Whole genome sequencing highlights genetic changes associated with laboratory domestication of C. elegans

Abstract

Defining the mutational landscape when individuals of a species grow separately and diverge over many generations can provide insights into trait evolution. A specific example of this involves studying changes associated with domestication where different lines of the same wild stock have been cultivated independently in different standard environments. Whole genome sequence comparison of such lines permits estimation of mutation rates, inference of genes' ancestral states and ancestry of existing strains, and correction of sequencing errors in genome databases. Here we study domestication of the C. elegans Bristol strain as a model, and report the genome sequence of LSJ1 (Bristol), a sibling of the standard C. elegans reference wild type N2 (Bristol). The LSJ1 and N2 lines were cultivated separately from shortly after the Bristol strain was isolated until methods to freeze C. elegans were developed. We find that during this time the two strains have accumulated 1208 genetic differences. We describe phenotypic variation between N2 and LSJ1 in the rate at which embryos develop, the rate of production of eggs, the maturity of eggs at laying, and feeding behavior, all the result of post-isolation changes. We infer the ancestral alleles in the original Bristol isolate and highlight 2038 likely sequencing errors in the original N2 reference genome sequence. Many of these changes modify genome annotation. Our study provides a starting point to further investigate genotype-phenotype association and offers insights into the process of selection as a result of laboratory domestication.

Figure 1. Laboratory strain cultivation history (a) and chromosomal distribution of N2/LSJ1 polymorphisms (b).

(a) Red lines indicate that worms were maintained in solid culture, and blue lines in liquid culture. (b) SNPs are indicated by red hatches, and small insertions and deletions by blue hatches. Changes are distributed uniformly on chrmosomes I, II and X and slightly enriched on the arms of chromosomes III, IV and V (K-S test, P values adjusted using FDR).

Figure 2. N2 and LSJ1 show life trait differences.

(a and b) Although N2 and LSJ1 have similar brood sizes (a), LSJ1 lays eggs over a longer period (b) (n = 11 for both strains). (c) LSJ1 embryos (n = 64) develop slower than N2 (n = 78) but (d) show no difference in lifespan N2 (n = 76) and LSJ1 (n = 85) (K-S test). Asterisks correspond to significances of differences from N2 under identical conditions; ** indicates P<0.001, *** indicates P<0.0001; error bars indicate s.e.m.

Figure 3. N2 and LSJ1 show behavioral variation.

(a) Variation in egg laying (n = 25 for both strains). Egg laying results were fitted to a binomial model and assessed with an ANOVA (2 degrees of freedom, F = 71.05, P = 2.2×10⁻¹⁵). Significances of differences in egg laying were assessed using a Mann-Whitney test with a posthoc Bonferroni correction. (b) Variation in aggregation and bordering behaviors (n = 116 for N2, n = 121 for LSJ1). (c) Variation in oxygen responses on food (n = 30 for both strains). (d) Variation in oxygen response off food (n = 40 for LSJ1, n = 30 for N2). In all panels, asterisks indicate significances of differences from N2 under identical conditions, while plusses indicate significance of speed alteration within strains in response to oxygen changes; ns = not significant, */+ indicates p<0.01, **/++ indicates p<0.001, and *** indicates p<0.0001; error bars indicate s.e.m.

Figure 4. Laboratory cultivation selects for solitary feeding.

The flowchart describes construction of advanced intercross lines, which yielded many more strains carrying the 215V allele of npr-1 than the npr-1 215F variant.