Andrew Paul DeFilippis

Identification and characterization of a human antibody profile inversely associated with adverse cardiovascular events

The paper identifies a human antibody profile inversely associated with adverse cardiovascular events in two cohorts, showing it predicts higher risk when levels are low.

Inferring metabolite interactomes via molecular structure informed Bayesian graphical model selection with an application to coronary artery disease

The paper develops a Bayesian graphical model to infer metabolite interactomes using molecular structure information and applies it to coronary artery disease data.

Wisdom of artificial crowds feature selection in untargeted metabolomics: An application to the development of a blood-based diagnostic test for thrombotic myocardial infarction

This paper applies a Wisdom of Artificial Crowds (WoAC) strategy to feature selection in untargeted metabolomics for developing a blood-based diagnostic test for thrombotic myocardial infarction, achieving lower misclassification rates than Lasso and Elastic Net.

Evaluation Of Classifier Performance For Multiclass Phenotype Discrimination In Untargeted Metabolomics

The paper evaluates classifier performance for multiclass phenotype discrimination in untargeted metabolomics using simulated and real data, finding that SVM and Random Forest outperform PLS-DA and Neural Networks in most scenarios.