Nontargeted chemical analysis reveals significant blind spots: study

A computational framework has shown that widely used environmental screening methods capture only a minute fraction of chemical space with implications for public health and lifetime chemical exposure

A recent study has revealed that so-called nontargeted chemical analysis captures only a small fraction of the chemicals present in the environment, despite its broad remit. Researchers at the University of Amsterdam’s Van ’t Hoff Institute for Molecular Sciences, Netherlands, have shown that methodological constraints impose substantial limitations on what instruments can detect, leaving large areas of chemical space effectively invisible.

Nontargeted analysis aims to screen environmental samples without prior selection of compounds, with the intention to map the full range of chemicals that may be present. However, the Amsterdam team has demonstrated that this ambition does not align with practical reality. Their findings have indicated that widely used analytical workflows fail to access the majority of chemical diversity, which introduces systematic gaps in environmental monitoring data.

To quantify these limitations, the researchers have developed a computational approach termed Measurable Feature Prediction. This framework has enabled the team to estimate, in advance of experimental analysis, which regions of chemical space a given method can detect. The work focused on liquid chromatography–electrospray ionisation–high-resolution mass spectrometry, a technique widely regarded as a gold standard for environmental screening.

The study has shown that inherent physical and chemical constraints, including ionisation efficiency and chromatographic retention, restrict the number of detectable compounds. These constraints create what researchers describe as ‘blind spots’ – that is, regions of chemical space that remain inaccessible to the method. In this context, ‘chemical space’ refers to the theoretical universe of all possible chemical compounds, a vast domain that far exceeds the reach of any single analytical technique.

“The numbers were much smaller than we expected”, said Dr. Saer Samanipour, who leads the Environmental Modelling and Computational Mass Spectrometry group.

“This may sound like a lot but compared to the vast chemical space it is about 0.01%, which is a minute amount,” he said.

The research has centred on analysis of internal standards within liquid chromatography–electrospray ionisation–high-resolution mass spectrometry workflows.

Dr. Lapo Renai, a postdoctoral researcher supported by the European Union’s Marie Skłodowska-Curie Actions and the University of Amsterdam Data Science Centre, has developed a similarity-based modelling strategy. This approach has combined molecular fingerprinting with predicted retention indices and ionisation efficiencies to estimate method-specific chemical coverage.

The Measurable Feature Prediction framework has therefore provided a means to determine which compounds are likely to be observable before any real-world sample undergoes analysis. For liquid chromatography – electrospray ionisation – high-resolution mass spectrometry, the results have suggested that fewer than a few thousand compounds can be measured in a single run. When placed against the immense scale of chemical space, this figure represents only a negligible proportion.

Samanipour argued that the findings underline a need to adopt complementary analytical strategies.

“We also need to map the blind spots of each method, as those are the real human and environmental health issues of the future”, he said. The reference to ‘orthogonal approaches’ denotes the use of multiple analytical techniques with differing detection principles, which together can expand overall chemical coverage.

Renai emphasised that the concept of comprehensive nontargeted analysis requires reconsideration.

“Chemical-space-aware frameworks such as we present in the paper can help guide smarter method development and reduce method-specific measurability uncertainty in exposomics and environmental screening”, he said.

The study has highlighted a fundamental limitation in current environmental analytics. While nontargeted workflows have expanded detection capabilities, they do not provide a complete picture. Instead, they operate within defined boundaries shaped by instrument physics and chemistry. The authors have concluded that recognition and systematic mapping of these boundaries will prove essential to improve analytical strategies and to address emerging risks in environmental and human health.

For further reading please visit: 10.1021/acs.analchem.5c07705