Only through machine learning and AI can we come to grips with all chemicals around us – Technology Org

Only through machine learning and AI can we come to grips with all chemicals around us – Technology Org


The Open-Access Journal of the American Chemical Society (JACS AU) has published an invited perspective by Dr. Sawyer Samanipour and his team on the daunting challenge of mapping all the chemicals around us.

Samanipour, assistant professor at the Van ‘t Hoff Institute for Molecular Sciences at the University of Amsterdam (UVA), takes inventory of the available science and concludes that a true active chemical management is not currently possible. To truly get a grip on the vast and expanding chemical universe, Samanipur advocates the use of machine learning and AI to complement existing strategies to detect and identify all the molecules we come in contact with.

Only through machine learning and AI can we come to grips with all chemicals around us – Technology Org

Description of the existing problem. Out of the vast amount of molecules in chemical space, current technology can detect only a limited number. The fraction of molecules actually identified is even smaller. The exposome chemical space – the molecules we are exposed to – extends far beyond the scope of these measurable, measured and identified molecules. Image credit: HIMS/JACS.

In the language of science, the set of all the molecules we come in contact with is called the ‘exposome chemical space’ and it is the center of Samanipur’s scientific efforts. It is their mission to explore this vast molecular space and map it to the most ‘remote’ corners. He is driven by curiosity, but even more by necessity. Direct and indirect exposure to numerous chemicals, most of which are unknown, pose a significant threat to human health. For example, it is estimated that 16% of global premature deaths are linked to pollution. There is also damage to the environment, which can be seen, for example, in the loss of biodiversity. According to Samanipur, a case can be made that mankind has exceeded the safe operating space for introducing man-made chemicals into planet Earth’s system.

The current approach is inherently passive

“It’s quite disconcerting that we know so little about it,” he says. “We know so little about the chemicals already in use, let alone how we can adapt to new chemicals.” which are currently being created at an unprecedented rate.” In a previous study, they estimated that less than 2% of all the chemicals we come into contact with have been identified.

“The way society looks at this issue is inherently passive and reactive at best. It is only when we see certain types of effects of exposure to chemicals that we feel the need to analyze them. We try to determine their presence, their effects on the environment and human health, and we try to determine the mechanisms by which they cause any harm. This has created many problems, the most recent of which is the crisis of PFAS chemicals. But we have also seen major issues with flame retardants, PCBs, CFCs, etc.

Furthermore, regulatory measures are mainly targeted at chemicals with very specific molecular structures that are produced in large quantities. “There are countless numbers of other chemicals out there that we don’t know much about. And these are not just man-made; Nature also produces chemicals that can harm us. Through purely natural synthetic routes, or through the transformation of man-made chemicals. The latter category in particular has been systematically ignored, according to Samanipur. “Traditional methods have cataloged only a fraction of the exosomes, overlooked transformation products and often given inconclusive results.”

We need a data-driven approach

The JACS AU paper provides an in-depth review of the latest efforts in mapping the exposome chemical space and discusses their results. A main limitation is that conventional chemical analysis is biased towards known or proposed structures, as this is key to interpreting data obtained from analytical methods such as chromatography and mass spectrometry (GC/LC-HRMS). Thus the more ‘unexpected’ chemicals are ignored. This bias is avoided by so-called non-targeted analysis (NTA), but still the results are limited. Over the past 5 years, 1600 chemicals have been identified, while approximately 700 new chemicals are introduced into the US market every year. Samanipur: “When you take into account the potential transformation products of these novel chemicals, you have to conclude that the pace of NTA studies is too slow to catch up. At this rate, our chemical exposome will remain unknown.”

The paper lists these and many other obstacles in current analytical science and suggests ways to improve results. Samanipur argues that the use of machine learning and artificial intelligence in particular will really advance the field. “We need a data-driven approach across multiple lines. First, we must intensify datamining efforts to obtain information from existing chemical databases. Already documented relationships between the structure, exposure and effects of the identified chemicals will lead us to new insights. For example, they can help predict the health effects of related chemicals that are not yet known. Second, we have to perform retrospective analysis on already available analytical data obtained by established methods, expanding the identified chemical space. We will certainly find molecules there that have been overlooked so far. And third, we can use AI to understand the structure and scope of the exposome chemical space.

work hard to deal with it

Of course, this is all a very complex, challenging matter, Samanipur realizes. But as a kind of astronaut in molecular space – like the explorers of the factual universe – he will not let that complexity elude him. “We have to work hard to deal with this. I have no illusions that during my scientific career we will be able to completely chart the exposome chemical space. But it is important that we face its complexity, discuss it and take the first steps towards dealing with it.

Source: University of Amsterdam