Plant breeding and molecular biology appear to be enormously different in the way in which problems are attacked. The objective of the plant breeder is to 'find a good one' and release it to the public for production. The molecular biologist tries to identify an important phenomenon and describe it to his colleagues. The plant breeder looks at thousands of different lines each year and tries to identify a few which merit close attention. He then looks carefully at those few, and how they perform when compared to the local 'industry standards' in replicated experiments which cover his environments of interest. Plant (and animal) breeders use statistics to determine which lines or populations they should focus interest on. The average molecular biologist remains nearly completely anumerate, focusing on single base changes in a sequence, and "yes or no" decisions regarding sequence performance. We plant breeders value a statistically significant 'maybe'.
Statistical analysis permits us to decide what level of risk of failure we are willing to accept. The molecular biologist unwittingly accepts risk of failure when he overgeneralizes or misinterprets an experiment. The plant breeder increases his risk of releasing a poor variety by increasing his acceptable p-value or decreasing his acceptable LoD score (we can also overgeneralize or misinterpret experimental results, but these failings are generally pretty easy for an outsider to see and we wind up looking pretty foolish). From my perspective, the molecular biologist takes inestimable levels of risk, while we estimate the amount of risk we take every time we publish a paper or release a variety. Knowing how much risk you run is rarely comforting.
The primary technique that permits the estimation of risk of making a poor decision is replication. Plant breeders spend much of their time in graduate school learning how to utilize replicated experiments to help in the process of decision-making. If you have not already taken his course in experimental design, I strongly recommend Jack Martin's course. In this lecture I will cover how we use a randomized complete block design to estimate within and among genotype variance, and how we use that information to decide how much of the variation in an experiment is due to genetic differences among genotypes and how much is due to measurement error and environmental variation.
One of the admirable characteristics of barley is its flexibility as a genetic system.Dr. Ken Kasha discovered that barley, when pollinated by its
second-nearest wild relative, Hordeum bulbosum, underwent transient fertilization. Following fertilization, the H. bulbosum chromosomes fail to effectively replicate and are lost during the first few cell divisions of embryogenesis. This results in a haploid embryo, one in which each cell contains only seven chromosomes. The embryo develops relatively normally during the first twelve days of embryogenesis, but after that endosperm development fails and without help the embryo dies. If the embryo is collected before endosperm necrosis begins it can be grown to maturity on artificial media and then germinated. The plant is haploid, and if permitted to grow fails to set seed. If the plant is treated with colchicine
(a spindle-formation inhibitor extracted from a beetle), mitotic spindle formation is inhibited in some dividing cells, chromosome doubling occurs during mitosis and fully diploid tillers are produced. These then set seed normally. They are now diploid, and since each chromosome is a perfect copy of its' sister, the plants are completely homozygous. This means that in a self-pollinating species like barley, each individual produced by the 'doubled haploid' process is essentially immortal. Its seed have its genotype, as do their seen, ad infinitum.
The Steptoe/Morex experiment was conceived at a small meeting in the faculty club at MSU in 1987. We selected these parents because they they differed for many characteristics associated with agronomic performance and grain quality. Steptoe is one of the most broadly adapted cereal varieties known. Steptoe performs well in almost any environment, and consistently produces grain of miserable quality. Morex produces grain of the highest quality (both in terms of feed and malt quality) but is very narrowly adapted. It fits well into the Red River Valley, but fails to compete in Western environments. Also, we demonstrated that Steptoe and Morex varied sufficiently at the nucleotide level to construct a linkage map.
One hundred fifty doubled haploid lines were derived from gametes derived from Steptoe/Morex hybrids. These gametes went through one round of meiosis prior to haploidy and chromosome doubling. This means that the measured frequency with which genes recombine is directly equivalent to recombination frequency. In most species, gametic recombination frequency must be addressed indirectly.
