Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Lance D. Hentges, Martin J. Sergeant, Damien J. Downes, Jim R. Hughes, Stephen Taylor (University of Oxford MRC WIMM Centre for Computational Biology, Oxford, UK)
Despite the ease with which humans can visually identify peaks, converting signal from ATAC, ChIP, and DNase-seq into meaningful genome-wide peak calls requires complex analytical techniques. Current methods rely on statistical frameworks to identify peaks as sites of significant signal enrichment, discounting that the analog data do not follow any archetypal distribution. Recent advances in artificial intelligence have shown great promise in image recognition, on par or exceeding human ability, providing an opportunity to reimagine and improve peak calling. We present an interactive and intuitive peak calling framework, LanceOtron, built around image recognition using a wide and deep neural network. We hand-labelled 736,753 regions covering 499Mb of genomic data, built 5,000 models, and tested with over 100 unique users from labs around the world. LanceOtron outperforms the long-standing, gold-standard peak caller MACS2 with its increased selectivity and near perfect sensitivity. In addition to command line accessibility, an intuitive and freely available web application was designed allowing any researcher to easily generate and visualise optimal peak-calls.