Bioinformatics Seminar
Time:
Venue: Na
22 March 2016
NaStatistical methods and software for functionally characterizing every single mutation in Your Favorite Gene
Alan RubinWEHI Bioinformatics
Although high-throughput DNA sequencing has rapidly expanded catalogues of normal and disease-associated variation ; the functional consequences of most mutations are unknown. In deep mutational scanning ; selection for protein function applied to a library of protein variants is combined with high-throughput DNA sequencing ; allowing direct measurement of the activity of hundreds of thousands of variants of the protein easily and cheaply. This approach helps to bridge the gap between variant identification and interpretation ; enabling researchers to elucidate sequence-function relationships at high resolution. The resulting data can be used in a variety of contexts ; from aiding the assessment of clinical variants to guiding protein engineering. Despite the growing popularity of deep mutational scanning ; there are few formal statistical methods available to help analyze these complex datasets. Alan will present a novel method based on linear regression and accompanying software implementation for assigning functional scores and statistical significance to all variants in a deep mutational scanning dataset. For illustration ; he will show the results of applying this method to deep mutational scans of diverse targets and demonstrate how these new methods allow researchers to obtain more information from these high-dimensional datasets.;