.

Objectives and overview of the presentation

The material reviews practical guidance for searching low-resolution MS/MS spectra in the NIST26 release and related public libraries, outlines trade-offs between accurate-mass and nominal-mass searches, describes library selection and search-mode choices, documents import workflows, and highlights software behaviours (including bugs) that affect results. It also explains scoring metrics used by NIST search tools and suggests best practices for routine identification when working with low-resolution data.

Methodology and instrumentation

• Scope of libraries: NIST26 integrates both high-resolution (accurate-mass) MS/MS entries and substantial numbers of experimental low-resolution LC-MS/MS spectra drawn from public sources (estimated tens of thousands in MoNA and MS-DIAL) and older commercial/user libraries in NIST format.
• Applicable instruments: single-quadrupole LC-MS with in-source CID, triple-quadrupole instruments producing product-ion spectra, and high-resolution platforms (QTOF, Orbitrap) — all can benefit from NIST26 searches when spectra are represented in the libraries.
• Common vendor integration: Waters, Agilent (MassHunter/OpenLab), and Thermo commonly include a “Send NIST Lib” export pathway for direct transfer into NIST’s Librarian/Spec List. Sciex and Bruker may require ASCII (txt or msp) exports and manual import.
• Import settings: for low-resolution spectra use the EI (normal m/z) option to retain nominal m/z values; for accurate-mass spectra select Accurate m/z and appropriate decimal places. Be aware of known software bugs that may force nominal mass truncation in certain import or threshold settings.

Main results and discussion

• Utility of low-resolution MS/MS libraries: The NIST26 MS/MS libraries are valuable for identification even without accurate-mass instrumentation. If an unknown’s fragmentation pattern is present in the library, identification can succeed regardless of whether the library spectrum was acquired at high or low resolution.
• Search-mode effects: EI-type searches (Identity EI Normal, Similarity → Simple EI) include low-resolution spectra in results when the search mode treats the query as nominal-mass/EI-like. Identity MS/MS and HiRes NoPrecursor restrict results to accurate-mass spectra unless the special NIST low-resolution library (lr_msms_nist) is included. Selecting Spectrum Type = Any simplifies processing by allowing mixed-spectrum searches.
• Search-mode guidance: Identity EI Normal places greater emphasis on higher m/z product ions (often structurally diagnostic) and is useful when higher-mass fragments are expected to be informative. Similarity → Simple EI performs a more balanced peak-by-peak comparison and is a good default when higher-mass bias is not wanted. MS/MS EI Similarity serves specifically to compare MS/MS spectra against EI spectra for cross-technique resemblance checks.
• Library composition considerations: Public libraries like MoNA and MS-DIAL contain many accurate-mass entries but also tens of thousands of low-resolution experimental spectra. Additionally, user and legacy commercial libraries in NIST format may lack precursor information, affecting search options and the interpretability of results.
• Scoring interpretation: NIST provides a composite Score plus directional dot-product metrics: Dot (bidirectional dot product measures overall similarity), Rev-Dot (reverse dot emphasizes presence of library peaks in the query), and PSS-Dot (Partial Spectrum Search dot product useful when only part of the library spectrum appears in query). Practical interpretation: Dot is used for initial ranking; high Rev-Dot relative to Dot suggests additional peaks in the query (co-fragmentation/contamination); high PSS-Dot relative to Dot indicates an incomplete or low-abundance query spectrum that still matches a portion of the library entry.
• Practical search settings: keep Best Matching Only toggled off initially to avoid inadvertent removal of useful duplicate records; add columns such as molecular weight (MW), PSS Match, and Rev-Dot to result displays to aid review; toggling filtering and display options can help reduce candidate sets (e.g., InChIKey-based de-duplication).

Benefits and practical applications of the method

• Broader laboratory applicability: enables confident identifications from low-resolution LC-MS platforms common in many routine labs without investing in high-resolution systems.
• Flexible workflows: supports both in-source CID and collision-cell product-ion spectra; combining low- and high-resolution libraries increases chances of finding matching fragmentation patterns.
• Operational efficiency: familiar search strategies (EI Normal, Similarity) and judicious use of scoring columns provide robust triage of candidate matches for follow-up confirmation steps (retention time, standards).
• Legacy-data utilization: incorporates older user-contributed and commercial low-resolution spectra so historical data remain useful for modern identification tasks.

Instrumentation used

• Nominal/low-resolution platforms: single-quadrupole LC-MS (in-source CID), triple-quadrupole instruments (product-ion spectra, collision cell).
• High-resolution platforms for comparison/reference: QTOF, Orbitrap (accurate-mass MS/MS entries present in libraries).
• Software and vendor integrations: NIST Librarian/Library Search in NIST26; vendor export features (“Send NIST Lib”) available in Waters, Agilent (MassHunter/OpenLab), Thermo; Sciex and Bruker may require manual export/import via ASCII formats.

Future trends and opportunities for application

• Library growth and harmonisation: continued accumulation of experimentally acquired low-resolution spectra in public repositories (MoNA, MS-DIAL) and improved curation will increase identification coverage for nominal-mass searches.
• Software improvements: resolving import/threshold bugs, richer metadata handling (collision energy, ionization mode), and better indexing of low-resolution libraries (as exemplified by lr_msms_nist) will improve search accuracy and interoperability.
• Hybrid and AI-driven search approaches: machine-learning classifiers and hybrid scoring that combine spectral pattern matching with retention time prediction or in silico fragmentation will raise confidence for low-resolution identifications.
• Cloud and federated libraries: centralised searchable collections with standardised metadata and provenance will make cross-instrument searches faster and more reproducible.
• Expanded use cases: routine QA/QC screening, environmental monitoring, forensic screening, and metabolomics labs with limited resources can apply low-resolution MS/MS library search workflows to broaden identification capacity.

Conclusion

NIST26’s inclusion of both high- and low-resolution MS/MS content makes it a practical resource for a wide spectrum of laboratories. Even without accurate-mass instrumentation, identifiable fragmentation patterns in the library frequently provide sufficient structural information for useful identifications, especially when appropriate search modes, scoring interpretation, and import settings are used. Awareness of software behaviours and careful selection of library/search options (Spectrum Type = Any, correct EI/Accurate-m/z import settings, and inspection of Dot/Rev-Dot/PSS metrics) are essential to maximise reliability when working with low-resolution data.