Bring Out Your Data

David Balshaw (Program Director at the Center for Risk and Integrated Sciences, National Institute of Environmental Health Sciences) would rather “not define exposome at all” because of ongoing questions as to what may or may not be included (from pre-birth to death), so I too shall refrain from defining it. At one point, the routine ‘food’ we intake is also assigned the status of ‘exposome’ – and rightly so, in my opinion; developmental biologists would see the gut as a hollow tube outside the biological organ system. Given the enormous amounts of exposomics information being generated, it is prime time that community standards – otherwise known as MIAEE (Minimum Information About an Exposomics Experiment) – are introduced and adhered to.

At the American Society of Mass Spectrometry (ASMS) 2016 conference, an oral session entitled “Exposomics: Targeted, Untargeted and Bioinformatics Technologies” (chaired by Dr. Balshaw himself) and the following “Exposomics Interest Group meeting” (coordinated by Professor H. M. “Skip” Kingston, Duquesne University, and Dr. Anthony Macherone, Agilent Technologies and Johns Hopkins School of Medicine) left the audience with more questions outstanding than answers. Most of them are valid and need to be addressed.

Who would lay the foundations for MIAEE? The metabolomics society or the exposomics community? Should there be a formal MIAEE, in the first place? Petabytes of mass-spectrometry (MS) and NMR-based exposomics datasets are being generated and published every year. But are they being archived properly? The answer to the latter question is no, if we look to popular metabolomics community adopted archives, such as MetaboLights (www.ebi.ac.uk/metabolights/index) or Metabolomics Workbench (www.metabolomicsworkbench.org). Surely, authors could at least take the initiative and archive datasets on their webpages or institutional repositories for community access!

Is it mandatory to go for standard compound-based authentication and validation of detected and quantified metabolites? How is it possible to obtain the plethora of chemical constituents reported in single studies where the number of compounds detected can be a few thousand? How do we know what is noise (from instrument and handling) and what is from environmental/biological systems; for example, does blank air serve as the correct negative control in studies where disease biomarkers are probed in smokers? What about the platform-dependency of the exposome or for that matter geography, socio-economy-diet-genotype-race-gender affected differences in exposomes? How do we account for inter-individual differences in exposome data? Where are the databases for spectral comparisons for exposome studies, especially given that the list is ever-expanding in terms of cosmetics, antibiotics, pesticides, industrial chemicals, petrochemicals, adulterants, heavy metals, phthalates, and unknowns? Existing databases, such as NIST and HMDB, are like drops in the ocean by comparison.

What open source and commercial tools are in place to allow visualization of this peta-byte scale data on a human body or ecosystem? How can we connect exposomes to cell- or tissue-specificity or to the genotype of individuals? How can we make sure that exposure-wide association studies [EWAS] and the resulting biomarkers discovered stand the test of time?

Finally, do we need to set different workflows for metabolomics and exposomics workflows? If so, on what basis and, if not, are they interchangeable? The Metabolomics Standards Initiative (MSI) does not yet specifically mention ‘exposome’ – and journals do not ask for data archiving in open-access and public repositories. Exposome studies have a history of political ramifications and generate huge commercial interest, as well as playing an enormous complementary role in translational and clinical medicine. Is now not the right time to start answering some of these important but evidently ethereal questions?