Greater wheat yields are of critical importance to meet the projected production targets for world food security. Lodging, the permanent displacement of shoots from their upright position, is a common phenomenon in high yielding crops, reducing yield and quality of irrigated and rainfed wheats alike. In this multi-year study, low-lodging high-yielding wheat germplasm and SNP-tagged novel alleles for lodging were identified by combining (a) donor selection through functional phenotyping for underlying traits with a designed phenotypic screen, and (b) a crossing strategy involving multiple-donor x elite populations.

Firstly, local and imported germplasm was screened to identify consistently higher-yielding genotypes with low lodging for north-eastern Australia. Using fertilisation and tactical irrigation to simulate lodging scenarios, lodging rankings were highly reproducible in multi-environment experiments. In separate field experiments, selected genotypes were phenotyped for traits underpinning stem, shoot and root type lodging. Multi-environment and phenotyping experiments ranked genotypes similarly for lodging score. A phenotyping screen including functional traits underpinning lodging, in relation to root, stem and shoot features was carried out on field plants. Root plate spread from field grown plants consistently discriminated low lodging, high yielding germplasm from a multi-trait analysis quantifying genotypic correlations. When the root plate spread was greater than or equal to 5.5 cm, the lodging scores were small, and yield was high. Root plate spread was highly heritable (above 0.80), with a high genotypic correlation (0.80) across environments.

Secondly, populations were developed crossing 4 low-lodging high-yielding donors based on lodging related traits, and three elite Australian varieties (EGA Gregory, LPB Spitfire and Suntop). Lodging was evaluated in single rows in an early generation and subsequently in plots in 2 years contrasting for lodging. A large number of lines lodged less than their recurrent parents. Heritability for lodging was high, but the genetic correlation between contrasting environments was intermediate-low. Lodging genotypic rankings in single rows did not correlate well with plots.

Finally, lines from the highest lodging background, EGA Gregory, were genotyped (90K iSelect beadchip array). Fourteen markers on 9 chromosomes were associated with lodging, differing under high vs. low lodging conditions. From those, 10 co-located with previously identified QTL for lodging-related traits or at homoeologous locations, while the remaining 4 (in chromosomes 2D, 4D, 7B and 7D) appear to map to novel QTL for lodging. Lines with more favourable markers lodged less, suggesting value in these markers as a selection tool. Our work shows that the combination of donor functional phenotyping, screen design and crossing strategy can help identify novel alleles in a complex trait without requiring expensive phenotyping in extensive bi-parental populations.