A5092249392- Computational Drug Discovery Methods
- vaccines and immunoinformatics approaches
- Biomedical Text Mining and Ontologies
- Machine Learning in Bioinformatics
- Receptor Mechanisms and Signaling
- Advanced Graph Neural Networks
- Genetics, Bioinformatics, and Biomedical Research
- Topic Modeling
- Metabolomics and Mass Spectrometry Studies
- Protein Structure and Dynamics
- Fault Detection and Control Systems
- Chemical Synthesis and Analysis
University College Dublin
2023-2025
IBM Research - Ireland
2023-2025
Conway School of Landscape Design
2023
Bioactive peptides are an important class of natural products with great functional diversity. Chemical modifications can improve their pharmacology, yet structural diversity presents unique challenges for computational modeling. Furthermore, data canonical (non-modified) is more abundant than non-canonical (chemically modified). We explored whether current methods sufficient to generalize from datasets. To do this, we first considered two critical aspects the modeling problem, namely choice...
Molecular modelling is a vital tool in the discovery and characterisation of bioactive peptides, providing insights into their structural properties interactions with biological targets. Many models predicting peptide function or structure rely on intrinsic properties, including influence amino acid composition, sequence, chain length, which impact stability, folding, aggregation, target interaction. Homology predicts structures based known templates. Peptide–protein can be explored using...
Automated machine learning (AutoML) solutions can bridge the gap between new computational advances and their real-world applications by enabling experimental scientists to build own custom models. We examine different steps in development life-cycle of peptide bioactivity binary predictors identify key where automation not only result a more accessible method, but also robust interpretable evaluation leading trustworthy
Quantifying model generalization to out-of-distribution data has been a longstanding challenge in machine learning. Addressing this issue is crucial for leveraging learning scientific discovery, where models must generalize new molecules or materials. Current methods typically split into train and test sets using various criteria, like temporal, sequence identity, scaffold, random cross-validation; before evaluating performance. However, with so many splitting criteria available, existing...
Recent research on predicting the binding affinity between drug molecules and proteins use representations learned, through unsupervised learning techniques, from large databases of molecule SMILES protein sequences. While these have significantly enhanced predictions, they are usually based a limited set modalities, do not exploit available knowledge about existing relations among proteins. In this study, we demonstrate that by incorporating graphs diverse sources modalities into sequences...
Abstract Motivation Automated machine learning (AutoML) solutions can bridge the gap between new computational advances and their real-world applications by enabling experimental scientists to build own custom models. We examine different steps in development life-cycle of peptide bioactivity binary predictors identify key where automation not only result a more accessible method, but also robust interpretable evaluation leading trustworthy Results present automated method for drawing...
Recent research on predicting the binding affinity between drug molecules and proteins use representations learned, through unsupervised learning techniques, from large databases of molecule SMILES protein sequences. While these have significantly enhanced predictions, they are usually based a limited set modalities, do not exploit available knowledge about existing relations among proteins. Our study reveals that representations, derived multimodal graphs describing proteins, lead to...