Previous PageTable Of ContentsNext Page

Linkage mapping of QTLs for seed yield, yield components and developmental traits in pea (Pisum sativum L.)

Gail Timmerman-Vaughan1, A Mills1, TJ Frew1, R.C. Butler1, J.A. McCallum1, Sarah Murray1, C.P. Whitfield1, A.C. Russell2 and D.R. Wilson1

1 New Zealand Institute for Crop & Food Research, Private Bag 4704, Christchurch, NZ. Email:
Plant Research (NZ) Ltd, PO Box 19, Lincoln, NZ. Email:


Improvement of yield in pea is a major objective. Yield is a complex trait that is strongly influenced by both genotype and environment. The yield of pea crops can be described in terms of its components which include plant number per unit area, seed weight and seed number. These yield components show interdependence or “plasticity” which makes it very unlikely that selection for increased yield based on a single component will succeed. Using molecular linkage maps of the pea genome and QTL mapping we have characterised the genetics of yield per se, yield determinants (seed weight, seed number and harvest index) as well as developmental traits. The traits were measured in three replicated field trials managed to ensure maximum expression of yield potential. Genetic loci affecting yield-related traits occur in 19 pea genomic regions. In addition, QTLs for different yield-related traits coincided in clusters, providing the basis for understanding at the genetic level the “plasticity” that is observed among yield components.

Media summary

Yield determination in pea is complex, and involves both genotype and environment. Research identifying multiple genetic loci for yield and related traits is presented.


quantitative trait loci, seed yield, Pisum sativum, linkage mapping


Pea (Pisum sativum L.) is grown worldwide for diverse uses as food and feed. When developing pea cultivars, plant breeders aim to obtain high yield, disease resistance, quality attributes, and adaptation to a range of environmental conditions. Reducing yield variability is an important goal because pea exhibits poor stability of yield across years and environments compared with other crops. Yield is a complex trait which, from a crop physiology perspective, is the culmination of a sequence of processes (phenological and canopy development, radiation interception, and biomass production and partitioning) that are driven by environmental influences (Charles-Edwards 1982). Ultimately, a genotype’s performance is determined by how it integrates all the influences it experiences. The end-result is seed yield which often is described as the product of four components: number of plants per unit area, mean number of pods per plant, mean number of seeds per pod, and mean seed weight (Moot and McNeil 1995). Timing is important because similar influences at different times can have very different effects on the structure of yield in terms of these components. As a result, the yield components show great interdependence or “plasticity” (Wilson 1987). For example, compensation occurs between numbers of pods per plant and seeds per pod (Moot and McNeil 1995) and between seed number and seed weight (Sarawat et al. 1994).

With the development of molecular linkage maps and statistical approaches for mapping quantitative trait loci (QTLs), it is possible to identify and tag genetic loci that are associated with complex traits such as yield. Often, QTL mapping studies examining complex and inter-related traits identify coincident QTLs and, therefore, may provide information on the genetic basis of phenotypic correlations that are observed between traits. In this paper, we report the results of a QTL mapping study that has characterised the genetics of yield, yield components (seed weight, seed number and harvest index) and developmental traits in a population derived by crossing two ‘adapted’ field pea genotypes, a marrowfat pea cultivar (Primo) and a blue pea breeding line (OSU442-15).


The ‘Primo’ X ‘OSU442-15’ (PX4) F2 population used for linkage and QTL mapping was developed as described by Timmerman-Vaughan et al. (1996). DNA marker methods were carried out as described previously (Timmerman-Vaughan et al. 2002).

Field trials were conducted during the summers of 1997-98, 1998-99 and 2002-03 in irrigated fields near Christchurch, New Zealand. Each trial contained two replicates of 227 F2-derived families, as well as a number of replicate check line plots (OSU442-15 and Primo) distributed to give approximately even placement across rows and columns. The trials were managed intensively, using best practices for pea crops, to minimise biotic and abiotic stresses and thereby to ensure maximum expression of yield potential. The traits measured were: yield (yld), 1000 seed weight (wt), seed number (num, derived from yield and mean seed weight), harvest index (hi), node of first flower (nff), total node number (tnn), and number of flowering nodes (nfn, derived from nff and tnn). Field data were analysed to estimate line means after adjustments for any spatial patterns within the trial using residual maximum likelihood (REML) as implemented in GenStat (GenStat Committee 2002 and previous releases).

Linkage maps were constructed using MAPMAKER/EXP ver. 3.0 as described by Timmerman-Vaughan et al. (2002). Composite Interval Mapping (CIM) was applied to map QTLs using QTL Cartographer ver. 2.0 software. Permutation tests were used to determine chromosome-wise significance thresholds (α=0.05) for QTL detection using QTL Cartographer

Results and discussion

Phenotype analysis

Trait means for progeny families showed continuous variation typical of quantitative traits, suggesting that multiple genetic loci are likely to be involved in determining phenotype. Transgressive segregation of progeny family means was observed for the developmental traits. Trait correlations for the progeny families were compared within environments. In all three environments, seed number was strongly positively correlated with yield, and strongly negatively correlated with seed weight. Seed weight was more weakly inversely correlated with yield. While developmental traits were not strongly correlated with seed weight, number or yield, tnn was strongly positively correlated with both nff and nfn.

