Christian Grønbæk

The paper trains a neural network to predict DNA bases from context, achieving 54% accuracy on the human genome, outperforming Markov models and revealing periodic signals linked to nucleosome positioning.

The paper develops context-dependent nucleotide models for the human genome, showing they can predict bases with over 50% accuracy and fit somatic variants better than simpler models.

GISA identifies rare protein conformations by computing Gauss integrals as topological-geometric descriptors, revealing new entangled structures and potential errors in protein models.