The Steptoe/Morex doubled haploid population was an exceptional population to work with. The doubled haploid lines derived from this population show enormous variation for productivity and almost any morphological character one would like to measure. The parents are both relatively normal barley varieties. Both flower relatively early, but well within the realm of 'normal' behavior for 6-rowed barley varieties grown in temperate environments. The progeny lines often vary far more than their parents differ. This is called 'transgressive variation'. During the first laboratory exercise, you will measure a character which you select, analyze the amount of variation, determine how much of the variation appears to be the result of genetic variance, and then estimate the number of genes likely to contribute to variation.
In the field, the experiment is organized in 4 row plots, 3 meters long. Each row is separated by 30cm. We have border plots (to minimize 'edge effect', and the plots are organized in our irrigated field (the weather station field at the Post Farm) 7 plots wide, 22 ranges long. The last two plots are Steptoe and Morex, making the total field 154 plots in size. Replication 1 is in block three of the weather station field, while replication 2 is in block four. You will go to the Post farm (hopefully while I'm out there), look at the experiment, identify a character you think varies among lines, and measure it in each of the 304 plots within the experiment. This will comprise your Quantitative Trait Locus (QTL) dataset.
The fieldbook files are .The Steptoe/Morex fieldbooks
The Process:
N.B. This represents a pretty tight lab schedule. Work must be done on time. I expect your writeup of your QTL analysis by the end of lecture, October 15. I expect the lab report on your primer design/amplification results to be turned in Thursday, Nov. 15. Sequence alignment, polymorphism identification and cladistic analysis is due December 6. The grant is due in Thursday, Dec. 13.
The Steptoe/Morex Population
This is a population of doubled haploids (and their parents). Doubled haploids are produced in two ways. Immature microspores can be encouraged to develop into small embryos, and then germinated, and ovules may be encouraged to believe that they have been fertilized, and harvested as young embryos. In each case, the young embryos must be germinated on artificial media, then the developing haploid plants (seven chromosomes) must be made diploid through the use of colchicine. These particular doubled haploids were produced from the ovules of Steptoe/Morex F1s. This means that each gamete has undergone one round of meiosis. This makes linkage analysis simple, and reduces the number of genetic classes at each locus to 2. This is the simplest sort of genetic population around.
My fieldwork is nearly finished, so you will need to arrange with me to walk through the experiment. I will show you the 304 4 row plots which comprise the 2-replication Steptoe/Morex doubled haploid population. This population was derived from a cross between one inbred variety of barley which has its origins in Manchuria (Morex) and another inbred (Steptoe) which derives from North African germplasm. These two inbred varieties are representatives of different barley germplasm groups. Each is a fully competent variety, and both perform well in the United States. Morex is a malting variety with limited agronomic performance, while Steptoe is a broadly adapted feed variety with very limited grain quality. The progeny of this population vary for almost any character imaginable. As an example, although the parents differ in flowering date by 1-2 days, their progeny vary (in the right environment) as much as three weeks.
Your job will be include 1) meeting me in the field (the weather station field at the post farm, just West of the road as you enter the farm; 2) observing the population and seeing a character which appears to vary from line to line; 3) moving from plot to plot, giving each plot a numerical rating for that character in your fieldbook. You will then transcribe that data to a field in your electronic fieldbook file.
Step two will involve performing an analysis of variance of the data you've gathered, working through an estimation of the amount of genetic variation you've uncovered, and preparing the data for quantitative trait analysis. I expect each of you to have completed the first phase of this analysis by Sept. 11. We will move on to linkage analysis in the second week of the course.
Objectives for the Week:
1. Meet me in the field sometime between today (Sept. 4) and Wednesday (Sept. 12). Be certain to meet with me and take a quick runthrough of the field experiment.
2. Gather your data for your trait of interest. You need a numerical value for your trait in each of the 304 cells of your fieldbook.
3. Be prepared to do the ANOVA (I use SAS and will assist you in getting acquainted with SAS) by 9/11. On 9/18 we will assemble our means files and begin linkage map construction.