Effect of Phenotype Selection on Genome Size Variation in Two Species of Diptera

Abstract

Genome size varies widely across organisms yet has not been found to be related to organismal complexity in eukaryotes. While there is no evidence for a relationship with complexity, there is evidence to suggest that other phenotypic characteristics, such as nucleus size and cell-cycle time, are associated with genome size, body size, and development rate. However, what is unknown is how the selection for divergent phenotypic traits may indirectly affect genome size. Drosophila melanogaster were selected for small and large body size for up to 220 generations, while Cochliomyia macellaria were selected for 32 generations for fast and slow development. Size in D. melanogaster significantly changed in terms of both cell-count and genome size in isolines, but only the cell-count changed in lines which were maintained at larger effective population sizes. Larger genome sizes only occurred in a subset of D. melanogaster isolines originated from flies selected for their large body size. Selection for development time did not change average genome size yet decreased the within-population variation in genome size with increasing generations of selection. This decrease in variation and convergence on a similar mean genome size was not in correspondence with phenotypic variation and suggests stabilizing selection on genome size in laboratory conditions.

Keywords: Cochliomyia; Drosophila; artificial selection; blow fly; body size; development time; stabilizing selection.

Figure 1

Range of development time and genome size for generations 1, 10, and 32 for different populations of the blow fly Cochliomyia macellaria. Hours of development time are plotted on the X-axis and genome size (Mbp) is plotted on the Y-axis. Points represent mean of the phenotype and lines represent the range of each trait. Colors represent development and shapes represent origin city. No change in development time by generation was seen in control lines, increases in development time was seen in slow selected lines, and decrease in development time was seen in fast selected lines. Variation in genome size reduced with generation and converged on a mean size of approximately 530 Mbp. Variation in development time increased from generation one and was maintained.

Figure 2

Pupal cases for D. melanogaster selected for large and small body size. All strains from this picture were maintained together, fed from the same batch of medium and the vials established for this image were started on the same day.

Figure 3

Boxplot comparisons of genome size and cell-counts between D. melanogaster differentially selected for body size. Genome size and cell-counts plotted by strain for outbred strains and isolines. Different letters above each box represent values significantly different according to Tukey HSD at the p < 0.05 level. (A) Genome size variation for outbred populations in control lines (C1, C2), large selected lines (L1, L2), and small selected lines (S1, S2). (B) Variation in cell-count ratio for outbred populations in control lines (C1, C2), large selected lines (L1, L2), and small selected lines (S1, S2). (C) Genome size variation for isolines developed from outbred populations for large-body size selected lines (L1, L2) and small-body size selected lines (S1, S2). (D) Variation in cell-count ratio for isolines developed from outbred populations for large-body size selected lines (L1, L2) and small-body size selected lines (S1, S2).

Figure 4

Distribution of genome sizes and relative cell-count ratio for large and small body size selected lines of D. melanogaster. Relative cell-count ratio (X-axis) plotted against genome size in Mbp (Y-axis) for large and small body size selected isolines. Large body size selected lines are represented in red circles (L1) and green triangles (L2), small body size lines represented in blue squares (S1) and purple diamonds (S2). Ellipses represent 95% confidence ellipses determined using the stat_ellipse() function in the ggplot2 package of R. Only a subset of large body size selected D. melanogaster isolines showed an increase in genome size. A two-dimensional Kolmogorov-Smirnov test using the function peacock2() in the package ‘Peacock.test’ found the distributions of large and small body size flies to be significantly different (p < 0.001) [41].