Homepage of Heiko Dunkel


Ongoing Research Projects

ncRNA classification | GitHub

This research is focussed on the classification of ncRNA sequences using the primary sequence and the secondary structure. The two-armed deep learning classifier gets the sequence as input for a CNN model and the graph-encoded secondary structure as input for a fully-connected NN. The two arms are concatenated in the end (Late Integration) and the input is predicted into one of the six ncRNA classes lncRNA, miRNA, rRNA, snoRNA, snRNA or tRNA. For more info see

Gene Expression Analysis of mice with cognitive disabilities

The FTSJ1 gene encodes for a 2'-O-Methyltransferase and FTSJ1 deficient humans have been shown to develop intellectual disabilities (Jensen et al. 2007). Additionally, Ftsj1 deficient mice have also been observed to show further symptoms including altered pain sensing, bone and energy metabolism (Jensen et al. 2019). In this project we are performing Gene Expression Analysis of Ftsj1 deficient mice to analyze the transcriptomic differences between mutant and wildtype mice.


Heiko Dunkel, Henning Wehrmann, Lars R. Jensen, Andreas W. Kuss, Stefan Simm (2023)
MncR: Late Integration Machine Learning Model for Classification of ncRNA Classes Using Sequence and Structural Encoding
Int. J. Mol. Sci. 2023, 24(10), 8884;

Teaching Experience