Population Sequencing for Studying Natural and Artifcial Variation in C. elegans

Abstract

The advent of high coverage and low cost sequencing technologies has allowed fornewer and more powerful approaches in molecular and population genetics. Transposonsequencing, where genome-saturated mutant populations allele frequencies aremeasured before and after selection, functionally characterizes each and every genein the genome in a single experiment. The approach has been successfully appliedto a variety of phenotypes in a variety of unicellular systems: growth and motilityin E. coli, synthetic genetic interactions in yeast, and in vitro pathogen-resistance inmammalian cell lines. However, transposon insertion typically produces null alleles,which can be valuable to identify gene function, but evolutionary insight relies onidentifcation of naturally occurring polymorphisms affecting the trait of interest.Genome-wide association studies (GWAS) can be used to study the effect of naturalgenetic variation on a trait, but they grow prohibitively expensive if the number ofindividuals to genotype and phenotype becomes large.Here I describe the application of transposon sequencing and pooled sequencingGWAS in the whole metazoan model, Caenorhabditis elegans. Transposon sequencinghas not been previously implemented in an animal model. I have sequenced a controllibrary using our method, C. elegans transposon sequencing (CeTnSeq). We haveconstructed a new Mos1 transposon mutator strain that is more convenient to usethan the existing strain and allows for extra-chromosomal insertions to be degradedby restriction digest. My preliminary results show that our method is qualitativelyeffective at identifying transposon insertion sites, but suffers from PCR duplicationerror. I propose to optimize the number of PCR cycles in the library and to includeunique molecular identifiers (UMI) in the library adaptor. I also show that therestriction digest is effective at removing extra-chromosomal array insertions fromthe library.I constructed simulation models to help design optimal Ce-TnSeq experimentswith respect to statistical power for a proposed starvation survival assay. I consideredmany parameters affecting the design, including: culture size, number of generations,expected effect size, sequencing coverage, and sample size. I show that the numberof homozygous mutant animals in the screen is a critical factor in the design ofexperiments. I also saw diminishing returns with respect to increasing sample sizeand sequencing depth. These simulations will be invaluable in designing future Ce-TnSeq experiments and identifying critical aspects of the protocol to optimize.We performed pooled sequencing (using restriction-site associated DNA sequencing)on a population of 95 wild isolates subjected to starvation. I identified strainsthat were resistant and sensitive to starvation, and we verified these results usingtraditional methods. We used our population sequencing data to perform an associationstudy of starvation survival across the 95 strains, and identified two statisticallysignificant quantitative trait loci.