Abstract & Bio
 Title: “Sequencing Populations to Find Causal Genetic Variants ”




Dr Richard Mott
UCL Genetics Institute
Gower Street


We describe two approaches, based on sequencing a population at low coverage, to impute single nucleotide polymorphisms (SNPs) in the absence of a haplotype reference panel, and to detect structural variation (SVs).  These methods are applicable to any species including crops, and in particular populations descended from known inbred lines such as MAGIC populations. The first approach is described in (Davies et al 2016 Nat Genetics.48:965-969 PMID: 2737623). It assumes an unknown set of haplotypes (the reference panel) segregate at each locus in the population, and that each chromosome is a mosaic of these unknown haplotypes. It then simultaneously estimates the panel and the individual chromosome mosaics using a hidden Markov model that utilises features of low-coverage sequence data. The number of unknown haplotypes at each locus is a variable dependent on the population history. We apply this method to humans, mice and Arabidopsis MAGIC and show that common variants can be imputed accurately using only modest levels of sequence coverage (between 0.1x an 1.5x depending on the population history). The second approach also uses low-coverage sequence to identify regions that are structurally variant, and if they are due to a transposition, the originating locus of the transposition. Applying this method to the Arabidopsis MAGIC population (Imprialou et al 2017 Genetics 205:1425-1441. PMID: 28179367) identified thousands of SVs of which about a quarter were transpositions, and mapped SVs that were likely to be functional for fungal resistance and germination. Of interest, for some phenotypes structural variation explains more heritability than does SNP variation.


Richard Mott is Weldon Professor of Computational and Statistical Genetics in the research department of Genetics, Evolution and Environment at University College London. He was previously at the Wellcome Trust Centre for Human Genetics and a Professor by Research at Oxford University.

He has worked on physical mapping with Hans Lehrach at Imperial Cancer Research Fund laboratories in London, where he developed a suite of software tools for the construction and validation of physical maps in 1995, he moved to the Sanger Centre to work on DNA sequence assembly where he wrote software that automatically analysed sequencing trace data in order to edit DNA sequence assemblies. This was used extensively to accelerate sequence production. He wrote the sequence CAFtools assembly pipeline which was used for the pipeline assembly of the human and other genomes at Sanger, and developed software for spliced alignment of EST to genomic DNA.

Between 1999 and 2015 he worked at the Wellcome Trust Centre for Human Genetics where he served as Head of Bioinformatics and Statistical Genetics. In 2010 he stepped down to concentrate on his own research leading a group working on quantitative genetics in plants and mice. He moved to UCL in November 2015.

He has developed methods for mapping in an outbred stock of mice (the heterogeneous stock). He developed the HAPPY software package used for high-resolution QTL mapping which led to the identification of a quantitative trait gene underlying behavioral variation in mice. As part of an international collaboration he is developing a genetic reference panel of recombinant inbred lines of mice, known as the Collaborative Cross. With Dr Paula Kover, Bath University, he has developed a genetic reference panel in Arabidopsis thaliana