Improvise, Adapt and Overcome: retention time prediction in the context of proteins and biomolecules
This presentation features the work of Robbin Bouwmeester, a senior postdoc in the CompOmics group in Ghent, Belgium, who completed his PhD in 2020. Robin is an expert in artificial intelligence and retention time prediction. His presentation, titled "Improvise, Adapt and Overcome," focuses on retention time prediction in the context of proteins and biomolecules.
He discusses the importance of identifying and quantifying biomolecules in cells, and the use of graph networks to improve prediction accuracy. Robbin highlights the benefits of separating amino acids as individual graphs and combining them with code enrollments, resulting in an 11% performance improvement.
He also touches on the use of chemical descriptors and graph representations to better distinguish peptides, and the potential of training on broader peptide sets before fine-tuning to smaller, more complex data sets. Additionally, he showcases a successful example of accounting for modifications in retention time prediction with high correlation coefficients.
Learning points:
- Importance of Transfer Learning
- Role of Modifications in Retention Time Prediction
- Application of Machine Learning
Who should attend:
- Researchers and professionals in the fields of bioinformatics, computational biology, and analytical chemistry, particularly those who are involved in proteomics and the study of biomolecules
- Individuals working with machine learning applications in biological sciences or those interested in retention time prediction and data-driven approaches to understanding biomolecular behavior would find the presentation valuable.
Presenter: Robbin Bouwmeester (CompOmics Group, Ghent University/VIB, Belgium)
