Maximizing Genetic Gain I
Maximizing genetic gain I
🧭 Overview
🧠 One-sentence thesis
Effective cultivar development requires understanding and manipulating the components of genetic gain per year—especially heritability—to design breeding strategies that maximize selection reliability for quantitatively inherited traits.
📌 Key points (3–5)
- Core tool: A formula for genetic gain per year (from Chapter 17 of Principles of Cultivar Development) helps compare different breeding strategies by examining variance components and heritability.
- What heritability measures: the numerical reliability of selection for a quantitative trait; higher heritability means selection is more dependable.
- Realized heritability in practice: comparing selection among F₂ seeds vs. F₂ plants shows which stage gives more reliable prediction of progeny performance (F₂:₃ lines).
- Common confusion: heritability is not an absolute measure—it is used to compare strategies; breeders may still select even when heritability is low if it helps discard unpromising individuals efficiently.
- Practical threshold setting: instead of relying solely on heritability calculations, breeders often set minimum trait values (e.g., oleic acid %) to discard seeds or plants unlikely to meet breeding goals.
🧬 What is heritability and why it matters
🧬 Definition and purpose
Heritability: a numerical measure of the reliability of selection for a quantitative trait.
- It is not a measure of how much a trait is "genetic" in absolute terms; it quantifies how reliably selection at one stage predicts performance in the next generation.
- The excerpt emphasizes that heritability's primary value is to compare expected gain from different selection strategies.
- Example: Should a breeder select among individual seeds or among individual plants? Heritability calculations help answer this.
🔍 Heritability as a comparison tool
- The excerpt states that "no two breeding programs are designed the same" because resources (facilities, time, money) and traits of importance differ.
- Heritability helps breeders decide which method (e.g., seed vs. plant selection) will give more genetic improvement per year.
- Don't confuse: heritability is strategy-specific and context-dependent, not a fixed property of a trait.
🧪 Realized heritability: seed vs. plant selection
🌱 The experimental setup (Applied Learning Activity 1)
The excerpt describes an experiment to compare selection at two stages for oleic acid content (a quantitative trait):
- F₂ seeds were analyzed non-destructively for oleic acid.
- Each analyzed seed was planted; the resulting F₂ plants were harvested individually and analyzed (using a bulk sample of five F₃ seeds).
- Progeny of each F₂ plant were grown as F₂:₃ lines in multiple locations, and oleic acid content was measured for each line.
The goal: determine whether selecting among F₂ seeds or F₂ plants is more reliable (i.e., which has higher realized heritability).
📐 How realized heritability is calculated
The excerpt provides a step-by-step method:
For F₂ seeds:
- Denominator = (mean of top 5 selected F₂ seeds) − (mean of all F₂ seeds)
This is the selection differential at the seed stage. - Numerator = (mean of F₂:₃ lines tracing to those 5 seeds) − (mean of all F₂:₃ lines)
This is the response to selection in the progeny. - Realized heritability for seeds = numerator ÷ denominator
For F₂ plants:
- Denominator = (mean of top 5 selected F₂ plants) − (mean of all F₂ plants)
- Numerator = (mean of F₂:₃ lines tracing to those 5 plants) − (mean of all F₂:₃ lines)
- Realized heritability for plants = numerator ÷ denominator
🔬 Why plant selection is expected to be more reliable
The excerpt asks: "Why would you expect the heritability to be greater for selection among individual plants than individual seeds?"
- Plants represent a later generation and have been grown in an environment; their phenotype reflects both genetic and environmental effects averaged over development.
- Seeds are measured before planting; their phenotype may be less representative of the genetic value that will be expressed in the field.
- The excerpt does not give the explicit answer, but the implication is that plant measurements integrate more information and reduce measurement error relative to seed measurements.
❓ Why heritabilities are less than 100%
The excerpt asks: "What are reasons to explain why the heritabilities were less than 100%?"
Possible reasons (implied by the context of quantitative genetics):
- Environmental variation: F₂:₃ lines were tested in multiple locations; environment affects trait expression.
- Sampling error: only five F₃ seeds were bulked to measure each F₂ plant; this introduces sampling variation.
- Genotype × environment interaction: the relative performance of lines may change across environments (this is a major theme in the next section, "Maximizing genetic gain II").
🎯 Practical selection without calculating heritability
🎯 Setting minimum thresholds
The excerpt emphasizes that breeders often decide on selection reliability without calculating a heritability value.
- Instead, they set a minimum acceptable trait value (e.g., 50% oleic acid).
- They then look at the seed and plant values for lines that met the threshold in the F₂:₃ generation.
- Example from the excerpt:
- Lines with >50% oleic acid in F₂:₃ had a minimum seed value of 40.86% and a minimum plant value of 42.42%.
- The breeder may decide to discard all future seeds below ~41% and plants below ~42%, reasoning that individuals below these thresholds have "little promise of being useful."
