Despite advances in DNA sequencing, full accurate measurement of complex genomes remains a huge challenge. We have discovered that certain 3D structural patterns can be used to solve a range of problems in the field of genome assembly, including the identification of disease mutations that are currently difficult to detect. Using machine learning models, we are developing new ways to utilize data from 3D genome measurements to better characterize its 1D sequence in healthy and disease genomes.
The 3D organization of genomes is tightly linked to how the genetic information is accessed, regulated and propagated. Using machine learning, with a special emphasis on probabilistic models, we build computational models aimed to gain mechanistic insights of how 3D genome structures are specified and how they change in disease.