November 28, 2016 – Lila is a third year graduate student in Dr. Shaun Mahony’s lab. This week she presented her work aimed at 3-D genome reconstruction from kilo base-pair resolution Hi-C data. The knowledge of a three-dimensional chromatin structure is important for studying many biological processes at the molecular level such as transcription regulation, repression, and stress response. The resolution of Hi-C data has improved continuously over the past few years. We now have whole genome interaction data at up to kilo-base resolution. However, this also results in an increase in data size, and traditional methods for 3-D reconstruction such as classical multi-dimensional scaling (MDS) cannot handle such large input matrices, for lack of both time and memory. Further, a large part of the Hi-C matrix is extremely sparse, requiring interpolation. Turning this to her advantage, Lila developed a method called miniMDS to reconstruct the genome in a piece wise fashion. Since TADs are less sparse, miniMDS splits the Hi-C distance matrix into subsets based on TADs, and attempts to recreate these smaller genomic segments in three dimensional space. It then uses a low-resolution whole genome structure as a scaffold to align these various complex smaller structures, thus improving accuracy of the over all global structure. It is much faster than the existing methods, as well as memory efficient. Further, since this method first recreates local structures, it is able to capture functionally relevant intricacies within TADs that would be missed if we directly try and create a global structure. Thus, it is a very useful tool for genome reconstruction at high-resolutions, and we look forward to seeing where this project goes!