Manual and Automated Hybrid Searches in NIST26

Beyond Exact Matches: Manual and Automated Hybrid Searches in NIST26

Conventional MS/MS library searching works best when the spectrum of an unknown compound closely matches a reference spectrum already stored in the library. But real samples frequently contain compounds that are absent from available libraries. In his presentation Manual and Automated Hybrid Searches in NIST26, James Little of Mass Spec Interpretation Services demonstrates how the hybrid-search capabilities in NIST26 can extend effective library coverage by finding structurally related compounds and using their spectral relationships to guide the identification of unknowns.

Extending MS/MS library searching beyond exact matches

A conventional library search attempts to find spectra that closely resemble the measured spectrum. This approach can provide highly confident identifications when an appropriate reference spectrum is available, but its usefulness decreases when the unknown compound itself is missing from the library.

The Hybrid MS/MS Search in the NIST Mass Spectral Search Program takes a different approach. Rather than requiring an exact spectral match, it can recognize spectra of structurally related compounds and account for systematic mass differences between the unknown and library candidates. This makes it possible to use existing reference spectra to investigate compounds that have not themselves been measured and deposited in the library.

In NIST26, hybrid searching can be performed in two ways:

• manually by sending individual spectra to the Library Search window,
• automatically as part of the Chromatogram workflow.

The two approaches complement one another: automated searching can rapidly screen many chromatographic components, while manual searching provides more opportunities to examine individual candidates and investigate their structures in greater detail.

Using mass differences to recognize structural relationships

An important element of the hybrid search is the interpretation of the mass difference between an unknown spectrum and a related library spectrum.

NIST26 reports this difference together with Δform (dForm), which represents a plausible elemental composition associated with the observed mass shift. Instead of treating a different precursor mass as evidence that two compounds are unrelated, the software can therefore evaluate whether the difference corresponds to a chemically reasonable structural modification.

The presentation also highlights pctRefForm, a statistical measure related to how frequently a particular dForm occurs relative to commonly observed modifications. Higher values indicate more frequently observed and therefore generally more plausible formula differences, whereas extremely rare differences may be accidental and should be interpreted cautiously.

This additional information helps transform a library hit from a simple similarity result into a starting point for structural interpretation.

A manual workflow for investigating unknown chromatographic components

James Little demonstrates a practical workflow in which a standard chromatogram search is first performed and components without a confident identification are selected for further investigation.

The unidentified components can be sorted by relative abundance, allowing the analyst to prioritize the most significant signals. Individual spectra are then reviewed, including through butterfly plots, and promising components can be manually transferred to the Library Search window for a hybrid search.

This approach is particularly useful when an automated conventional search has failed to provide a satisfactory identification but the spectrum still contains enough structural information to point toward a related compound class.

Identifying a PEG-related unknown

The presentation illustrates the process using an unknown associated with polyethylene glycol (PEG)-type materials. The spectrum suggested an ammonium adduct, (M+NH₄)⁺, followed by the loss of ammonia.

Hybrid searching produced candidates related to PEG ethers, with the most reasonable interpretation being a dodecyl ether of polyethylene glycol. The proposed library candidate was then examined in the NIST MS Interpreter to understand its fragmentation behavior.

A particularly informative result came from the observed mass difference. The dForm indicated a loss corresponding to three ethylene glycol repeat units, with an observed value of 132.079 matching the theoretical value of 132.079. This agreement provided additional evidence that the unknown and the library candidate belonged to the same structural family but differed in PEG chain length.

The candidate structure could then be modified by removing three ethylene glycol repeat units and transferred back to the MS Interpreter for further evaluation. The example shows how a hybrid search can move the analyst beyond simply asking “What compound is in the library?” toward the more useful question “How might my unknown differ from the closest known structure?”

Why ammonium adducts require special attention

Adduct chemistry can complicate MS/MS library searching. In the example presented, the experimental precursor was observed as (M+NH₄)⁺, whereas the NIST MS/MS Library contains more spectra acquired from (M+H)⁺ precursors.

The presentation notes that NIST does not intentionally add ammonium when acquiring reference spectra. When an analyst observes a loss of 17 Da, corresponding to NH₃, it can therefore be useful to recognize that the underlying fragmentation may effectively correspond to the protonated molecule.

For such cases, Little recommends replacing the precursor value used for searching with the corresponding (M+H)⁺ value rather than relying solely on the precursor displayed in the experimental spectrum.

In the PEG example, doing so still returned two monodecyl ether candidates as the leading hits. Importantly, the presentation notes that the major MS/MS fragments generated from the ammonium and protonated precursor forms were essentially the same, supporting comparison with the protonated reference spectra.

Automating hybrid searches across a chromatogram

Manual inspection is valuable when investigating a small number of important unknowns, but chromatographic datasets may contain hundreds of components. NIST26 therefore also supports automated hybrid searching within the Chromatogram workflow.

The analyst selects the input file, enables Hybrid Search, and can choose to run the process in the background. Because hybrid searching performs substantially more processing than a standard search, the presentation notes that it can require significantly more time. NIST26 displays the background-task status and alerts the user when processing is complete.

The resulting components can again be filtered to focus on unidentified compounds and sorted according to abundance. In the example shown, the automated search returned the same PEG-related candidate found during the manual investigation, demonstrating consistency between the two approaches.

Automated hybrid searching can therefore act as an efficient first-pass screening tool, while selected results can subsequently be transferred to the Library Search window for closer examination of candidate structures and fragmentation.

From library matching to structural investigation

Hybrid searching does not eliminate the need for expert interpretation. Instead, it expands the information available when an exact reference spectrum is missing.

By combining spectral similarity with precursor-mass differences, plausible dForm values, fragmentation interpretation, and structural modification, NIST26 provides a workflow for moving from an unidentified chromatographic feature toward a chemically meaningful candidate.

The presentation demonstrates that the strongest workflow is often a combination of automation and manual interpretation: automated searching identifies promising relationships across large datasets, while targeted manual analysis helps determine whether those relationships make chemical sense. In this way, hybrid searching extends the practical value of spectral libraries beyond the compounds they contain directly and makes them more useful for the investigation of previously uncharacterized analytes.