💡 Why select even when heritability is low
- The excerpt notes that a breeder may choose to practice selection "even if the heritability is low" in order to discard seeds or plants that have very little chance of being effective.
- This is a practical efficiency measure: it saves time and resources by eliminating the least promising individuals early, even if the selection is not highly reliable.
Don't confuse:
- High heritability = selection is reliable; chosen individuals are likely to produce superior progeny.
- Low heritability = selection is less reliable, but may still be worth doing to eliminate obvious poor performers.
📊 Variance components and the genetic gain formula
📊 The formula and its components
The excerpt refers to a formula for genetic gain per year in Chapter 17 of Principles of Cultivar Development, based on the variance component method of calculating heritability.
- The formula's components are derived from analyses of variance (ANOVA).
- The excerpt reproduces two ANOVA tables (Tables 17-6 and 17-7) for seed weight in soybeans, showing sources of variation:
- Environments (E)
- Replications within environments (R/E)
- Lines (L) — this is the genetic variance among lines
- Environment × Line interaction (E × L) — how line performance changes across environments
- Replication × Line interaction within environments
- Plants within plots — within-plot sampling error
🔢 What the tables show
| Source | What it represents |
|---|---|
| Lines (L) | Genetic differences among lines; the target of selection |
| E × L | Genotype × environment interaction; lines perform differently in different environments |
| Plants/plots | Sampling variation within plots; measurement error |
- Expected mean squares columns show how each variance component contributes to the observed mean square.
- The excerpt notes that the number of plants per plot (3), replications (2), and environments (2) are used in the formulas.
🧮 How breeders manipulate components
The excerpt states: "Each of the components in the formula can be influenced by the choices breeders make in carrying out their cultivar development programs."
Examples of breeder choices (implied by the ANOVA structure):
- Number of environments: testing in more environments reduces the relative size of E × L variance.
- Number of replications: more reps reduce error variance.
- Number of plants per plot: more plants per plot improve the reliability of line means.
Chapters 6, 7, 18, and 19 of Principles of Cultivar Development (mentioned in the excerpt) explain "how each of the components can be manipulated to maximize genetic improvement."
📋 Applied Learning Activity 1: step-by-step
📋 The task
The excerpt provides a 10-step assignment to calculate and compare realized heritabilities for seed vs. plant selection, using oleic acid data from 50 F₂ individuals.
Steps 1–3: Realized heritability for F₂ seeds
- Highlight the 5 F₂ seeds with highest oleic acid; calculate their mean and the mean of all 50 seeds; subtract to get the selection differential (denominator).
- Highlight the F₂:₃ lines tracing to those 5 seeds; calculate their mean and the mean of all 50 F₂:₃ lines; subtract to get the response (numerator).
- Divide numerator by denominator to get realized heritability for seeds.
Steps 4–6: Realized heritability for F₂ plants 4. Highlight the 5 F₂ plants with highest oleic acid; calculate selection differential (denominator). 5. Highlight the F₂:₃ lines tracing to those 5 plants; calculate response (numerator). 6. Divide to get realized heritability for plants.
Steps 7–10: Interpretation 7. Explain why plant heritability is expected to be greater. 8. Explain why heritabilities are less than 100%. 9. Choose a minimum oleic acid % for discarding seeds and justify. 10. Choose a minimum oleic acid % for discarding plants and justify.
📊 The data
The excerpt includes a table (Table 1) with 50 entries, each showing:
- F₂ seed oleic acid %
- F₂ plant oleic acid %
- F₂:₃ line oleic acid %
Example rows:
- Entry 1: seed 43.43%, plant 37.51%, line 46.21%
- Entry 16: seed 65.28%, plant 62.70%, line 49.72%
The data show that seed and plant values do not always predict line values perfectly, illustrating the concept of heritability < 100%.
🔗 Context and next steps
🔗 Broader breeding program design
The excerpt opens by noting that "no two breeding programs are designed the same" because:
- Available resources (facilities, time, money) differ.
- Traits of importance differ.
- Breeding strategies differ (as seen in publications like Journal of Plant Registrations and Horticultural Science).
The goal of this section is to help breeders understand variables that need to be considered in designing an effective breeding program for selection of traits that are quantitatively inherited.
🔗 Preview of "Maximizing genetic gain II"
The excerpt ends with a brief introduction to the next section, which emphasizes genotype × environment interaction:
- The importance of this interaction is "highly dependent on the trait under selection."
- Example: days to maturity show much more consistent relative differences among genotypes across environments than yield does.
- As a result, the genotype × environment component in the genetic gain equation is smaller for maturity than for yield.
- Therefore, fewer environments are needed to obtain reliable values for maturity.
Don't confuse:
- Low G×E traits (e.g., maturity): genotype rankings are stable across environments; fewer test locations needed.
- High G×E traits (e.g., yield): genotype rankings change across environments; more test locations needed for reliable selection.