The Conner lab studies the mechanisms by which natural selection on plants produces (sometimes very rapid) adaptation to a variable environment, as well as possible constraints on this adaptation. We measure the strength of selection acting in present-day populations and combine this with quantitative and molecular genetic and genomic analyses to predict short-term evolutionary change and identify the genetic mechanisms underlying adaptation and constraint. Major projects focus on floral evolution, weed adaptations to agricultural habitats, and fitness effects of duplicate genes.  Approaches we use include genomics, QTL mapping, recombinant inbred lines and nearly-isogenic lines (RILs and NILs), gene knockouts, field studies of fitness and natural selection, and ecological studies to determine the agents of selection.

All of our research areas include multiple potential thesis projects for graduate students as well as undergraduate research:

Stamen evolution in the mustard family (Brassicaceae)

  1. Mechanisms of adaptation in anther exsertion in wild radish: Past work in the lab has focused on adaptation and genetics of long-stamen anther exsertion in wild radish (see Publications).  We are currently using genotyping by sequencing (GBS/Rad-Seq) approaches to perform QTL mapping to understand the genetic mechanisms underlying rapid evolution of this trait in response to our artificial selection (Conner et al. 2011).  The ultimate goal is to find genes causing natural variation in anther exsertion, and track changes in allele frequencies of these genes in the artificial selection lines after they are returned to the field environment.
  2. Adaptation and constraint in a conserved trait: evolution of tetradynamy across mating systems: The mustard family Brassicaceae is large (over 3700 species) and includes many important crops (e.g., canola, mustard, broccoli, cabbage, radish), invasives (e.g., garlic mustard), serious agricultural weeds, especially wild radish (see below), and the model plant Arabidopsis thaliana.  A conserved trait that is diagnostic for the family is the presence of four long and two short stamens within each flower, called tetradynamy (Greek for characterized by four and two).  We have done a number of experimental and observational studies to understand the function, genetics, and selection on this trait in the obligately outcrossing wild radish.  The weight of the evidence at this point suggests that selection is likely more important than genetic constraint in the conservation of the trait, but we are continuing this work.

    We are also pursuing complementary studies of tetradyamy in the highly selfing model plant Arabidopsis thaliana.  My previous student Anne Royer showed that the short stamens do not significantly increase selfed seed set; thus, the transition to selfing, which is one of the most common evolutionary changes in flowering plants, seems to have eliminated the function of the short stamens.  From this we predict that short stamens should be lost in this species; consistent with this, Anne found widespread but incomplete loss showing a latitudinal cline, with stamen loss more common in the south.  In parallel, my current student Sam Perez has found an altitudinal cline in stamen loss, with more loss at lower elevations.  Using recombinant inbred lines (RILs) from Doug Schemske’s lab, Anne also found three quantitative trait loci (QTL) for stamen loss that interact epistatically to constrain loss; using an independent set of lines Sam has found four QTL, two that are in the same locations as those in Anne’s study.  By searching for parallel changes in the parental DNA sequences of these two RIL sets, Sam has identified two strong candidate genes at the shared major-effect QTL.

    Short stamen loss in A. thaliana is an excellent model system for improving our understanding of evolutionary trait loss. The trait loss seems to be ongoing, as it is incomplete in all individuals and population studied to date.  Stamen loss is easy to score and manipulate, large numbers of experimental plants can be placed in the field and lifetime total fitness (selfed seed set) estimated, and A. thaliana is a genetic model system with abundant resources.  Current and future work has the aims of finding the genes underlying stamen loss, quantifying the fitness effects of loss at high and low altitudes and latitudes, and understanding constraints on stamen loss.  Approaches include near isogenic lines (NILs), association and fine mapping, use of T-insertion knockout lines, whole-genome and GBS sequencing, and artificial selection on stamen loss.

Radish weed and crop evolution

An estimated 40% of the earth’s land surface has been converted by humans to agriculture, making this one of the largest sources of global environmental change.  This massive change is both a challenge and an opportunity for organisms; an opportunity, if they can adapt to this new agricultural habitat.  Those that do can become agricultural pests, which cause massive economic losses worldwide. Agricultural weeds cause 35 billion in economic losses in the US alone, but the mechanisms of adaptation of weeds to farmer’s fields is virtually unknown.
We work on adaptations to agriculture in one of the world’s worst weeds, wild radish; this is a closely related species to the garden radish.  Compared to the native European populations of wild radish, weedy radish has evolved a very rapid life cycle, flowering and producing seeds at the same time or even ahead of the crop, without producing a rosette, which is an adaptation for overwintering.  The two major groups of crop radish, the round red European radish and the large white Asian daikon have undergone parallel phenotypic evolution, with the European radish flowering quickly without a rosette like the weeds, and the daikon having delayed flowering with a rosette like the native wild plants.Our goal is to uncover the parallel phenotypic and genetic changes that occurred during the rapid evolution of the pest weed and agricultural crop from their wild ancestors.  Current thinking is that both the weed and crop evolved from the same wild species, but our work aims to test this rigorously.  A better general understanding of crop and weed adaptation should benefit agriculture.  More generally, this project should give us a better basic understanding of how organisms can rapidly evolve to adapt to global environmental change.

Fitness effects of duplicate gene knockouts

Gene duplication is ubiquitous in nature, and whole-genome duplication (polyploidy) is especially common in plants.  Thus, many duplicate pairs of genes exist in most genomes, but the effects of these copies on fitness in the field is poorly understood. The overall goal of this project is to use the model plant Arabidopsis thaliana as a model to measure how knocking out one or both duplicate copies of a gene impacts fitness. The role of the Conner lab in this collaborative project is to estimate lifetime total selfed fitness of single knockout (where each copy is rendered nonfunctional on its own), double knockout (both copies nonfunctional), and wild type (both copies functional) plants in the field for 240 pairs of A. thaliana duplicate genes. Fitness effects of gene knockouts have rarely been assessed in the field, and lifetime seed set in a highly selfing annual like A. thaliana seed set includes all female and 95% of male fitness; thus, these are among the best field fitness estimates ever produced.  To date, we have planted 115 sets of single and double knockout lines of duplicate gene pairs in the field, with each line replicated 44-50 times for a total of almost 22,000 plants. One planting of 27 sets was repeated in the spring and fall to test for genotype-environment interactions, because A. thaliana populations can be either winter or spring annuals. Preliminary results show large variation in the degree of redundancy among duplicate pairs, that is, where one copy can partially or completely compensate for the effects of knocking out the other, and also large variation in epistasis, where the fitness effects of the double mutant are different from the additive effects of each single knockout.