Linkage map

The linkage map of the 227 F2 families from the PX4 cross used for QTL mapping was constructed using 108 molecular markers and covers 1369 cM (Haldane function) on 11 linkage groups. Linkage groups were identified by mapping anchor loci from the consensus map for P. sativum (Weeden et al. 1998) or from other linkage maps in the published literature.

QTL discovery and coincidence of QTLs

Using molecular linkage maps and composite interval mapping, we have identified 19 pea genomic regions that have significant marker-phenotype associations (QTLs) for yield per se, yield determinants and/or developmental traits (Table 1). In most cases, each of these genomic regions contains clusters of coincident QTLs detected using distinct traits or, in some cases, the same trait in different environments (years). Particularly notable examples are found on linkage groups I, III and VII (Fig. 1). Coincidence of QTLs may indicate either that single genes found in these genomic regions have pleiotropic effects or that the genomic regions associated with these QTLs carry groups of linked genes associated with contributing to yield per se, yield determinants and/or developmental traits. The strongest trait correlations were observed for seed yield and seed number (positively correlated) and seed weight and seed number (negatively correlated). Seed weight, seed number, harvest index and/or yield appear to be coordinately controlled by genetic loci that occur in at least eight distinct genomic regions. QTLs for developmental traits are also associated with genomic regions that carry QTLs for seed yield and yield determinants.

QTLs for seed weight and seed number coincide in association with five genomic regions, on linkage groups I (2 QTLs), IV (2 QTLs) and VII. In all cases, the marker alleles associated with the genetic loci have inverse effects on the seed weight versus seed number phenotype means, which is consistent with the observed negative correlation between the trait phenotypes, and provides a genetic basis for the compensation that occurs between these yield components. Likewise, seed yield QTLs coincide with seed weight, seed number and/or harvest index QTLs in genomic regions of linkage groups III, IV and VII. Although seed yield and seed number are highly correlated, QTLs for these traits only coincided in two genomic regions, on linkage groups III and VII.

Table 1. Summary of QTL peak clusters for seed yield, yield determinants and developmental traits. Trait designations are num (seed number), wt (1000 seed weight), hi (harvest index), yld (seed yield), nff (node of first flower), tnn (total node number) and nfn (number of flowering nodes).


QTL Cluster


LOD range

R2 range (%)








num, wt, hi, nff, nfn





num, wt





num, tnn















yld, num, hi, tnn, nff, nfn










wt, tnn





num, hi





tnn, yld





wt, yld





wt, num





wt, num, nff










wt, tnn





nfn, tnn





yld, num, wt








Figure 1. Peak locations of QTLs detected on linkage groups I, III and VII. Bubbles represent the 1 LOD confidence intervals for QTL peaks and solid circles represent the QTL peak location.

QTLs detected by developmental traits (nff, tnn and/or nfn) also map to genomic regions that carry QTLs for yield and yield determinants. The coincidence of QTLs for these developmental traits and yield-related QTLs also confirms the role of plant development, including the duration of flowering (Bourion et al. 2002), in yield determination through its influence on determining the timing of events.


By identifying clusters of coincident QTLs for different yield-related traits, we have shown that there is a clear genetic basis for trait correlations (e.g. between seed weight and seed number, or between seed number and seed yield) that are observed in pea and contribute to the interdependency, or “plasticity”, of yield components. QTL clustering may indicate that a single gene has a pleiotropic effect that is expressed as a number of phenotypes, or that a number of distinct genes are linked. These separate mechanisms will not be distinguished in most cases without further genetic and molecular characterisation.


We thank Michael Lakeman for assistance with marker analyses. This research was funded by the New Zealand Foundation for Research, Science and Technology.


Bourion V, Fouilloux G, Le Signor C and LeJeune-Henaut I (2002). Genetic studies of selection criteria for productive and stable peas. Euphytica 127, 261-273.

Charles-Edwards DA (1982). Physiological determinants of crop growth. Academic Press, Sydney. 161 pp.

Genstat Committee (2000). Genstat 5, Release 4.2, Reference Manual Supplement. Numerical Algorithms Group Ltd. Oxford, United Kingdom.

Moot DJ and McNeil DL (1995). Yield components, harvest index and plant type in relation to yield differences in field pea genotypes. Euphytica 86, 31-40.

Sarawat P, Stoddard FL, Marshall DR and Ali SM (1994). Heterosis for yield and related characters in pea. Euphytica 80, 39-48.

Timmerman-Vaughan GM, McCallum JA, Frew TJ, Weeden NF and Russell AC. (1996). Linkage mapping of quantitative traits controlling seed weight in pea (Pisum sativum L.). Theoretical and Applied Genetics 93, 431-439.

Timmerman-Vaughan GM, Frew TJ, Russell AC, Khan T, Butler R, Gilpin M, Murray S and Falloon K. (2002). QTL mapping of partial resistance to field epidemics of ascochyta blight of pea. Crop Science 42, 2100-2111.

Weeden NF, Ellis THN, Timmerman-Vaughan GM, Sweicicki WK, Rozok SM and Bernikov VA. (1998). A consensus linkage map for Pisum sativum. Pisum Genetics 30, 1-4.

Wilson DR (1987). New approaches to understanding the growth and yield of pea crops. In ‘Peas: Management for quality’. (Ed. W.A. Jermyn and G.S Wratt) pp. 23-28. Agronomy Society of New Zealand Special Publication No. 6.

Previous PageTop Of PageNext Page