PNNL

The Science

Researchers and investigators use non-targeted analysis (NTA) to identify chemicals present in an unknown sample. These chemical components can be known or yet-to-be-known compounds, all of which makes up “chemical space.” NTA is widely used in threat detection, forensic investigations, and studies of biological systems.

Unfortunately, no one NTA method can fully map (detect) all of chemical space, so researchers are routinely creating and refining NTA methods. And because researchers use varying methods and instrumentation to conduct NTA studies, there is a great need to define and communicate the subsets of chemical space that are detectable by a method.

In response to this challenge, a team of researchers, including data scientists from Pacific Northwest National Laboratory (PNNL), proposed the ChemSpace Tool framework to advance reporting of NTA chemical identifications through a more standardized, transferable, and accurate process. The envisioned tool will enable any laboratory or practitioner to apply a standardized and systematic workflow to easily identify and communicate the detectable chemical space applicable to their study.

The Impact

The framework proposed by this work kicks off development of a first-of-its-kind chemical space amenability tool, or ChemSpace Tool, to better define the detectable chemical space for various NTA approaches. As the ChemSpace Tool is developed, it will improve NTA accuracy, streamline method development, improve reproducibility of experiments or analysis, and enhance method transferability between laboratories and practitioners.

Summary

The framework proposed here will make it possible for multiple data-driven chemometric filtering steps to determine which compounds or components are likely detectable in the chemical space analyzed.

The ChemSpace Tool, when fully developed, is intended to divide chemical space into three subsets: the detectable space, the identifiable space, and the region that includes compounds that are not detectable or identifiable. By understanding chemical space coverage and the differences between detectable, undetectable, and unidentified space, the ChemSpace Tool will improve the accuracy of NTA, both prospectively and retrospectively.

Prospectively, the tool will be able to organize and present screening libraries and set identification limits before the NTA process has started. Retrospectively, this advancement will be able to identify false detections to further confidence in NTA workflows. Because the ChemSpace Tool workflow better defines the chemical space of a method, it offers a distinct opportunity to increase performance assessment, compare results on an inter-laboratory or inter-project basis, and adapt existing methods to new projects with more confidence.

PNNL Contacts

Jessica Bade
Pacific Northwest National Laboratory
jessica.bade@pnnl.gov, (509) 371-7843

Funding

This research is partially supported by PNNL’s m/q internal investment in mass spectrometry for molecular characterization and quantification.