Hypothesis
Our central hypothesis is that proteins are key regulators of human health and can serve as markers of traumatic injury and cardiometabolic health. Proteomics analysis in the ADVANCE study, therefore, provides a novel avenue for biomarker identification and potential therapeutic targeting of cardiometabolic health-related conditions.
Summary
Our objective is to analyze the proteomics data in combination with the clinical information, blood tests and imaging/PWA results to identify novel markers of cardiometabolic health and disease in ADVANCE participants. Our specific aims are as follows:
1. Exploratory analysis of the proteomics data for any technical variation.
2. Integrative and confounder analysis of the proteomics and phenomics (clinical and imaging) data for marker identification.
3. Validation/replication of identified markers and exploration of underlying mechanisms using external cohorts and advanced statistical methods.
The following methods will be employed for each of the aims described above:
1. QC analysis using analytical tools provided by Somalogic here (https://github.com/SomaLogic).
2. Linear methods including univariate and multivariate approaches such as linear regression, partial correlations, principal component analysis as well as discriminant analysis using PLS or OPLS methods.
3. Machine learning-based classification and regression approaches such as gradient boosting decision trees and t-SNEs as well as pathway analysis, mediation analysis or Mendelian Randomization as needed.
Keywords
Proteomics, Biomarkers, Machine Learning, Cardiovascular, Cardiometabolic