SOT Supports Graduate Student Training in Environmental Mixtures
My name is Lorraine N. Vélez-Torres, and I am currently a PhD candidate in the Microbiology and Medical Zoology Department at the Universidad de Puerto Rico Recinto de Ciencias Médicas. My research project focuses on assessing the effect of Hurricane María on indoor air concentrations of fungal spores among homes affected by different degrees of water damage and evaluating the immunotoxicity potential of Aspergillus species in humans. In Puerto Rico, fungal spores have been found to be triggers of asthma and allergies. However, other aeroallergens that are present in the air, such as pollen, Saharan dust, and pollutants, may contribute together with fungal spores in worsening respiratory conditions symptoms.
While exploring how to analyze this complex mixture of aeroallergens and particles, I found the Columbia University Environmental Mixtures Workshop. This course is a two-day intensive training that provides basic concepts, techniques, and data analysis methods for the assessment of environmental mixtures (multiple pollutants at once) in environmental health studies. Although the workshop was going to be offered virtually because of the COVID-19 pandemic, I still had to cover the registration fee. I applied for and received the SOT Supplemental Training for Education Program (STEP) award, which allowed me to attend this workshop and pursue my goal of learning new statistical approaches for the analysis of environmental mixtures (such as air samples).
Dr. Brent Coull, Dr. Chris Gennings, Dr. Jeff Goldsmith, and Dr. Marianthi-Anna Kioumourtzoglou were the team of experts in statistics and environmental epidemiology that led the workshop. The goal of this workshop was to provide some examples of different approaches to answer various types of environmental health research questions and apply these methods using tools such as RStudio. The first thing I learned from this training was that there’s no strict definition for a mixture, but it can be defined as a combination of three or more chemicals, chemical groups, or nonchemicals. Studying exposure to environmental mixtures is very important since we are exposed to multiple (hundreds, even thousands) chemicals at a single time and this can have an effect on our health. Next, an important statement that the presenters asked us to consider is what mixture analysis method to select according to the primary research question in your study. The team emphasized this by presenting a “bubble” illustration of potential questions in environmental analyses (exposure pattern recognition, toxic agent identification, interactions and nonlinearities, a priori defined groups, and overall mixture effect estimation) with an overview of their corresponding existing mixtures statistical methods (e.g., BKMR, WQS, K-means, Clustering, PCA, Factor Analysis, Group Lasso).
During the two training days, I became familiar with the statistical approaches of Principle Component Analysis (PCA), Factor Analysis (FA), Clustering, Variable Selection (Lasso, elastic net), Bayesian Kernel (BKMR), and Weighted Quantile Sum Regression (WQS). The presenters explained these complex methods in a simple and understandable way. Throughout the entire virtual training, the instructors answered questions during and at the end of each topic.
The meeting also fostered networking, allowing further conversation between the participants and the instructors during short breaks after every two topics: the lunch break and the interactive panel discussion at the end of the day. During the last day’s interactive panel discussion, I asked a question regarding the use of the presented techniques in the context of my research field. Dr. Brent Coull kindly answered my question and provided me with scientific journal examples via email. All the course resources and materials were provided beforehand, including the email of the course in case of further questions after the end of the training.
I had a very enriching experience with this course. I was able to learn about current new and most-used methods for statistical analyses in environmental health. Furthermore, it is very important to use these techniques in research, since traditional epidemiological studies have only focused on single chemical analyses, which do not represent the reality. My goal is to apply the learned approaches to my current research, publish the results, and share the learned techniques to my lab mates, mentors, and other students. Also, I plan to use the knowledge from this workshop in my future career as a scientist in the fields of mycology, immunotoxicology, and aerobiology.
I’m extremely grateful for the opportunity SOT provided me with the STEP award! I encourage all students to take advantage of the programs offered by SOT and specifically to apply for the STEP award to gain professional development.