Weak response to selection on stigma-anther distance in a primarily selfing population of yellow monkeyflower
Data files
May 14, 2024 version files 18.73 MB
Abstract
Stebbins hypothesized that selfing lineages are evolutionary dead ends because they lack adaptive potential. While selfing populations often harbor limited nucleotide variability compared to closely related outcrossers, reductions in the genetic variability of quantitative characters remain unclear, especially for key traits determining selfing rates. Yellow monkeyflower (Mimulus guttatus) populations generally outcross and maintain extensive quantitative genetic variation in floral traits. Here, we study the Joy Road population (Bodega Bay, California, USA) of M. guttatus, where individuals exhibit stigma-anther distances typical of primarily selfing monkeyflowers. We show that this population is closely related to nearby conspecifics on the Pacific coast with a modest 33% reduction in genome-wide variation compared to a more highly outcrossing population. A five-generation artificial selection experiment challenged the hypothesis that the Joy Road population harbors comparatively low evolutionary potential in the stigma-anther distance, a critical determinant of selfing rate in Mimulus. Artificial selection generated a weak phenotypic response, with low realized heritabilities (0.020-0.028) falling 84% below those measured for floral characters in more highly outcrossing M. guttatus. These results demonstrate substantial declines in evolutionary potential with a transition toward selfing. Whether these findings explain infrequent reversals to outcrossing or general limits on adaptation in selfers requires further investigation.
README: Weak response to selection on stigma-anther distance in a primarily selfing population of yellow monkeyflower
To recreate our results, use the following scripts and associated data files. In all data files, missing values are 'NA'.
1. GATK script.txt: This file contains the arguments we used when calling variants in GATK, which generated our variant call file (VCF). We processed the resulting VCF, using the steps outlined in the paper's methods, to generate our clean file of synonymous SNPs: 'platy_syn_snps.csv'.
In this file, the rows are positions in the reference genome, and the columns are: Chromosome, position, reference base pair, alternate base pair, J2 genotype, J33 genotype, J58 genotype, J60 genotype, annotation, and putative impact. Each genotype cell corresponds to reference homozyote (0), heterozygote (1), and alternate homozygote (2). The last two columns are annotations added from SnpEff.
2. 'platy_polymorphism.R' is a script that uses 'platy_syn_snps.csv' to estimate nucleotide diversity and Tajima's D.
3. To make Figure 3, use the files in the 'Fig3' folder. 'G1 platycalyx' contains floral trait data (in mm) measured on individuals from the G1 generation in M. platylcalyx. 'G1 rabbit ck' contains data on the same floral traits measured on individuals from the G1 generation of M. guttatus. 'g1means' are means calculated for each M. platycalyx individual. The script 'compare platy_rabbitck.R' uses data from the above files to compare stigma-anther distances between the Joy Road (M. platycalyx) and Rabbit Creek (M. guttatus) populations.
4. To make Figure 4, use the files in the 'Fig4' folder. All of these files were generated by Treemix analysis on an alignment containing whole-genome sequences of several Mimulus species (13mimulus_4treemix.gz). Columns are formatted as in 'platy_syn_snps.csv'. The file 'make_tree.R' generates Figure 4. The files '13mimulus_input.gz' and 'prepTreemix.R' were used to format files so that treemix would run on our files.
5. To make Figure 5, use the files in the 'Fig5' folder. 'allgens.csv' is a summary of the mean trait value in generations 1-6 (G1-G6). These summaries are taken from the full datafile, 'allparents_parents_marked_g1g2g3g4g5g6.csv', which reports trait means for all individuals. Columns denote: individual (X), population (line), an unused indentifier based on individual (ID), days to flower (d2flower), leaf width in mm (max.leaf.W), and floral traits (corolla length in mm, corolla width in mm, tube length in mm, throat width in mm, stigma height in mm, stamen height in mm, stigma-anther distance in mm), parent (FALSE=no, TRUE=yes), and generation of the experiment. The script 'Make_Fig5.R' generates Figure 5 from the above data files.
6. To make Figure 6, use the files in the 'Fig6' folder. The file 'g7means.csv' contains mean floral traits for all individuals in the G7 populations plus the G1 ancestor raised in a common environment. The columns match those described above in 'allparents_parents_marked_g1g2g3g4g5g6.csv,' with the exception that parents are not marked. The script 'Make_Fig6.R' generates Figure 6 from the raw data file.
Methods
Data come in two forms.
- DNA sequence data was generated on an Illumina Novaseq machine. These raw sequences were aligned to a Mimulus guttatus reference, followed by haplotype calling and variant calling in GATK. The resulting sequence files are provided, and their format is described in the README file.
- Phenotypes were measured on plants across 7 generations in a greenhouse. The description of the traits found in these data files is available in the README file.