Work package 5
Analyzing the exposome in the PRIL cohort
The exposome of PRIL participants (WP4) will be explored through examining a combination of pre-existing data (e.g. map data, LUR [land use regression]-derived exposure data) and the generation of new data via wearable devices (WP5) and biological testing (WP6). Together, this will give a better insight into risk factors for the development of lung diseases, and lead to the detection of biomarkers and treatable targets for early lung damage.
The wish to non-invasively monitor one’s personal health status and to assess the quality of the environment is increasing in both the general population and many patients with chronic lung diseases. These measurements and their analyses could lead to practical advice for both patients and healthy subjects in order to prevent respiratory disease and improve its treatment and prognosis. It could also lead to the development of new and innovative tools for health scientists and governments to develop new strategies, policies and therapies.
New tools and devices are being developed to assist in the generation of such measurements and assessments. Also, ever increasing data sets are becoming available to evaluate population health and the quality of the environment (e.g. LUR models of ambient air pollution, pollen dispersal maps). The necessary next step is to integrate these datasets and connect them to the data generated from measurement devices and biomonitoring. For the analysis of this wealth of data, simple correlation metrics are often insufficient due to the often-high co-linearity of variables (e.g. ambient air pollution constituents commonly show high co-linearity). Therefore, new and innovative data management techniques and software are needed.
We will develop advanced data management techniques, including software machine learning algorithms. The second aim is to develop and deploy non-invasive devices, minimally invasive diagnostic tests and simple environmental sensors for the day-to-day use of healthy subjects and patients with respiratory conditions. The third aim is to develop tools and apps to integrate population level data with personalized data and generate lifestyle and medication advice. The last aim is to generate data on the exposome of participants in the PRIL cohort.
The first aim of this work package is to improve data management. As stated, co-linearity of variables, especially in a data-rich environment such as the exposome, limits the effectiveness and interpretability of standard statistical techniques such as regression analyses. Therefore, advanced data management techniques including “big data” approaches, machine learning algorithms, dimension reduction techniques (e.g. PLS models) and differing variable selection approaches (e.g. penalized regression) will be explored and developed. Furthermore, we will work on an administrative and legal framework, as well as standard operating procedures necessary for large scale data collection, and user feed-back.
The second aim is to develop innovative environmental monitors. The development of new wearable sensors focuses on detection in changes in the environment which can be connected in a personalized way to perturbations in bodily functions. e.g. connect pollen or pollutants to respiratory rates. Recent advances means that a variety of different sensors are both under development and available for use, which can be deployed for further testing and generating new data within the PRIL cohort.
Personal air pollution monitors4,5
The deployment of light-weight real-time (e.g. DiscMini) and time-weighted (e.g. UPAS) air pollution monitors will be used to further identify components of the external exposome and their relationship to human health. Linking these components (e.g. the DiscMini) directly to the ExpoApp (aim 3) allows for an integrated real-time overview of participant activity and exposure.
Polydimethylsiloxane (PDMS) wristbands (see Figure 6) can be rapidly and easily deployed into the field for the passive sampling of several hundred organic compounds, including PCBs, PAHs, and pthalates. Participants wear wristbands for days to weeks before they are collected for analysis.
Propeller Health integration7,8
The propeller sensor is a blue-tooth enabled sensor which can be attached to the head of a typical inhaler (see Figure 7). When deployed, the sensors record the date, time, and number of puffs each time a participant uses their medication. This information can be supplemented with (amongst others) geolocation, temperature, humidity and real-time air pollution (e.g. via personal DiscMini) information to identify specific environmental contributors to inhaler use.
The third aim is to develop tools and apps to integrate data. The data gained above represents a wide array of biological and environmental data. This data will be supplemented via integration with existing environmental data, covering a wide range of components including annual average ambient air pollution, green spaces, and pollen density.
Multiple tools are required for the integration of the wealth of population level data with individual data. One of such tools is the ExpoApp9; a smartphone app developed to assist in the assessment of environmental exposures at the individual level. It integrates the following components to give an overview of the individual environment:
- Geolocation and accelerometery via a waist-mounted smartphone
- Data from portable air pollution monitors
- Geographic information systems
- Participants’ information
Additionally, several new components for the app are being developed including new measurement capabilities (noise, electromagnetic fields, and temperature) and the ability to transmit short burst questionnaires to participants (e.g. individual level behaviour, social interactions, general wellbeing).
Personalized data obtained from the PRIL participants can also be integrated into population level data. The Exposome Hub10 is a new multidisciplinary structure which facilitates interaction between stakeholders. The goal of the hub is for researchers from different research areas and partners such as TNO, RIVM and Nutricia Research to investigate the exposome and map the influences on our health. The data within the hub will furthermore be supplemented by a wealth of already generated population level data that is currently available.
Integration of established air pollution metrics into the Hub
There are multiple established models and measurements of ambient air pollution within the Netherlands which can be integrated into the above active monitoring devices. Annual average ambient concentrations of multiple pollutants, including PM2.5, black carbon, and NO2, at the residential address of participants can be achieved via integration with the developed GIS models within the ESCAPE11 and ELAPSE12 projects. Further, the integration of real-time measurements from official monitoring stations13 provides validation for wearable sensors and additional sources of exposure data.
Integration of established pollen metrics into the Hub
Pollen data will be derived from European pollen-load maps14. Pollen data was measured at approximately 300 monitoring stations across Europe and long-term averaged have been estimated for multiple allergens over 3 periods each month. This data will be integrated into the above additional data sources to identify the contribution of pollen to respiratory health.
Integration of key emission sources into the Hub
Utilizing GIS and mapping technologies (e.g. OpenStreetMap, Google Maps) the distance from participants’ residence (or any required GPS co-ordinate) to key emission sources will be calculated and integrated. The following sources will be examined:
- Industrial sites 15
- Petrol stations
- Green spaces
- Blue spaces
Integration of personalized data (the exposure and biomarker metrics of PRIL participants) into the hub requires developing communication networks with different partners and stakeholders, allowing communication within the network and rapid collaboration across disciplines.
The forth aim is to collect data on the exposome of participants of the PRIL cohort. One of the unique features of the PRIL cohort is that it will include a population who do not currently have lung disease but who have been a-priori identified as being susceptible. Collection of data on the exposome at baseline and during follow up will allow for the building of a database of individuals without disease.
The first deliverable of this work package is a data management system. The second deliverable is new wearable sensors that detect changes in the environment and connect these in a personalized way to perturbations in bodily functions (e.g. connect data on pollen or pollutants to respiratory rates). The third deliverable is tools and apps to integrate population level data to personalized data and generate lifestyle and medication advice. The last deliverable is data on the exposome of participants in the PRIL cohort. This data will be linked to the other biomarkers acquired in the PRIL cohort (WP6/7) in WP8.
6. O'Connel et.al. Silicone Wristbands as Personal Passive Samplers. Environ. Sci. Technol. 2014;48(6):3327-3335.
9. Donaire-Gonzalez et al. ExpoApp: An integrated system to assess multiple personal environmental exposures. Environ Int. 2019;126:494-503