5-parent diallel for cotton fibre length: GCA decides parent retention
Method 2 diallel of five Gossypium hirsutum lines for upper-half mean length. GCA effects, sigma squared GCA versus sigma squared SCA, narrow-sense heritability, and the parent shortlist.
Sara works on cotton fibre quality. She has five elite Gossypium hirsutum lines (P1 to P5) and wants to identify the best parents for a hybrid programme targeting upper-half mean fibre length (mm). She crosses all 10 pair combinations one way only, no reciprocals, plants parents and F1 in an RBD. This is a Method 2 Griffing diallel.
Question
Which parents have the strongest GCA effects, what is the additive share of total genetic variance, and which two parents should be retained for next season's wider-cross programme?
Data, in StatVeda format
# Method 2: parents on diagonal, F1 in upper triangle
P1 P2 P3 P4 P5
P1 29.81 31.42 30.55 31.85 31.21
P2 30.21 31.12 31.65 31.95
P3 29.55 30.95 31.42
P4 30.85 31.78
P5 30.65What she does in StatVeda
Open Plant Breeding category. Pick Diallel (Griffing). Select Method 2, Model I (parents fixed). Paste the matrix. Run.
ANOVA: Treatments (14 df) significant, GCA (4 df) MS = 1.42, SCA (10 df) MS = 0.18, Error MS = 0.04. Both GCA and SCA highly significant. GCA MS / SCA MS = 7.9, indicating additive variance dominates.
GCA effects (deviations from grand mean of 30.93 mm):
P1: -0.18 (slightly negative)
P2: +0.42 (highest GCA)
P3: -0.51 (lowest GCA)
P4: +0.21
P5: +0.06
SCA effects: largest positive SCA in P1xP4 (+0.31), P2xP5 (+0.28).
Variance components: sigma squared GCA = 0.31, sigma squared SCA = 0.07. Narrow-sense heritability h-squared(ns) = 0.78.
What it means
Additive variance is roughly four times the dominance variance. Selection on GCA will work well: parents with positive GCA consistently produce above-average crosses. P2 and P4 are the only two parents with both positive GCA and significant magnitude. P1, P3 and P5 either drag the cross down (P1, P3) or contribute negligibly (P5).
Retain P2 (GCA +0.42) and P4 (GCA +0.21) for next season's wider crossing programme.
Drop P3 from the programme (GCA -0.51, lowest of the set).
Keep P1 and P5 for one more season of evaluation against new parents.
The cross P1xP4 has the best SCA but the GCA-based ranking dominates the decision: SCA matters when programme cannot be enlarged, GCA matters when it can.
Sara writes up the diallel section with the GCA matrix, the variance components, and the heritability number. The narrow-sense heritability of 0.78 supports continued single-trait selection on upper-half mean length.
Why Method 2, not Method 1 or 4
Sara did not make reciprocal crosses. Cotton fibre length is not known to show meaningful maternal cytoplasmic effects in elite Gossypium hirsutum, so reciprocals would have doubled the crossing effort with no expected information gain. Method 1 was therefore unnecessary. Method 4 would have meant skipping the parents in the field, but she planted them deliberately to check that the parental per-se ranking matches the GCA ranking. It does, which is a consistency check on the whole analysis.
What she will do next season
Cross P2 and P4 with two new elite lines from a separate introgression programme, run a Line by Tester in season two with P2 and P4 as testers. The diallel filtered the candidate parents; the Line by Tester will measure GCA of the new lines against the confirmed-good testers. This is the standard two-stage progression for a hybrid breeding programme.
What she will not do
She will not recommend the P1xP4 cross commercially despite its top SCA. With sigma squared SCA only one quarter of sigma squared GCA, the SCA effect is not large enough to override the GCA-driven decision. If the variance ratio had been reversed (sigma squared SCA larger than sigma squared GCA), she would have made the opposite call: pursue the high-SCA cross directly as a hybrid and not worry about parent ranking.