Periodic Reporting for period 1 - EpiBigDatainWomen (Epigenetic Data Integrated in a Big Data Approach to Unravel Novel Pathzays of CV Risk independent of classical CVD Risk Factors in Women.)
Período documentado: 2019-04-01 hasta 2021-03-31
CVD is the most prevalent cause of morbidity and mortality among women (and men) worldwide. Impacting on CVD prevention in this population can help identifying specific disease risk factors and signs of reversible early disease, in turs improving healthcare with important repercussions on public health costs.
Machine learning techniques have been developed to address prediction of health-related outcomes, however, multifactorial diseases such as CVD, still lack the integration of environmental data with the clinical information of the patients.
The project EpiBigDataInWomen aimed at identifying novel CVD-related gender-specific risk factors by combining genome-wide DNA methylation data with clinical, environmental, socio-economic information through machine learning statistical approaches in women and men with different CVD risk profiles.
In a subsequent phase of the project, we were able to retrieve data for each of the participant in the study relative to mortality and hospitalization rate, air pollution of the area where the individual lives, lifestyle and nutrition information.
From a preliminary analysis, we observed that the number of hospitalizations for a cardiovascular event was higher in the risk individuals in both sex groups, with men doubling the frequency between risk subjects and controls. Women with a low cardiovascular risk leave more frequently in urban areas compared to high-risk ones, while no significant difference was observed in the two men subgroups.
We used a total of 188-food items questionnaire, classified into 83 predefined food groups on the basis of similar nutrient characteristics or culinary usage to investigate the associations between global DNA methylation and nutrition measured at different levels of complexity (from micronutrients up to main food groups and adherence to Mediterranean Diet). We found that global DNA methylation is significantly associated with zinc and Vitamin B3 dietary intake via both classical and machine learning statistical approaches (Figure 1) (Noro F, … Izzi B. Fine-grained investigation of the relationship between human nutrition and global DNA methylation patterns. Eur J Nutr. 2021 Nov 6).
All the above-mentioned data, in combination with the drug prescription records, will be used to run more complex statistical multivariable approaches and will be used in the machine learning phase of this project.
The work management of EpiBigDataInWomen has been strongly impacted by the COVID-19 pandemic emergency, making it impossible to perform all tasks described in the WPs. Specifically, WP1 suffered from significant delays as lab activities were suspended for several months in 2020 and started again and at a reduced pace only in the second half of 2020. As a consequence, also machine learning analysis of the study cohort data, foreseen in WP2 still need to be performed at the time of submission of this report. This delay also affected the exploitation and dissemination of the results, activities that are currently on-going and will be further developed at the end of the planned experimental analyses.
Big data have the potential to improve medical care and reduce costs, both by individualizing medicine, and bringing together multiple sources of data about individuals. The data produced within this project will help predict whether a woman is likely to experience CVD in the near future, so that appropriate and personalized preventive/life saving strategies can be recommended to the patient. Therefore, the final findings from this project will serve as important information to help cardiologists in their daily follow-up of women, with enormous social impacts and economic benefits in terms of public health system expenses.