Robert W. Ness, Ph.D. student

Department of Ecology and Evolutionary Biology
University of Toronto
25 Willcocks Street
Toronto, Ontario, Canada
M5S 3B2
Phone: (416) 978-5603


  B. Sc. Biology

(2002) Queen’s University, Canada
Thesis title Evolution of apomixis in Antennaria parlinii (Asteraceae) : a phylogenetic investigation

  Ph.D. Botany

(Defending Dec 2010) University of Toronto, Canada
Thesis title: Consequences of self-fertilization for molecular evolution in Eichhornia paniculata. (Pontederiaceae)

Current Position

Post Doctoral Fellow, with Peter Keightley, The Institute of Evolution, School of Biological Sciences, University of Edinburgh.



I am interested in understanding the consequences of multiple transitions from outcrossing to selfing on the patterns and magnitude of molecular genetic variation in the annual aquatic Eichhornia paniculata (Pontederiaceae). Variation among populations in mating system in this species offers a valuable opportunity to compare independently derived selfing populations and their outcrossing ancestors (Husband & Barrett 1993). Specifically, within my PhD thesis I plan to address questions pertaining to: (1) population history and geographical subdivision of genetic diversity; (2) comparisons of the magnitude of variability between outcrossing and selfing populations; (3) the roles of genetic drift and selection in shaping the patterns of genetic variation in selfing and outcrossing populations.
These questions will be addressed by analyzing multilocus sequence data gathered from both coding and non-coding loci throughout the E. paniculata genome. Loci are being identified and isolated by developing an expressed sequence tag (EST) library. From the EST data I have designed PCR primers that amplify loci directly from E. paniculata genomic DNA for sequencing. In the description below I summarize some of the motivations and theory behind the three subject areas I outlined above:



Populations of E. paniculata are located primarily in northeastern Brazil, Jamaica, Cuba and Central America (Nicaragua and Mexico). While the majority of populations in Brazil are tristylous and predominately outcrossing, modification of floral traits that increase selfing rates commonly occur in dimorphic and monomorphic populations (Barrett et al. 1989). Selfing modifications, largely confined to the mid morph, involve elongation of the filaments of short-level stamens to a position adjacent to the mid-level stigma. A cross between two selfing variants from different regions of Brazil revealed the existence of at least two distinct recessive modifier genes (Fenster & Barrett 1994). Further, allozyme polymorphism data from Jamaican populations suggest multiple independent colonization events. (Husband & Barrett 1991) These results raise the question; what are the historical relationships among populations in Brazil and the Caribbean, and what is the origin of populations that occur in Mexico and Nicaragua?
Sequence polymorphism data can be exploited to elucidate relationships among populations from across the geographic range of E. paniculata. Identifying independently derived selfing lineages will allow more sophisticated analyses and comparisons between outcrossing and selfing populations in later components of my Ph.D. thesis.



Intraspecific variation in mating system allows the comparison of sequence variability without the confounding effects that are often associated with interspecific comparisons of outcrossing and selfing species. A large body of population genetic theory predicts a reduction in genetic variability within inbreeding populations as a result of three processes.

i. Selection
Increased homozygosity due to self-fertilization leads to reduced effective recombination in selfers (Charlesworth et al. 1993; Nordborg 2000). Due to the reduction in recombination experienced by selfers there is a consequent increase in linkage among loci. When loci are selected for or against the alleles associated with them are also indirectly selected, the two processes have been termed selective sweeps and background selection, respectively (Charlesworth et al. 1993). Both processes result in the overall reduction in genetic variability (Maynard Smith & Haigh 1974; Charlesworth, et al. 1993; Charlesworth et al. 1995). Therefore in inbreeding populations with high linkage disequilibrium there is predicted to be lower levels of genetic variation.

ii. Demographic
Completely selfing populations are expected to have a two-fold reduction in effective populations size (Ne.) (Charlesworth et al 1993; Nordborg 2000). This can be conceptualized by imagining a halving of the number of segregating genomes in the population. Smaller Ne strengthens the effect of genetic drift and can reduce the amount of genetic variation maintained in a population of selfers . Further, genetic drift interacts with selection to reduce the ability of purifying selection to purge deleterious mutations or directional selection to fix advantageous ones (Birky et al 1988; Charlesworth 1992; Charlesworth & Charlesworth 1997).

iii. Life history
A suite of life history characteristics commonly associated with the transition to selfing can exaggerate the reduction of Ne and genetic variability. Both increased population subdivision due to isolation and frequent population bottlenecks due to marginal habitats and extinction-recolonization dynamics lower Ne and lead to a reduction in genetic variability within a population via genetic drift

I plan to measure DNA sequence variation in 16 populations (4 trimorphic, 4 dimorphic and 8 monomorphic) including several independently derived selfing populations.



The shift from outcrossing to selfing in E. paniculata is thought to involve a process of genetic drift resulting from bottlenecks during colonization, followed by selection on selfing variants for reproductive assurance in the absence of reliable pollinator service (reviewed by Barrett 1993).
On a single locus the footprint of positive selection or a bottleneck is the same. In the short term only a small subset of polymorphism survives. As time progresses new mutations accumulate and there is an increase in the frequency of rare polymorphisms (Braverman et al. 1995). With a population bottleneck this effect is evident throughout the entire genome and is expected to affect all loci equally. However, a selective event affects only loci directly involved with coding the trait under selection and loci linked to them. These ‘selective sweeps’ can be detected as a local excess of rare polymorphisms or a negative skew in the frequency spectrum relative to the rest of the genome (Wright et al 2005, Ometto et al 2005).
As outlined above we expect there to be evidence of both a bottleneck and selection in selfing populations of E. paniculata. In collaboration with Dr. Stephen I. Wright (York University), I plan to model the effects of population bottleneck on molecular variation using an outcrossing population as a reference. We will investigate if any loci from selfing populations exhibit reductions in variability beyond that which can be explained solely by bottlenecks and drift.



  • Ness. R.W., S.I. Wright & S.C.H. Barrett. (2010). Mating-System variation, demographic history and patterns of nucleotide diversity in the tristylous plant Eichhornia paniculata. Genetics 184: 381–392. DOI: 10.1534/genetics.109.110130 (pdf)

  • Barrett. S.C.H., R.W. Ness & M. Vallejo-Marín (2009). Evolutionary pathways to self-fertilization in a tristylous plant species. New Phytologist 183: 546-556. (pdf)

  • Wright, S.I., R.W. Ness, J.P. Foxe & S.C.H. Barrett (2008). Genomic consequences of outcrossing and selfing in plants. International Journal of Plant Sciences 169:105–118. (pdf)

Eichhornia paniculata, Brazil

E. paniculata