04/21/2026
By Irma Silva

The Kennedy College of Sciences, Department of Biological Sciences, invites you to attend a Ph.D. Dissertation Defense in Applied Biology by Panagiotis Nikolaos Lalagkas entitled: “Pleiotropy-based methods for discovering therapeutic and adverse effects of drugs on disease”

Candidate: Panagiotis Nikolaos Lalagkas
Date: Wednesday, May 6
Time: 11 a.m.-1 p.m.
Location: Olsen 234

Committee members:

  • Rachel Melamed, Assistant Professor, University of Massachusetts Lowell
  • Frederic Chain, Associate Professor, University of Massachusetts Lowell
  • Teresa Lee, Assistant Professor, University of Massachusetts Lowell
  • Jonine Figueroa, Senior Investigator, National Cancer Institute, Division of Cancer Epidemiology & Genetics, NIH

Title: Pleiotropy-based methods for discovering therapeutic and adverse effects of drugs on disease

Brief Abstract:
Drug discovery is slow, expensive, and with a failure rate exceeding 90%. Human genetics has helped identify variants associated with disease risk, but linking these signals to causal genes driving disease remains a major challenge. This limits the ability to translate genetic findings into effective drug gene-targets. In this thesis, I address this challenge using pleiotropy, the phenomenon in which a genetic variant or gene influences multiple diseases. By leveraging these shared genetic effects, I develop methods to transfer biological insights from one disease to another and accelerate drug discovery. First, I show that biological knowledge from monogenic (Mendelian) diseases can be used to identify promising drug targets for clinically associated (comorbid) complex diseases. I then extend this approach to pairs of complex diseases, using breast cancer as a case study, and identify shared (pleiotropic) genes that can be used to prioritize existing drugs approved for predisposing diseases that target the identified shared genes and pathways they participate in. Next, I develop metrics to quantify genetic similarity across 178 common diseases and show that diseases with higher genetic similarity are more likely to share both drug treatments and side effects, even when analyzing clinically dissimilar disease pairs. Using this insight, I then use genetic similarity to predict new uses and side effects for 1,711 drugs. Notably, predicted drug indications are twice as likely to progress from Phase I trials to regulatory approval and predicted side effects are enriched for true adverse drug outcomes. Finally, I show that drug efficacy and toxicity often share a common genetic basis. Building on this insight, I develop a neural network model that learns gene‑specific scores quantifying each gene’s relative contribution to therapeutic and adverse effects of a drug on disease. Overall, this work presents pleiotropy as a scalable and interpretable framework for improving drug discovery, enabling better target selection, drug repurposing, and earlier detection of safety risks.