Welcome to Loving Science!

Hi, I'm Héléna, a chemoinformatician currently doing a postdoc at King's College London. This website gathers some of my research projects.

I'm interested in using omics and ML to better understand how drugs work - and hopefully find new treatments with less side effects!

For any information or request, contact Héléna A. Gaspar, PhD, King's College London.

Email: helena.gaspar@kcl.ac.uk

Github: https://github.com/hagax8


Drug Targetor: genetics-driven drug-target networks

This website harnesses results from genome-wide association studies (GWAS), and drug bioactivity data, to prioritize drugs and targets for a given phenotype. Drug Targetor networks are constructed using genetically scored drugs and genes, connected by the type of drug-target or drug-gene interaction. The interface currently contains information for 530 phenotypes.


Navigome: navigate the human phenome

The website is based on a collection of 465 phenotypes from different GWAS studies. Here, you will be able to:

  • Visualize an interactive 2D map of phenotypes based on genetic correlations
  • Browse pathways, gene and tissue analyses, using interactive visualizations
  • Generate gene profiles across phenotypes and tissues

ugtm: a python package

ugtm: a python package for Generative Topographic Mapping

ugtm is a python package to generate GTMs (generative topographic maps) and GTM-based classification and regression models. GTMs (by C. Bishop) are an interesting and probabilistic alternative to Kohonen maps - they usually give results comparable to t-SNE.