SKM 2023 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

BP: Fachverband Biologische Physik

BP 11: Poster Session I

BP 11.16: Poster

Dienstag, 28. März 2023, 12:30–15:30, P1

Machine Learning Guided RNA Contact Prediction — •Utkarsh Upadhyay¹, Oskar Taubert², Christian Faber³, and Alexander Schug⁴ — ¹Forschungszentrum Jülich, Jülich, Germany — ²Karlsruher Institut für Technologie, Karlsruhe, Germany — ³Forschungszentrum Jülich, Jülich, Germany — ⁴Forschungszentrum Jülich, Jülich, Germany

For around 50 years, the primary focus of genomic research has been the development of efficient and accurate methods to predict the structure of proteins, which led to the birth of better sequencing techniques and databases. About 98% of the human genome(RNA, DNA) during this action was overlooked.

RNA is not merely a messenger for making proteins, in the past few years, studies have revealed the existence of many non-coding RNAs which catalyse various biological processes; to gain detailed insights into these roles, we require the appropriate structure. Recent years have led to breakthroughs in protein structure prediction via Deep Learning. The scarcity of RNA structures, however, makes a direct transfer of these methods impossible.

We predict contact maps as a proxy to understand and predict RNA structure, they provide a minimal representation of the structure. We have worked on methods that took accuracy from 47%(DCA) to 77%(CoCoNet) and now to 87%(Barnacle). Further, we are trying to create more efficient neural networks for working with limited data, using statistical physics and ML techniques, to substantially reduce the sequence-structure gap for RNA.