A5071954806- RNA and protein synthesis mechanisms
- Monoclonal and Polyclonal Antibodies Research
- Machine Learning in Bioinformatics
- Genomics and Phylogenetic Studies
- Protein purification and stability
- Glycosylation and Glycoproteins Research
- Machine Learning in Materials Science
- Software Engineering Research
- vaccines and immunoinformatics approaches
- Biochemical and Structural Characterization
- Protein Structure and Dynamics
- demographic modeling and climate adaptation
- Mathematical and Theoretical Epidemiology and Ecology Models
- COVID-19 epidemiological studies
- Web Data Mining and Analysis
University of Oxford
2023
Vrije Universiteit Amsterdam
2020-2022
Antibodies are the largest class of biotherapeutics. However, in recent years, single-domain antibodies have gained traction due to their smaller size and comparable binding affinity. (Abs) (sdAbs) differ structures sites: most significantly, lack a light chain so just three CDR loops. Given this inherent structural difference, it is important understand whether Abs sdAbs distinguishable how they engage partner thus, suited different types epitopes. In study, we use non-redundant sequence...
Antibodies play an important role in clinical research and biotechnology, with their specificity determined by the interaction antigen's epitope region, as a special type of protein-protein (PPI) interface. The ubiquitous availability sequence data, allows us to predict epitopes from order focus time-consuming wet-lab experiments toward most promising regions. Here, we extend our previously developed sequence-based predictors for homodimer heterodimer PPI interfaces residues that have...
Self-supervised language modeling is a rapidly developing approach for the analysis of protein sequence data. However, work in this area heterogeneous and diverse, making comparison models methods difficult. Moreover, are often evaluated only on one or two downstream tasks, it unclear whether capture generally useful properties. We introduce ProteinGLUE benchmark evaluation representations: set seven per-amino-acid tasks evaluating learned representations. also offer reference code, we...
Protein protein interactions (PPI) are crucial for functioning, nevertheless predicting residues in PPI interfaces from the sequence remains a challenging problem. In addition, structure-based functional annotations, such as interface scarce: only about one-third of all structures residue-based annotations available. If we want to use deep learning strategy, have overcome problem limited data availability. Here multi-task strategy that can handle missing data. We start with model...
Abstract Motivation Antibodies play an important role in clinical research and biotechnology, with their specificity determined by the interaction antigen’s epitope region, as a special type of protein-protein (PPI) interface. The ubiquitous availability sequence data, allows us to predicting epitopes from order focus time-consuming wet-lab experiments onto most promising regions. Here, we extend our previously developed sequence-based predictors for homodimer heterodimer PPI interfaces...
Abstract Antibodies are the largest class of biotherapeutics. However, in recent years, single-domain antibodies have gained traction due to their smaller size and comparable binding affinity. (Abs) (sdAbs) differ structures sites: most significantly, lack a light chain so just three CDR loops. Given this inherent structural difference, it is important understand whether Abs sdAbs distinguishable how they engage partner thus, suited different types epitopes. In study, we use non-redundant...
Mathematical models play a crucial role in understanding the spread of infectious disease outbreaks and influencing policy decisions. These aid pandemic preparedness by predicting outcomes under hypothetical scenarios identifying weaknesses existing frameworks. However, their accuracy, utility, comparability are being scrutinized. Agent-based (ABMs) have emerged as valuable tool, capturing population heterogeneity spatial effects, particularly when assessing intervention strategies. Here we...
Abstract Protein protein interactions (PPI) are crucial for functioning, nevertheless predicting residues in PPI interfaces from the sequence remains a challenging problem. In addition, structure-based functional annotations, such as interface scarce: only about one-third of all structures residue-based annotations available. If we want to use deep learning strategy, have overcome problem limited data availability. Here multi-task strategy that can handle missing data. We start with model...
ABSTRACT Self-supervised language modeling is a rapidly developing approach for the analysis of protein sequence data. However, work in this area heterogeneous and diverse, making comparison models methods difficult. Moreover, are often evaluated only on one or two downstream tasks, it unclear whether capture generally useful properties. We introduce ProteinGLUE benchmark evaluation representations: set seven tasks evaluating learned representations. also offer reference code, we provide...
Abstract Self-supervised language modeling is a rapidly developing approach for the analysis of protein sequence data. However, work in this area heterogeneous and diverse, making comparison models methods difficult. Moreover, are often evaluated only on one or two downstream tasks, it unclear whether capture generally useful properties. We introduce ProteinGLUE benchmark evaluation representations: set seven tasks evaluating learned representations. also offer reference code, we provide...