The conventional process for breeding a new variety of wheat starts with setting out breeding objectives, selecting initial germplasm, developing lines that breed true with the desired traits, field testing to assess performance in real-world conditions, and then selecting lines to grow out and release as new varieties. Here is how I currently see this playing out for my perennial wheat project…
- Lifespan of at least 5 flowerings
- High genetic diversity
- Acceptable breadmaking quality (measured by a SDS sedimentation test)
- High content of nutrients and micronutrients
- Horizontal disease resistance
- Low lodging
- High threshability
- High nutrient yield
- High weight yield
- High biomass / forage yield
- Large seed size
That is a lot of traits to work on – the more traits that are selected for, the more demanding the process becomes – it either needs more power, or progress slows significantly. I’ll be using speed breeding (more on that later) and a selection pipeline operating within each generation. I also be prioritizing these traits, so that I can be more assured of progress in the most important ones.
I have several samples of Ezeer perennial wheat developed by Tim Peters, a variety of perennial wheat obtained from Caleb Warnock of unknown descent, several species of wheatgrass, a couple heirloom varieties of wheat (Triticum aestivum) einkorn wheat (Triticum monoccocum), and a shipment coming of USU perigee variety wheat.
Most of my efforts initially will be in development of the existing lines of perennial wheat that I have. The USU perigee wheat was originally developed for use growing food for astronauts in outer space: it matures very quickly, flowering fewer than 30 days after germination, and it grows very short, about 16 inches tall, and is well-adapted to 24 hour light and hydroponic setups.
Those traits are *not* what I want in a field variety of wheat. But I’m not at the field stage yet, I am at the stage where I just want something that behaves as much like annual wheat as possible, but is long lived. There is a lot of selection that is needed. Given that I have very limited space available, and am planning to operate a speed breeding process with quick generation times and 24 hour light anyway, if I can incorporate these traits into a perennial line of wheat then it could significantly accelerate the process.
Approach to line development
I will use speed breeding, developed by the Hickey lab at the University of Queensland, to grow up to 6 generations a year, with multiple consecutive screenings during each individual generation. See their video to get an idea of what this is like – in one generation, they can screen 40,000 seeds into 100 desired plants.
This will give rapid turnaround, and allow me to screen very large numbers of seeds and seedlings. I am hoping that this additional power will allow reasonably timely progress on the project despite the large number of traits.
Another method of speeding up the line development process is to use doubled haploids, which in one generation create lines that breed true to seed. In the case of domesticated and perennial wheat hybrids, however, at least for early line development, there is a *ton* of selective pressure that needs to be applied. The odds of getting a desirable line out of a doubled haploid early in the process is very low, and I do not have the resources to create hundreds or thousands of doubled haploid lines. Once I’ve done early work and some promising lines start to emerge, I may make some doubled haploids to lock down set of true-to-seed lines.
High diversity is an important breeding goal, though, so that whatever I come up with can be more quickly adapted to different regions and growing conditions. If I use doubled haploids, my hope will be to get a number of acceptable doubled haploid lines and then release them as a cultivar mix.
Given that increased lifespan is one of the breeding goals, and in fact a lifespan of greater than one flowering cycle is the very central breeding goal, I will need to have assessment of lifespan in the breeding cycle. I anticipate having three channels of development running, with intercrosses between them:
- Speed breeding through single generations for rapid selection of desirable traits, with up to 6 generations a year.
- Lifespan selection, with speed breeding growing conditions but without culling the plants. More thoughts on this below.
- I will also be growing some of the plants the old fashioned way, in the ground, letting them flowering once a year. And I’ll count the number of years that they live, and consider each individual to have a lifespan equal to the number of years it lived. Crazy, right?
So, about lifespan selection. I’m going to try doing lifespan selection under speed breeding conditions. This could give 1 generation with 6 flowering periods per year (or more, if the regrowth to flower again is quicker than growth from seed). There are a number of uncertainties about this. I don’t know if it is even possible to force the plants to flower so frequently. I don’t know if lasting through multiple flowering cycles under those conditions actually indicates the ability to go through multiple years of yielding in the field. I don’t know if this will encourage selection of undesirable traits, like loss of cold hardiness.
Anyway. The quickest way to do this, if it works, would be to cut the plants back as soon as the process of dieback begins, also known as senescence. In annual plants, they should just die because as I understand it once senescence begins is irreversible and terminal, while perennial plants just have leaves die back and then should either go dormant or start sending up more tillers.
If cutting back immediately after senescence doesn’t work for some reason, I may have to let them go further through the process of ripening seed before cutting them back. If a period of cold that triggers dormancy (vernalization) is required, that would also slow the process, and if that is the case I’ll try to hack it to go faster. Again, that would introduce more unknowns that would need to be tested in the field.
Approach to basic field testing
It is not clear when I’ll need to do field testing. I have a few lines of perennial wheat, and I need to grow them out to get a rough idea of what stage of development they are at. It is possible that one or more of them may be ready enough to go to field testing. It is likelier, though, that they will all require more work first.
In any case, the first priority will be performing a distributed, amateur field test. Given that I do not have a research facility, this is probably the most realistic option. It also carries the benefits of starting to develop partnerships with gardeners and famers, serves to begin the process of disseminating experience with and germplasm for the perennial lines. I’m not overly concerned with unfinished lines being available to the public – it lets other people perform their own breeding work if they desire. And if I come out with something better in the future, if it serves their needs better than they could switch to it.
At this point the purpose would be simply to get a handle on if the grain performs well for people or not. I plan to measure some of the traits, but not in a way that is overly burdensome, complicated, or precise. Ballpark is fine.
Approach to rigorous field testing
I haven’t researched what is involved in this yet. I’ll cross that bridge if it seems worth it to do when the time comes.
Approach to release
I imagine the first targets for a more developed variety would be homesteaders and small market farmers. From there, the fantasy plan includes a global switch from annual to perennial on all 500+ million acres that are currently planted in wheat, resulting in the end of climate change, topsoil loss, malnutrition, hunger, and war.*
* Release and its ensuing impacts are at least one or two years in the future, so I’m not planning out the details too much yet. ; )
There are a few really crazy things I might try, like breeding perennials directly from annuals without crossing with perennials. These ideas have very, very little chance of success (and may in fact be ridiculous). But I might try them anyway – if you’re curious, here are the moonshot ideas.
All errors and inadequacies are my own. Credit and thanks to Nicolas Moreau, Klaus Brugger, (and anyone else who’s talked about this with me) for their input!