Backcrossing

[Minnesota Mato]

I do a lot of crosses with wild tomatoes and it takes me a long time and a lot plants to recover size. I have been reading about back crosses and how it is best to due two back crosses minimum to recover size. Is it best to start the back cross from the f1 or select a f2 with the most recessive genes to start your back cross?
Craig

[Darren Abbey]

Backcrossing an F1 to the domestic parent is an effective way to transfer over dominant traits from the wildling into the domesticated line.

P1 x P2 => F1 x P1 => BC1 x P1 => BC2 x P1 => BC3 ... etc.

After several generations, you will end up with a line that is essentially the same as the chosen backcross parent, but with the selected dominant traits from the the other parent.

----
The case for transferring a recessive trait is move involved.

P1 x P2 => F1 x F1 => F2 x P1 => BC1 x BC1 => BC1:F2 x P1 => BC2 x BC2 => BC2:F2 x P1 => BC3 ... etc.

In each F2 generation, you'd need to grow out enough plants to recover the double-recessive case for the traits of interest. If you don't, then you can easily lose the recessive trait at any stage in the backcross series.

The recessive case takes roughly twice the number of generations as the dominant case, because of the need to check all the time that you haven't lost the recessive traits. If you're lucky, you can identify the "recessive" trait by some subtle effect in the heterozygous condition and then the process will be like the dominant case.

[Darren Abbey]

So, in short, I would advise going from the F2 with the most recessive traits of interest... and then doing the same thing every two generations.

[Fusion_power]

Selecting in the F2, then doing the cross with the best F2 plant(s) is usually the best choice. There is a factor of how much chromosome fragmentation occurs from cross-over. Doing a backcross using an F2 allows for more effective selection as the fragments become smaller.

With modern DNA tools to assist selection, it is possible to dramatically reduce the number of generations to stabilize a given set of genes. As the cost of a dna profile goes down, we should see tools available that backyard breeders have not previously had access to. I'm still waiting for that $35 tomato DNA test. We may not see it for a few more years, but we will eventually see it.

[Minnesota Mato]

My primary project is crossing a large green tomato with a bright red pimpinellifolium to get a large black tomato. So I have to pick thru red, yellow, green and black. I hope I can cut down the number of plants because with my other crosses I have 200 plants already and things are getting out of hand. Lucky for me my wife likes salsa, ketchup, marinara ect. so I'm not in too much trouble yet.

[Darren Abbey]

green-fleshded (gf), yellow-flesh (r), and fasciated (fas) are all recessive.

P1 (gf/gf; r/r; fas/fas) x P2 (Gf/Gf; R/R; Fas/Fas) => F1 (Gf/gf; R/r; Fas/fas)

----
Examining the the F2s... F1 x F1 => 25% gf/gf; 25% R/R; 25% fas/fas

You'd need to grow at least 64 to expect to find the combination you're looking for.

----
Backcrossing to either parent...

BC1:: F1 (Gf/gf; R/r; Fas/fas) x gf/gf; r/r; fas/fas [beefsteak] => [50% Gf/gf, 50% gf/gf]; [50% R/r, 50% r/r]; [50% Fas/fas, 50% fas/fas]

BC2:: F1 (Gf/gf; R/r; Fas/fas) x Gf/Gf; R/R; Fas/Fas [pimpinellifolium] => [50% Gf/Gf, 50% Gf/gf]; [50% R/R, 50% R/r]; [50% Fas/Fas, 50% Fas/fas]

Will fail to get you the combination you're looking for.

----
But, backcross one (BC1) will get you closer. 1/8 of the first backcross will be brown (green-flesh with red) and have large fruit. ([50% gf/gf]; [50% R/r]; [50% fas/fas])

Self such a plant and 3/4 of the next generation will be brown (green-flesh with red) and have large fruit. A third of these will have the combination of alleles (gf/gf; R/R; fas/fas) you're looking for.

I suspect there may be another recessive trait resulting in the large fruit compared to the pimpinellifolium parent, so it may be a bit more complicated. (1/16 of the backcross would have the intermediate state).

[Darren Abbey]

In short, yes, you can seriously reduce the numbers you'll have to deal with... with a cost of an additional generation. (Actually, two, as you'll need at least one to identify a fully homozygous line.)

[Darren Abbey]

Originally Posted by Fusion_power

Selecting in the F2, then doing the cross with the best F2 plant(s) is usually the best choice. There is a factor of how much chromosome fragmentation occurs from cross-over. Doing a backcross using an F2 allows for more effective selection as the fragments become smaller.

If you select from the F2s before doing the backcross, you'll be at risk of losing one of the recessive traits being looked for.

Originally Posted by Fusion_power

With modern DNA tools to assist selection, it is possible to dramatically reduce the number of generations to stabilize a given set of genes. As the cost of a dna profile goes down, we should see tools available that backyard breeders have not previously had access to. I'm still waiting for that $35 tomato DNA test. We may not see it for a few more years, but we will eventually see it.

Once we know which gene is associated with which trait, we can definitely speed things up dramatically. Yes, eventually we'll have a simple test that will tell you which important alleles are found in any given plant... but it may never be all that cheap, because there are relatively few people who will be interested in buying the product.

