The whole story
Genes rarely change their expression alone. A regulator at work in many processes is constrained by all of them, so species differ in wiring more often than in genes. Expression levels cannot be compared directly between species, but co-expression (which genes rise and fall together) is measured within each. I try to align co-expression networks the way we align sequences. The unit is the co-expressologue, an orthologue that keeps its neighbourhood across species. Most neighbourhoods are conserved, yet the orthologue closest in sequence is often not the one that kept the job.
Crossing from tropical to temperate biomes is rare in plant evolution, as it is often easier to move to a familiar climate than to evolve for a new one. Grasses that made it had to fit flowering into a short season, and flowering is gene expression with a calendar. Vernalisation (a cold spell a plant needs before it can flower) and daylength responses both evolved repeatedly in grasses. My PhD traced which gene copies took on the job, from FRUITFULL-like genes through weeks of cold to CONSTANS-like genes across a single day.
There are two ways through a hard winter. The most elegant is to avoid it, and annual grasses spend it as seed. Perennials procrastinate. They keep some shoots from flowering until next year. Annuals evolved from perennials at least 48 times in temperate grasses, mostly in lineages that already invested more above ground. Each origin asks whether evolution reused the same genes. Comparing annual barley with its perennial relative foxtail barley (Hordeum jubatum) should show what in gene regulation makes an annual annual, and help breed perennial cereals.
Comparing genes by their activity means comparing shapes. Within a species, a gene’s activity through a day is a curve, not a handful of separate time points, and functional data analysis treats it as a whole. The main patterns among such curves, their principal components, capture when and how sharply a gene peaks and sort thousands of rhythms into a few types. Over a season, Bayesian non-linear growth models describe how plants develop, and over evolutionary time, phylogenetic analysis of gene families shows how the genes themselves evolved, duplications included.
Macroevolutionary processes rarely involve shifts from tropical to temperate biomes. This scarcity points to climate as a strong filter, which lets only lineages with specialised adaptations leap out of their ancestral ranges. Flowering at the right time is central to such a leap. Yet most of what we know about how flowering responds to cold and daylength comes from domesticated species. That knowledge need not hold for the adaptations that carried their untamed relatives into ever harsher habitats.
My favourite plants are grasses, which feed most of the world and, by one popular theory, are the reason we walk on only half of our limbs. They began in tropical forests, yet most northern grass species belong to one subfamily, the temperate grasses (Pooideae). Getting north meant surviving frost and waiting out winter before flowering. Evolution solved both problems more than once. It often rewired the regulation of genes the grasses already had and sometimes changed the genes themselves. Given the same problem again, does evolution find the same solution? This is exactly what I study in wild grasses.
How long a grass lives is another answer to winter. Some live fast and die right after flowering; others are long-lived and overwinter repeatedly. Perennial life is the rule among flowering plants, yet the temperate grasses have switched to annual life at least 48 times. For flowering and life history alike, I follow single genes through a day or a season and compare co-expression networks between species. Wheat, barley, and oats are grasses too, so what I learn helps predict how crops and their wild relatives will cope with a changing climate.
Sidequests have taken me to the genome sizes of brown seaweeds, faba bean metabolomics, the pollen chemistry of the Svalbard flora, and the taxonomy of green microalgae. The last has grown up into Untangling the knots, a project funded by Artsdatabanken.