Natural Neural Networks Inspire Artificial Neural Networks in Therapeutic Design

Mentored by Martin Bogdan and Jens Meiler
at Leipzig University, Department of Neuromorphic Information Processing or Leipzig University, Institute of Drug Discovery

Artificial intelligence revolutionizes biomedical research as we speak. Examples are the deep learning neural network AlphaFold that allows the prediction of protein structure as well as PorteinMPNN which designs protein therapeutic candidates in seconds. Nevertheless, the neural network implementations used typically for such impressive results do not primary intent to reflect the natural neural networks occurring in the brains of humans and vertrabtes. Effects such as 'spiking', temporal delays, and neuronal heterogeneity are central to natural neural networks but often ignored in even the most advanced artificial neural networks. We hypothesize that inclusion of these effects into natural neural networks has the potential to further improve their predictive power.

We plan to test this approach on concrete questions including simple proof-of-concept questions of learning neuronal spike sequences as well as questions in therapeutic design, in particular the design of vaccine candidates targeting viral infections such as SARS-Cov-2. In this project, the PIs Martin Bogdan and Jens Meiler bring their expertise together to address this need. In Martin Bogdan's group neuromorphic information processing has been studied (https://doi.org/10.1007/978-3-642-15819-3_23). Jens Meiler’s group focuses on understanding the molecular foundation of protein function through structural modeling using the Rosetta software package and deep learning methods (https://doi.org/10.1021/bi902153g). We will collaborate with Stefan Hallermann (Leipzig University, Medical Faculty) who studies neuronal networks in the vertebrate brain and implements the findings in biology-based neuronal networks (https://doi.org/10.7554/eLife.51771). In this interdisciplinary project, the required expertise is combined to develop novel approaches of artificial intelligence.

Work environment

Depending on final arrangements, appointment will in either of the labs at LU Leipzig. We expect the trainee to travel frequently between the laboratories. The Interdisciplinary Center for Bioinformatics in Leipzig and cooperate with the machine learning and AI experts at the ScaDS.AI in Leipzig and Dresden. You will have access to machine-learning HPC resources of the ScaDS.AI Center for Scalable Data Analytics and AI Dresden/Leipzig, offering state-of-the-art CPU (Intel and IBM POWER 9) and GPU (Nvidia A100 and V100) resources, as well as high-performance computing and storage systems for computational experiments.

Prerequisites

The candidate will need a solid background in bioinformatics or structural biology as well as proficiency in scientific programming with Python and interest in data analysis. Prior expertise in machine learning, genomics or transcriptomics, and deep learning will be an advantage.

Further details on the requirements and application process can be found in SECAI's announcement for open PhD positions in 2024.