[Minnesota Mato]

I will be trying this plan shortly, my F1 is starting to blush, I am hoping to find what I'm looking for with 40 plants, that's the most room I have. Just for fun how many plants would I need to grow out in the F2 without any backcrosses to get (gf/gf,R/R,fas,fas)

[joseph]

Originally Posted by Minnesota Mato

I will be trying this plan shortly, my F1 is starting to blush, I am hoping to find what I'm looking for with 40 plants, that's the most room I have. Just for fun how many plants would I need to grow out in the F2 without any backcrosses to get (gf/gf,R/R,fas,fas)

Depends on the concentration of those genes in the parents of the cross.

For the worst case scenario, let's assume that one parent provides all of the desired genes, and the other parent doesn't provide any. So with two recessive genes and one dominant the odds of finding the phenotype you are looking for in the F2 are: 1/4 * 1/2 * 1/4 = 1 in 32... However half of plants that match the desired phenotype will only have one copy of the R gene, so the odds are 1:64 that an F2 plant will be homozygous for all the genes of interest. But because the assortment of the genes, and the selection of seeds is random, I figure that you'd need about 320 plants to feel really secure about winning the genetic lottery.

[crmauch]

Originally Posted by joseph

For the worst case scenario, let's assume that one parent provides all of the desired genes, and the other parent doesn't provide any. So with two recessive genes and one dominant the odds of finding the phenotype you are looking for in the F2 are: 1/4 * 1/2 * 1/4 = 1 in 32... However half of plants that match the desired phenotype will only have one copy of the R gene, so the odds are 1:64 that an F2 plant will be homozygous for all the genes of interest. But because the assortment of the genes, and the selection of seeds is random, I figure that you'd need about 320 plants to feel really secure about winning the genetic lottery.

Would you not have to grow out subsequent generations to verify that the variety is homozyqous for the dominant R gene?

I would also refer you to Carol Deppe's book (unless someone knows an online version of her chart), where for a given odd she gives how many plants to plant out for a 95% certainty and a 99% certainty of getting what you're looking for.

[carolyn137]

Originally Posted by crmauch

Would you not have to grow out subsequent generations to verify that the variety is homozyqous for the dominant R gene?

I would also refer you to Carol Deppe's book (unless someone knows an online version of her chart), where for a given odd she gives how many plants to plant out for a 95% certainty and a 99% certainty of getting what you're looking for.

I think that same information is also given at Keith Mueller's website in a very different way where he explains and gives numbers as to how many gens need to be grown out to give different percentages of purity.

http://www.kdcomm.net/~tomato/

Go to Culture, there go to tomato gene basics, read that and then click on the segregation link at bottom of that.

Carolyn

[maf]

Originally Posted by Minnesota Mato

I do a lot of crosses with wild tomatoes and it takes me a long time and a lot plants to recover size.

Dissecting the Genetic Pathway to Extreme Fruit Size in Tomato Using a Cross Between the Small-Fruited Wild Species Lycopersicon pimpinellifolium and L. esculentum var. Giant Heirloom

Minnesota Mato, I don't know if you have seen the above study, but it uses a cross between a pimpinellifolium and a giant fruited domestic tomato to identify the QTL's that contribute to fruit size. Six loci were identified that made a significant contribution to fruit weight, two of the most significant mapping to the classical genes fasciated (f) and locule number (lc). Not all six of these may be in the large fruited variety in your cross but it is likely there are at least three or four including f and lc so it is going to take some serious selection to recover the size you desire.

If it was me, I would select for size first and color second. For example, if I found a type that was huge fruited but red, I would grow that out in preference to one that was medium sized and black, in the hope that it segregated for gf in the next generation. Actually, a very good backcross would be the largest fruited red (or black) F2 and the green parent, I think that would give you a very big push in the right direction.

[maf]

Originally Posted by Minnesota Mato

I will be trying this plan shortly, my F1 is starting to blush, I am hoping to find what I'm looking for with 40 plants, that's the most room I have. Just for fun how many plants would I need to grow out in the F2 without any backcrosses to get (gf/gf,R/R,fas,fas)

Another thought, you could start 200 seeds and plant them out five times as crowded as you normally do, then when the first fruits start to form it will be very easy to identify the majority of small fruited selections and yank them all out leaving your best 20%.

[Darren Abbey]

Originally Posted by maf

Dissecting the Genetic Pathway to Extreme Fruit Size in Tomato Using a Cross Between the Small-Fruited Wild Species Lycopersicon pimpinellifolium and L. esculentum var. Giant Heirloom

Minnesota Mato, I don't know if you have seen the above study, but it uses a cross between a pimpinellifolium and a giant fruited domestic tomato to identify the QTL's that contribute to fruit size. Six loci were identified that made a significant contribution to fruit weight, two of the most significant mapping to the classical genes fasciated (f) and locule number (lc). Not all six of these may be in the large fruited variety in your cross but it is likely there are at least three or four including f and lc so it is going to take some serious selection to recover the size you desire.

If it was me, I would select for size first and color second. For example, if I found a type that was huge fruited but red, I would grow that out in preference to one that was medium sized and black, in the hope that it segregated for gf in the next generation. Actually, a very good backcross would be the largest fruited red (or black) F2 and the green parent, I think that would give you a very big push in the right direction.

I knew that the f and lc traits were important in fruit size, as well as an unnamed gene that results in larger cells (and therefor larger seeds and fruit), but I didn't realize there were likely to be three others as well.

The more genes contribute to the large size you're looking for, the more plants (or years) you'll have to work on it... no matter if you can use backcrossing to reduce the numbers you'll need.

I'll set about checking the numbers for six traits, assuming all are recessive(?), and we shall see what sort of task you're engaging in.