14th May 2015
In my talk, I will discuss my collaboration with Nick Loman’s lab to develop de novo assembly methods for MinION data. We have built a pipeline to error correct nanopore reads using partial order graphs and the corrected reads are subsequently assembled using the Celera Assembler. The main focus of my talk will be on the post-assembly consensus calling stage, which uses the electrical current signal emitted by the MinION. I will describe in detail the probabilistic model we have developed for MinION signal data and our software tools for working in signal space. Our approach assembled the E. coli genome into a single contig with 99.4% accuracy after signal-level polishing.