Genetic and environmental influences on the distributions of three chromosomal drive haplotypes in maize

Abstract

Meiotic drive elements are regions of the genome that are transmitted to progeny at frequencies that exceed Mendelian expectations, often to the detriment of the organism. In maize there are three prevalent chromosomal drive elements known as Abnormal chromosome 10 (Ab10), K10L2, and the B chromosome. There has been much speculation about how these drivers might interact with each other and the environment in traditional maize landraces and their teosinte ancestors. Here we used genotype-by-sequencing data to score more than 10,000 maize and teosinte lines for the presence or absence of each driver. Fewer than ~0.5% of modern inbred lines carry chromosomal drivers. In contrast, among individuals from 5331 open-pollinated landraces, 6.32% carried Ab10, 5.16% carried K10L2, and 12.28% carried at least one B chromosome. These frequencies are consistent with those reported in previous studies. Using a GWAS approach we identified unlinked loci that associate with the presence or absence of the selfish genetic elements. Many significant SNPs are positively associated with the drivers, suggesting that there may have been selection for alleles that ameliorate their negative fitness consequences. We then assessed the contributions of population structure, associated loci, and the environment on the distribution of each chromosomal driver. There was no significant relationship between any chromosomal driver and altitude, contrary to conclusions based on smaller studies. Our data suggest that the distribution of the major chromosomal drivers is primarily influenced by neutral processes and the deleterious fitness consequences of the drivers themselves. While each driver has a unique relationship to genetic background and the environment, they are largely unconstrained by either.

Fig 1. Detection of CDHs in GBS data.

A. Min/max scaled tag index for Ab10, K10L2, and the B chromosome. The CDH status of each control sample is shown on the x axis. Relevant features of each CDH are highlighted on the y axis. A scaled tag index was calculated for each 1 Mb bin; here the CDH haplotypes are scaled to their relative lengths. Ab10-1, Ab10-II, and Ab10-III are different cytological types of Ab10 that were included in our control set. There are multiple replicates of the same CDH chromosomes in each panel. B. Stepwise manner of differentiating Ab10 from K10L2 (left) and general workflow for identifying CDHs using scaled tag index bins (right). Check indicates present, x indicates absence.

Fig 2. Detection of CDHs in experimental samples.

A. Summary table for previous studies of CDH distribution as well as the results from the work presented here for comparison. Numbers in parentheses indicate the number of individuals surveyed. The data from Yamakake 1975 [20] were incorporated into the McClintock et al. 1981 [21] report. The 314 teosinte individuals assayed in this study were sampled from 79 accessions (see S1 Table). B. Scaled tag index for all CDH positive lines and N10 controls. CDH-positive controls and CDH-positive experimental samples are intermixed: those from controls are indicated by colors in the bars below the dendrograms (that show relatedness), those with no color are experimental samples. The N10 control samples are plotted separately. C. Estimated copy number of all 3 CDHs in control and experimental samples. B Chr Pos. Control refers to samples that are B chromosome positive via PCR (copy number was unknown). Pos. Exp. refers to all CDH positive experimental samples identified in public GBS data. Dotted lines show approximate one copy increments as determined by the relationship between 1 and 2 copies in control samples. The values on the y axis (CDH/core gene relative tag index) are not integers because much of the CDH sequence is not single copy and poorly conserved, resulting in relatively fewer tags aligning to the reference.

Fig 3. Maps and PCA plots of maize landraces assayed as CDH negative or positive.

For each CDH, a location map is shown (left) and a plot of the first 2 principal components of population structure based on whole genome SNP profiling (right). Each grey dot indicates the location of a landrace that was assayed. Those with a CDH are highlighted in colors. The numbers in parentheses indicate the percent of variation that principal coordinates account for. Base maps from https://datacatalog.worldbank.org/search/dataset/0038272/World-Bank-Official-Boundaries were annotated using QGIS (http://qgis.org).

Fig 4. Relationship of CDH to the environment and genetic loci.

A. Manhattan plots of SNPs that associate with each CDH and B chromosome copy number. Dotted black line indicates a p value of 5x10^-8. The Ab10-associated SNP with very high p value (-(log)p close to 50) on chromosome 9 may be an artifact of read alignment to the Ab10 haplotype itself. B. Plots of fully simplified generalized linear models for each CDH including population structure, genetic loci, and environmental variables. Numbers beginning with PC are principal components of population structure. Color and orientation of the triangle indicates whether a SNP is positively or negatively associated with the CDH. The size of the triangle represents the effect size. Very Severely and Severely Limiting Soil Salt refers to growth limiting excess soil salts. Moderately Limiting Rooting Cond. refers to moderately growth limiting rooting conditions. The direction of the relationship between Ab10 and Chr3 SNP1 and Chr9 SNP 1 changes when considering each SNP in isolation (A) or with the other SNPs and environmental variables (B). C. Partitioned deviance of each model shown in B. The partitions do not sum to the full model due to shared variation among the partitions.