Diseases, the human microbiome and metabolites are densely linked. Abundance data and can help to understand the cause of a disease or ways to cure it.
Linear regression and multilayer perceptrons both are not suitable to describe the relation of abundace data and diseases or their symptoms, due to under- and overfitting.
The aim of this project is therefore to train and test new DNN architectures on different phylogenetic datasets at the HKI Jena in the next 3 years.
DNNs for Disease Classifaction on Phylogenetic Trees (LL DEEP - EGA)
Year of approval
2020
Institute
Leibnitz Hans KnölI Institute (GER)
Primary applicant
Hoffmann, S.