Every time I inspect a seedling from one of my crosses for the first time I need to answer one simple question:
Is it actually a hybrid?
More specifically, is the pollen that I applied to the flower actually the pollen that produced the seedling?
Sometimes the physical appearance of some trait or another reveals the pollen parent but often times, especially in crosses involving species, another generation of crosses and back-crosses is needed to prove hybridity. Even at that, it may not be possible to actually be certain - we still don’t know, for instance, whether Floradora is actually part roxburghii.
I can bias my crosses to improve my ability to distinguish hybrids by, say, avoiding the use of species as seed parents which works pretty well sometimes in the F1 generation but starts to become less effective in subsequent generations.
I also think that this little trick comes at a big cost because (am I correct about this?) it dictates that the mitochondria and chloroplasts of my seedlings can only come from the mother plant. Since these two organelles are intimately involved with energy metabolism I worry that I’m eliminating a lot of potential benefit (health and vigor) that might come from assimilating the species versions over the domesticated versions.
I would like to know (early on) more than just whether a seedling is a hybrid - does it have the genes for particular characteristics? Most of the time I’m interesting in knowing about basic things like petal count, remontancy, pigmentation, dwarfness and so forth. Wait and see will only give me the answer some of the time, as for dominant traits or for homologous recessives.
So what I need are (1) a way to tell who papa is for sure and (2) a way to test for a set of simple traits. This would shave years off of the hybridization process for me and I can use every year I can get.
Oh, one more thing - the technology has to be really easy to use and really inexpensive to buy.
Kim, I have software that predicts coat colour genotypes for my rabbits called iBreed It’s a super simple program in that it’s just an MSAccess database facilitated by a frontend software application. I’m only looking at about 20 genes though which are based mostly on visible characteristics (which I can do in my head), but it does provide an excellent means of keeping track of genes and predicting genotype possibilities from particular crosses especially when polygenes start affecting the expression of colours so that I start doubting the genotypes.
I think the answer to the first question can be mostly solved by changing pollinating techniques. I find that because I do a lot of breeding with roses like hybrid musks, hybrid multifloras and rugosa etc that their pollen is released very early and it is difficult to use them as seed parents because of this early dehiscence. I love using them as seed parents because their seeds are produced in large quantities and they invariably germinate very easily. I get around this by emmasculating flowers a lot earlier than I normally would, sometimes even before I can see any colour in the bud, and then I’ll pop a little organza drawstring bag over the bud and leave it for a day or two or I will pollinate it straight away, or repeatedly over a few days, on the understanding that it will only take a few days to become fully receptive and that the pollen will retain reasonable viability for that time. Emmsculating doesn’t seem to affect the maturing process of stigma that much and there always seems to be enough successful hips to go forward with. Paul once told us on here that he’s found leaving at least some petals on the flower to be pollinated seems to improve hip-take and I have been using this strategy since then as well with some good results.
On the second question… I think it would be useful to try and establish definate inheritance patterns for roses so that knowing whether a rose has certain traits/genes becomes more intuitive. I just don’t think we know enough about what genes roses have and which ones are responsible for the things we are trying to improve. By looking at our breeding results and by looking at HMF can’t we do some genetoc counselling for our roses to determine how genes are passed on and therefore what genes are present and be able to keep track of them more easily.
Don, things are moving in the right direction for you. Prices are rapidly falling on massive sequencing projects. Soon you should be able to get a draft sequence of your favorite species. Then pick say 10 QTLs for each linkage group and put them on a chip rigged so that when matching DNA binds, it lights up something or another. That way you could extract a little DNA from each seedling, shear it, and verify that it carries the various QTLs that you’re interested in having in offspring.
I’m not so sure about the plastid/mitochondrion argument. Mitochondria evolve so very slowly that there may be very little difference between even widely separate species. Plastids are a different thing and might have real differences. For sure corn breeders and wheat breeders recognized decades ago that the cytoplasmic genomes affect overall performance in hybrids. And either wheat or maize is a rather narrow, single species genetic base. So it’s possible you could have a big species contribution through the cytoplasm.
My observation of a few limited crosses is that the direction of cross makes little detectable difference in offspring. Certainly nothing like some claims by Pernet-Ducher and others that one parent controls flowers and the other the plant architecture for instance. But it is clear that with species in the mix there is huge linkage disequilibrium. So we get things that look like the species pollen donor even a couple/three generations on when you’d expect it competely diluted out into a pool of standard HTs. That’s what Pernet-Ducher struggled with trying to bring in the foetida colors separate from the plant type (including B.S.)
The technology can probably be made relatively easy to use. But I don’t know about inexpensive. A lot of labor goes in at some step and likely someone wants to reap some reward.
At one time an approximate minimum number of samples required to have an idea of what a given cross had to offer was floating around. I do not remember the number but it was large enough so that only the big firms like J and P could possibly consider using that approach. I think that the concesus then was that the amateur hybridizer should aim on adjusting one trait for one market. It appears the the sucess of the Knock Roses blows that theory out of the water.
Larry there is also an argument suggesting that BS was bought into modern day rose by introducing Damask lines, have you heard anything on this or anybody?. When I first started out rose breeding I thought it may be nice to bring in some of those old traites of OGR. I soon learnt that when using Gallica’s , there was an increase in PM, and with Bourbon’s BS. The only OGR I use now are Tea’s and China’s, and things are looking dandy.
It’s a super simple program in that it’s just an MSAccess database facilitated by a frontend software application. I’m only looking at about 20 genes though which are based mostly on visible characteristics (which I can do in my head), but it does provide an excellent means of keeping track of genes and predicting genotype possibilities from particular crosses especially when polygenes start affecting the expression of colours so that I start doubting the genotypes.