New statistical methods for the analysis of genome variation data - Richard Durbin
1,429 views · Published 16 April 2016 · 46:23 · Indexed 22 October 2026
Channel: bhadeshia123 · 2016 · Education
The last two decades have seen an exponential growth in the quantity of DNA sequencing data, at a rate more than twice as fast as Moore’s law for growth in computing power. Individual sequences at one position in the genome are related by descent from a common ancestor in a tree, but this genealogy is made much more complex across the whole genome by recombination, which means that you inherit different sections of your genome from different recent ancestors. A clean mathematical model exists for the full genealogy but it is too complex to use for inference with data sets available today. I will discuss various algorithmic and representational innovations introduced to model and exploit genetic relatedness in analysing ever larger amounts of genome data, illustrating them with our work on the Haplotype Reference Consortium, which has assembled over 65,000 whole genome sequences with data at nearly 40 million variable sites. Part of a symposium to celebrate the work of Professor Sir David MacKay FRS. The symposium was held over the period 14-15 March 2016. https://mlg.eng.cam.ac.uk/djcms2016/
