Research Report: Wheat Breeding Project
By David Van Sanford
2019 Results
Purification and Increase: Working with Clements Ag Supply and Peterson Brothers, we grew a .88 acre increase of our most advanced wheat breeding line that will likely be the next Pembroke variety: KY07C-1145-94-12-5. The increase, grown near Springfield, KY, produced > 100 bu. This was an exciting result in that it represents the first time we used the Yuma, AZ increase to produce genetically pure breeder seed from headrows; Pembroke 2016 came from a Yuma increase but not at the headrow level and it was too expensive to be sustainable. Now we know that we need to increase the number of headrows grown in Yuma to produce sufficient seed so that we need just one farmer level increase before the variety is released to growers. In short, this year’s increase allowed us to determine how many headrows to grow in Yuma so the system can be sustainable. Seed from Yuma headrows that were selected this year will be put into smaller increases of lines that are not as far along in testing but still show great promise.
Yield testing: Yields across our testing locations varied considerably this year, based mostly on time of harvest. Locations harvested early before the rains hit – Schochoh (97 bu/a) and Princeton (86 bu/a) – had excellent yields, while the two later harvested locations, Lexington (53 bu/a) and Woodford (74 bu/a), showed the impact of the rain on yield and test weight. The grain filling period was not shortened by a rapid increase in temperature, as was the case in 2018. Disease pressure was low, with little scab or stripe rust but lots of Septoria leaf blotch at Schochoh. We had a number of lines early in the testing phase that combined outstanding yield with excellent disease resistance; initial increases of these lines will be planted this fall.
Crossing: We produced a total of 475 crosses in the greenhouse this season, both single cross hybrids between 2 parental lines and 3-way crosses where we cross the hybrid with a third parent. The single cross hybrids will be planted in the greenhouse this fall for an additional round of crossing, while we will plant the 3-parent F1 hybrids in the field at Lexington this fall to produce F2 seed. One set of crosses was made on the basis of genomic and field performance information about the parents, using software that tells us which are the cross combinations that are most likely to produce high performing progeny. The majority of our crosses, though, are made on the basis of field performance combined with information about which disease resistance genes are present in which parents.
Line development: We grew F4 and F5 headrows at Lexington in 2019; vigor and yield levels of the headrows were outstanding. About 1500 rows were selected based on height, maturity, clean leaves and heads, and overall vigor. Many of these lines will go into a Preliminary trial to be planted at Princeton and Lexington this fall. We also had some very high yielding scab resistant lines in our Advanced, Max and Supermax tests; these have been advanced to the next stages in the testing system.
Scab screening: Disease pressure was high in our scab nursery – the susceptible check variety was destroyed, and yet overall scab level in our breeding lines was lower than it has been in recent years. This could be a sign that the resistance we have been working on for years is more and more widespread in our material, and our lines are just more resistant than they were even several years ago. DON data from 2019 is not yet available.
Doubled Haploids and Speed Breeding: We have been excited about some of the doubled haploid lines in our breeding program in that they exceed our requirements for height, maturity, and disease resistance; a number of them in this year’s Preliminary trial made the cut for additional testing. Doubled haploids save time by shortening the breeding process, but are very expensive. We are exploring speed breeding as an alternative that provides up to 3 generations per year which will still shave time off of the breeding process but at a much lower cost, so we can use it with many more crosses.
Genomic Selection: In 2019 we were able to use genomic selection to choose lines on the basis of genomic predictions and in doing so, were able to recover roughly 7 of the top 10 actual yielders in a given population. With this kind of data we are moving towards the goal of using genomic data to decide which lines to test in the field, which will save money and allow us to look at more potential lines because we will avoid testing lines that have no chance of performing well. To date this has been somewhat effective, but not perfect. The data from 2019 has prompted us to take some of our populations going into the Preliminary test and only test half of the lines, using genomic data to predict the other half. We are hopeful this will move us one step closer to our goal.
View the full report with tables
2018
Crossing: We produced a total of 550 crosses in the greenhouse this season. We made both single cross hybrids between 2 parental lines (138) and 3-way crosses (412) where the hybrid is crossed to a third parent. The single cross hybrids will be planted in the greenhouse this fall for an additional round of crossing, while we will plant the 3-parent F1 hybrids in the field at Lexington this fall to produce F2 seed.
Line development: We grew F3 and F4 headrows at Lexington in 2018 and selected approximately 1614 rows based on height, maturity, clean leaves and heads, and overall vigor. The 1 rep Preliminary trial has been planted at Princeton this fall and will also be planted at Lexington.
