Most human diseases arise from a combination of genetic and environmental influences. Genes contribute 30 to 70 percent of the risk for developing chronic diseases, such as type 1 diabetes, cardiovascular diseases, or Parkinson’s disease; the remaining risk likely comes from external, non-biological factors. The World Health Organization has estimated that the etiology of approximately 24% of human diseases worldwide can be attributed to potentially avoidable environmental factors. However, when it comes to understanding how these multiple exposures interact over time and measuring their synergistic effects, numerous methodological challenges arise, highlighting the need to move beyond the traditional “one exposure – one outcome” approach. Thus, this complex and poorly understood mix of environmental variables, unique to each person, makes up what is known as our individual exposome.
The concept of the exposome was introduced by Wild in 2005 to refer to “the totality of environmental (non-genetic) exposures to which an individual is subjected from conception onwards.”
Generally, the exposome is divided into three domains: the general external environment, which includes factors such as climate and urbanization, measurable at the population level; the specific external environment, including variables such as diet or smoking habits, measurable at the individual level through questionnaires, environmental sensors, or personal dosimeters; and the internal environment, which includes hormonal, inflammatory, and molecular processes, measured collectively through high-throughput “omics” techniques designed to globally describe the set of molecules in a biological system (e.g., proteome, transcriptome, metabolome).
The exposome approach opens new possibilities for the study and prevention of numerous diseases; unlike the traditional approach to environmental epidemiology, which focuses almost exclusively on exposures coming from the external environment (e.g. pollutants, chemical and physical agents), the exposome approach aims to capture the totality of exposures, including the components linked to the internal environment (e.g. specific markers of exposure such as genetic, epigenetic, metabolic an hormonal response). Characterizing and studying the internal environment is of fundamental importance and allows for the detection of previously unknown associations, improves the understanding of disease etiology, and identifies new biomarkers of exposure. Furthermore, although most environmental epidemiology studies focus on testing a well-defined hypothesis about the relationship between a single environmental factor and a health outcome, only a few diseases develop as a result of a single exposure. Thus, the characterization of the exposome seeks to overcome this limitation by integrating distal risk factors (external – general and specific – environment) with proximal risk factors (internal environment) to study their potential interactions and evaluate potential causal mechanisms in relation to various health outcomes.
However, due to the vast range of environmental exposures and the intricate ways they interact, scientists will increasingly rely on AI to unravel and understand the human exposome. Indeed, AI may be the key to understanding the exposome, having the ability to process and analyze diverse types of data simultaneously — such as genomic profiles, dietary and lifestyle habits, living environments, and biological samples.
As recently discussed by Chirag Patel, associate professor of biomedical informatics in the Blavatnik Institute at Harvard Medical School, in the future, AI tools could help healthcare professionals personalize disease risk predictions based on an individual’s unique exposome and genome. This could result in tailored interventions, including more frequent monitoring or earlier treatments. By leveraging AI, physicians would be able to make risk predictions based on a patient’s unique lifetime exposures. With AI analyzing both the exposome and genome, it could forecast long-term risks and suggest more proactive monitoring or earlier intervention when needed. Again, AI could uncover clues about a patient’s exposure that are hidden in existing data and often overlooked. For instance, if a patient arrives at the ER with certain symptoms, the personnel can input his/her medical, social and family history, blood tests and imaging results into an AI model which would then generate a report highlighting possible exposures that clinicians may not easily detect.
It’s important to recognize that AI models are still in development; researchers must continue to refine algorithms, resolve issues, and conduct thorough testing to ensure the accuracy of data analysis and predictions.
Furthermore, as recently highlighted once again by Chirag Patel, the use of AI to investigate the exposome could introduce new challenges and require new approaches: clinicians, healthcare professionals and policymakers will have to understand how to apply these insights to the prevention of diseases, since current prevention and regulatory efforts typically target one factor at a time.
References/instruments
- Prüss-Üstün, Annette, Corvalán, Carlos F & World Health Organization. (2006). Preventing disease through healthy environments : towards an estimate of the environmental burden of disease/Prüss-Üstün A, Corvalán C. World Health Organization. https://iris.who.int/handle/10665/43457
- The Human Exposome Project: EXPOSOME https://humanexposomeproject.com/
- Wild CP. Complementing the genome with an “exposome”: the outstanding challenge of environmental exposure measurement in molecular epidemiology. Cancer Epidemiol Biomarkers Prev. 2005 Aug;14(8):1847-50. doi: 10.1158/1055-9965.EPI-05-0456.
- Turning to Artificial Intelligence to Disentangle the Exposome – How AI can help analyze a lifetime of environmental exposures. Available at: https://hms.harvard.edu/news/turning-artificial-intelligence-disentangle-exposome
- CHAT GPT openAI – https://openai.com/chatgpt/
