Deep Generative Modelling of Epigenomics Data

Dr Mile Sikic

Agency for Science, Technology and Research

Project Description

DNA modifications contribute to diseases such as cancer and it has been found to be highly predictive of age, demonstrating the value of epigenomics data to understand the profile of each individual patient. We plan to use an anomaly detection method based on deep learning models that can identify DNA modification from nanopore sequencing data. Existing approaches are limited to specific type modification and work on a subset of modified and unmodified nucleotides or previously trained sequences. Successful detection of DNA modifications in an accurate manner would enable large scale profiling of patients in Singapore and opens new venues for large scale projects related to use of AI in genetics.

Research Technical Area

Machine learning

Benefits to the society

Precision Medicine is a key strategic area with huge potential benefits to improve health care due to the personalised evaluation and treatment of each patient. If we are successful with this project, this will open up many possibilities to apply AI to epigenomics data to improve healthcare in Singapore.

Team's Principal Investigator

Dr Mile Sikic
Genome Institute of Singapore
Agency for Science, Technology and Research

Introduction of the Principal Investigator

Mile Sikic currently works as a group leader at Genome Institute of Singapore, A*STAR. He also holds professorship for computer science at University of Zagreb, Croatia (leave). His research interests include use of string and graph algorithms, AI and complex network to deal with problems related to genetics, epidemiology and social networks.

Recent Notable Awards

  • 2015, University of Zagreb, FER – annual award for the paper “Identification of Patient Zero in Static and Temporal Networks ” published in Physical Review Letters

Team

Co-Principal Investigator

Dr. Lee Hwee Kuan

Bioinformatics Institute,
Agency for Science, Technology and Research

Research Areas:

  1. Machine learning
  2. Vision
  3. Heuristic search and optimisation

Dr. Jonathan Göke

Genome Institute of Singapore
Agency for Science, Technology and Research

Research Areas:

  1. Machine learning