ment. “It links them together so whenever you do a mass spectral
library search, you get more information that allows you to
reliably identify the compound.”
The search function also informs which peaks are shifted in
the mass spectrum, and which are not—alerting researchers as to
where a modification is located in a specific molecule. This can
help define the structure, as well.
But Stein wouldn’t recommend using the function without
an analyst’s judgement on its reliability. After all, libraries don’t
identify compounds—people do.
“When you get a hit list, it always needs a manual confirma-
tion,” Stein said. “Compounds’ spectra are too variable, they
can be too similar, or there could be an analytical problem. In
all cases, a human needs to look at a couple hits in the library
and make a decision based on that.”
Humans are the ones that fuel NIST’s rigorous quality
control measures for the Mass Spectral Library. According to
Stein, his team first tries to obtain the highest-quality original
spectrum “because there’s no substitute for that.” Every spec-
trum is then critically evaluated by at least two people. If they
agree, the spectrum gets added to the library. If they disagree,
the spectrum is either resolved with a third-party, or is com-
pletely left out of the library if a high level of confidence cannot
be reached. The last stop for a spectrum includes running it
through various software methods that match consistency with
structure or chemical formula. If a spectrum passes all these
tests, it’s added to the archives.
“We just want the compounds to give [scientists] good, confident identification when they find it in the lab,” Stein said.
More tools for forensic analysts
The NIST spectral library is used by scientists in virtually every industry—clinical, environmental, food, fragrance,
energy, etc. As Stein points out though, forensic scientists have
another tool at their disposal—one he believes is underuti-lized, especially in relation to the opioid crisis.
AMDIS, as it’s called, was developed at NIST with the
support of the DoD to reduce the effort involved in identifying compounds by gas chromatography mass spectrometry
(GCMS), which is often the method of choice for forensic
scientists given its reproducibility factor.
The program extracts the spectrum of each component
in a mixture analyzed by GCMS or liquid chromatography
mass spectrometry (LCMS), and identifies target compounds.
AMDIS is an integrated set of procedures for first extracting pure component spectra and related information from
complex chromatograms, then using this information to
determine whether the component can be identified as one of
the compounds represented in a reference library. The library
could be your own, it could be NIST’s, or it could be one of a
Be it targeted or untargeted analysis, Stein leans toward the
use of a large, comprehensive library.
“One interesting thing about a big library is all the compounds that could masquerade as false positives are probably
in a big library. So, you can show the uniqueness of your spectrum only by searching against a large library,” he said.
CATALYST FOR SUCCESS
Although the library upgrade is still new, NIST postdoc and
mathematical statistician Arun Moorthy is already working
on improvements to the search algorithm. He told
Laboratory Equipment he is busy deciphering a way to automatically
estimate the molecular weight of compounds by conducting a
hybrid search at multiple masses.
When the right mass is hit, the score will go up—indicating
a reliable match to the researcher. It’s an automation function
that will negate one of the only drawbacks of using the hybrid
search function—having to guess the molecular weight.
After work on the molecular weight estimation is completed,
Stein said they will turn their attention to the hit list.
“We want to improve the distance between the correct hit score
and the highest incorrect hit,” Stein said. Doing so will allow researchers to be even more confident in their mass spectrum analyses.
NIST has been publishing its Mass Spectral Library since
1989, with the latest update coming last year. At ASMS this
year, NIST will be giving three poster presentations on the
application of the hybrid search to various research areas,
including metabolomics, proteomics and forensics, as well as a
fourth presentation detailing mass estimation.