The Effect of Common Inversion Polymorphisms In(2L)t and In(3R)Mo on Patterns of Transcriptional Variation in Drosophila melanogaster

Abstract

Chromosomal inversions are a ubiquitous feature of genetic variation. Theoretical models describe several mechanisms by which inversions can drive adaptation and be maintained as polymorphisms. While inversions have been shown previously to be under selection, or contain genetic variation under selection, the specific phenotypic consequences of inversions leading to their maintenance remain unclear. Here we use genomic sequence and expression data from the Drosophila Genetic Reference Panel (DGRP) to explore the effects of two cosmopolitan inversions, In(2L)t and In(3R)Mo, on patterns of transcriptional variation. We demonstrate that each inversion has a significant effect on transcript abundance for hundreds of genes across the genome. Inversion-affected loci (IAL) appear both within inversions as well as on unlinked chromosomes. Importantly, IAL do not appear to be influenced by the previously reported genome-wide expression correlation structure. We found that five genes involved with sterol uptake, four of which are Niemann-Pick Type 2 orthologs, are upregulated in flies with In(3R)Mo but do not have SNPs in linkage disequilibrium (LD) with the inversion. We speculate that this upregulation is driven by genetic variation in mod(mdg4) that is in LD with In(3R)Mo We find that there is little evidence for a regional or position effect of inversions on gene expression at the chromosomal level, but do find evidence for the distal breakpoint of In(3R)Mo interrupting one gene and possibly disassociating the two flanking genes from regulatory elements.

Keywords: DGRP; DPGP; Drosophila; eQTLs; inversion polymorphism.

Figure 1

Manhattan plot of -log transformed q-values for In(2L)t (top) and In(3R)Mo (bottom) inversion effect on transcription for each probe set across the genome. The blue bar above the points indicates the genomic location of the inversion. The red horizontal line represents the genome-wide significance threshold of q=0.05.

Figure 2

Linkage disequilibrium (LD) between single nucleotide polymorphisms (SNPs) and inversion state. Disequilibrium is measured as r². Inversion breakpoints are depicted as red vertical lines. SNPs with at least 10% minor allele frequency, for inversion and SNP, and that are in significant LD with the inversion are shown in light blue (P < 0.05, Bonferroni correction for all SNPs detected by chromosome arm).

Figure 3

Cohen’s d of inversion effect for probe sets by location. Cohen’s d for In(2L)t effect across chromosome 2L (A) and for In(3R)Mo effect across chromosome 3R (D). Loess curves (polynomial line) and 95% C.I.s (shaded area) for 200 kb surrounding each breakpoint: In(2L)t (B) proximal and (C) distal, and In(3R)Mo proximal (E) and distal (F). Inversion breakpoints are depicted as vertical lines. Loess curves and C.I.s generated by (geom_smooth) function in R package (ggplot2). Positive values of Cohen’s d represent increased transcript levels associated with the inverted chromosome state and negative values represent decreased transcript levels.

Figure 4

Location of Npc2s and mod(mdg4) on chromosome 3R. Long vertical lines represent breakpoints of In(3R)Mo. Only mod(mdg4) contains a single nucleotide polymorphism in linkage disequilibrium with In(3R)Mo.

Figure 5

Correlation of single nucleotide polymorphism (SNPs) with In(3R)Mo by gene. All SNPs within gene regions and within 5 kb up- and downstream of each gene: Npc2b, Npc2c, Npc2f, Npc2g, and mod(mdg4). Horizontal bars on each plot represent the gene region for that gene.