Dr. Lars Gabriel

I am a computational biologist interested in how modern machine learning can be used to solve real problems in genomics. My work combines method development, large-scale data analysis, and research software engineering.

In recent years, I have worked primarily on gene structure prediction and genome annotation and led the development of the open-source genome annotation tools Tiberius and BRAKER3.

What I enjoy most is building methods that are scientifically solid, useful in practice, and accessible to other researchers.

My academic background is in computational biology and bioinformatics, with a strong emphasis on machine learning and genome analysis. During my doctoral and postdoctoral work, I have developed methods for genome annotation using deep learning and built software for large-scale sequence analysis.

Working daily on HPC systems has also made me care a lot about tools that are reliable in practice and easy for others to use.

My current work centers on deep learning for genome annotation across diverse eukaryotic clades. I am particularly interested in improving models and integrating different sources of evidence, such as genomic sequence, RNA-seq, and protein information, to improve annotation quality.

More broadly, I am interested in applying these models to biological sequence analysis tasks beyond genome annotation, especially where machine learning can support scalable and useful scientific tools.