Yield testing: Overall, our yields were down a bit in 2018. A very wet, cool spring was followed by a rapid increase in temperatures that reduced the length of the grain filling period and also reduced kernel size. Testing location averages were: Schochoh – 74 bu/a; Princeton 69 bu/a; Woodford 87 bu/a; Lexington 56 bu/a. Disease pressure was not bad, with little scab or stripe rust but a fair amount of Septoria leaf blotch at Schochoh and Princeton.
Scab screening: There was a high level of disease pressure in the irrigated, inoculated Lexington scab nursery, but it was not so high as to prevent us from finding lines with resistance. We are still waiting on DON data for the breeding material.
Purification and Increase: I traveled to Yuma, AZ in May to evaluate our headrows there for uniformity and overall vigor; seed from the selected rows was sent to us in early June. Seed from last year’s Yuma rows had been put into increase plots in 2018 to provide breeder seed of several lines. One of these lines will be increased in 2018-19 by Clements Ag. Supply.
Doubled Haploids: We grew a number of doubled haploid (DH) lines in headrows in 2018 and selected the most vigorous of those that met our requirements for height, maturity, and disease resistance. We selected 150 of these true breeding lines to test in our preliminary trials in 2018 from crosses that were made only 2 seasons ago. The time-savings in getting to the true breeding stage is tremendous.
Genomic Selection: 2018 gave us our third opportunity to evaluate genomic selection. We are particularly interested in using genomic data to select those lines that we have not yet tested in the field, so that we avoid testing lines that have no chance of performing well. To date this has been somewhat, but not completely effective. In other words, if our genomic model predicts that a line will perform poorly, most of the time it does – but not always. We are now, along with many other breeding programs, trying to figure out how to tweak the model so it is more reliable. When we achieve that, it will reduce our costs and make the program more efficient.
2017
Full Report (PDF)
Crossing: We produced a total of 435 crosses in the greenhouse this season; this total includes single cross hybrids and three-way crosses where the hybrid is crossed to another parent. The single-cross hybrids will go back into the greenhouse this fall for an additional round of crossing, and the 3-parent F1 hybrids will be planted in headrows at Lexington this fall to
produce F2 seed.
Line development: We grew F4 and F5 headrows at Lexington in 2017 and selected approximately 1,000 rows based on height, maturity, resistance to stripe rust and head scab and overall vigor. Seed has been packaged into envelopes will be put into a 1 rep Preliminary trial to be grown at Lexington and Princeton in 2017-18.
Yield testing: A very warm February followed by several nights of low temperatures led to concern about spring freeze damage to the wheat crop. However, after the threat of spring freeze, weather was ideal for grain filling and this was reflected in yields and test weights. Testing location averages were: Schochoh – 89 Bu/a; Princeton 87 Bu/a; Woodford 76 Bu/a; Lexington 93 Bu/a. Disease pressure was minimal, with little scab or stripe rust and less septoria leaf blotch than normal. We completed harvest at Spindletop farm on June 30, tied with 2016 for the earliest data of completion.
Scab screening: There was a high level of disease pressure in the irrigated, inoculated Lexington scab nursery, yet we were able to identify lines with resistance. For example, in one of our tests, scabby seed ranged from 5% to 72%. We are still waiting on DON data from the lab that processes our samples.
Purification and Increase: Headrows in Yuma, AZ were screened for uniformity and selected in mid May; seed was sent to us in early June. Seed from our top performing lines will go into large plot increases and will be treated as “breeder seed” that would be suitable for production of Foundation seed. We also grew small plot increases (~ 125 lb) at Spindletop farm for testing in the 2018 KY variety trial and USDA Uniform nurseries.
Genomic Selection: 2017 was our second year of using genomic selection where we take genomic information from lines that have been DNAsequenced and field-tested to develop a prediction model and apply it to new lines that have been sequenced and have just been put into their first year of testing. Some of our top yielding lines were those that were predicted to do well based on the genomic selection model. Other top performers were not predicted to do as well by the model. We will plant a study this fall that includes lines in both categories. By growing these lines at several locations in replicated tests we will learn how best to use the prediction model to make the breeding program more efficient.
Doubled Haploids: Doubled haploid (DH) lines were screened in headrows in 2017 based on height, maturity, vigor and disease resistance. We selected 288 of these true breeding lines to test in our preliminary trials in 2018; this means that approximately 25% of our new lines came from doubled haploids, even though the crosses that produced these lines were made
only 2 seasons ago. The time-savings in getting to the true breeding stage is tremendous.