• Genome Data Science

    We develop methods and tools to work with tens of thousands of genomes and analyze and integrate the corresponding data.

    Model of DNA double helix in front of a student.
    © Universität Bielefeld

Thesis Topics & Student Projects


GENERATING GENOMES USING LANGUAGE MODELS

  • Genomes, the DNA of living organisms, follow a grammatical structure
  • The rules are implied by the laws of genetics
  • Idea: Structure DNA in terms of applicable words and sentences
  • Make language model pick up the grammar that underlies the resulting sentences
  • Application: Language models capture the structure of diseases and other (biomedically) relevant particularities of (human) genomes

GRAPH DATABASES / DRUG REACTION PREDICTION

  • We know functional relationships for many drugs and genes I Would like to predict new drugs and their behaviour
  • Use advanced machine learning for prediction
  • Exploit knowledge arranged in graph databases
  • Project involves implementation of advanced prediction techniques not yet exploited in clinical practice

GRAPH NEURAL NETWORKS / POLYPHAMARCY SIDE EFFECTS

  • Identify biological pathway of interest
  • Screen drugs relative to their effect on pathway
  • Implement graph neural networks to systematically exploit effects
  • Investigate pathways using Gaussian graphical model
  • Project involves analysis of multilevel, complex data hierarchies

INDEX STRUCTURES FOR PANGENOME GRAPHS

  • Pangenome graphs are modern, advanced data structures for storing (hundreds of) thousands of genomes in compressed form
  • Indexes enable us to navigate these graphs systematically
  • Project involves algorithmically tricky techniques

PREDICTING ANTIMICROBIAL RESISTANCE USING TRANSFORMERS / CAPSULE NETWORKS

  • Idea: predict resistance behaviour of pathogenic strains
  • Use advance machine learning to predict this from genome (DNA) data
  • Project involves implementation of advanced machine learning techniques for novel, clinically relevant application scenarios


Just come and talk to us on in our offices at U10 and we will work out an exciting project with you!