Research topic: Decomposable latent representations for in-vivo toxicity prediction
Project description: In the perspective of improving drug design with AI support, it is important to incorporate information on possible compounds tox liability as early as possible in the design. While the most indicative data for compound toxicity is recorded as pre-clinical pathologies or further as clinical indications, this data is very scarce. However, we would aim to enhance that information by modeling in the context of other heterogeneous and multi-fidelity data sources. In the project decomposable latent representations for in-vivo toxicity prediction we aim at developing state-of-the-art, novel machine learning methods to learn chemical representations that factorize into intuitive chemical or pre-clinical pathologies from public and proprietary data. These models will be developed to assess tox liability for compounds to support drug discovery in the very early stage. The goals of the project include learning under data scarcity, integration of heterogenous and multi-fidelity data sources, and model interpretability.
Personal Introduction: I have a background in Chemical Engineering, and I completed my internship and master’s thesis at Theis Lab, HMGU Munich. My thesis involved representation learning from Multiplex protein data by using self-supervised deep learning methods to analyze the impact of cancer drugs at the single cell level.
AIDD consortium has provided me with a unique platform to combine my background and skills to fast-track the drug discovery process. My current project involves the development of Bayesian approaches for the reliable prediction of in-vivo toxicity. We aim to combine the chemical and biological space by leveraging the chemical structural information and the activity profiles of in-vitro assays. Moreover, another objective is to incorporate the developed methodology into one chemistry model that unifies the toxicity prediction task with other ADME properties tasks.
I am passionate to bridging the gap between life sciences and machine learning through my work because this integration has the potential to impact millions of lives.
Contact: GitHub LinkedIn Twitter
Presentations at conferences and meetings:
- Masood, A., Heinonen, M., Herman, D., Ceulemans, H. Kaski, S. Dos-Time dependent DILI modeling. In Janssen Discovery Data Science meeting. May 08, 2023.
- Masood, A., Heinonen, M., Herman, D., Ceulemans, H. Kaski, S. Dos-Time dependent DILI modeling. In Finnish Center of Artificial Intelligence Virtual Drug Design Lab seminars. Feb 08, 2022.
- Masood, A., Heinonen, M., Herman, D., Ceulemans, H. Kaski, S. Dos-Time dependent DILI modeling. In Finnish Center of Artificial Intelligence Virtual Drug Design Lab seminars. October 10, 2022.
- Masood, A., Heinonen, M., Herman, D., Ceulemans, H. Kaski, S. Dos-Time dependent DILI modeling. In STB, Helmholtz Zentrum München. November 08, 2022.
Aalto University, Finland, September 1st, 2021 - February 28, 2023
Janssen Pharmaceutica NV, Belgium, March 1st , 2023 - August 31th, 2024
Secondment: Helmholz Munich, Germany November 2022