Crop Science Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 27 May 2005
Published in Crop Sci 45:1336-1344 (2005)
© 2005 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (8)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Timmerman-Vaughan, G. M.
Right arrow Articles by Wilson, D.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Timmerman-Vaughan, G. M.
Right arrow Articles by Wilson, D.
Agricola
Right arrow Articles by Timmerman-Vaughan, G. M.
Right arrow Articles by Wilson, D.
Related Collections
Right arrow Other Legumes
Right arrow Crop Genetics

GENOMICS, MOLECULAR GENETICS & BIOTECHNOLOGY

Linkage Mapping of QTL for Seed Yield, Yield Components, and Developmental Traits in Pea

Gail M. Timmerman-Vaughana,*, Annamaria Millsb, Clare Whitfieldc, Tonya Frewa, Ruth Butlera, Sarah Murraya, Michael Lakemand, John McCalluma, Adrian Russelle and Derek Wilsona

a New Zealand Institute for Crop & Food Research Ltd., Private Bag 4704, Christchurch, New Zealand
b Plant Sciences Group, P.O. Box 84, Lincoln Univ., Canterbury, New Zealand
c Joint Nature Conservation Committee, Monkstone House, City Road, Peterborough PE1 1JY, UK
d Dep. of Biology, Univ. of Washington, Seattle, WA 98195-1800
e Plant Research (NZ) Ltd., PO Box 19, Lincoln, New Zealand

* Corresponding author (timmermang{at}crop.cri.nz)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
Seed yield in pea (Pisum sativum L.) is a physiologically complex trait that is strongly influenced by both genotype and environment. Seed yield 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 and plasticity in response to environment; therefore, selection based on a single component is unlikely to succeed in increasing yield in breeding programs. To improve our understanding of the genetic basis of seed yield determination in pea and to identify genetic loci involved, quantitative trait loci (QTL) for yield per se, yield components (seed weight, seed number, and harvest index [HI]) and developmental traits (node of first flower [NFF], number of flowering nodes [NFN], total node number [TNN]) were mapped. The QTL mapping was conducted using F2–derived families of a cross between Primo, a marrowfat cultivar, and OSU442-15, a blue pea breeding line. Linkage maps containing 108 loci on 11 linkage groups (LGs) were constructed for 227 families derived from this cross. Traits were measured in three replicated field trials conducted in New Zealand during the summers of 1997–1998, 1998–1999, and 2002–2003. The trials were managed to ensure maximum expression of yield potential. The QTL affecting yield-related traits were associated with 19 genomic regions on LGs I, II, III, IV, VI, and VII. The QTL for different yield-related traits were colocalized, suggesting some basis for understanding the reproductive plasticity that is observed in pea at the genetic level.

Abbreviations: AFLP, amplified fragment length polymorphism • agpL2, ADP glucose pyrophosphorylase large subunit L2 • CIM, composite interval mapping • HI, harvest index • LG, linkage group • MAS, marker-assisted selection • NFF, node of first flower • NFN, number of flowering nodes • NUM, seed number per plot or per square meter • Px4, Primo x OSU442-15 • QTL, quantitative trait loci • RAPD, random amplified polymorphic DNA • RFLP, restriction fragment length polymorphism • sAFP, allele-specific polymerase chain reaction marker derived from an amplified fragment length polymorphism • SPS, sucrose phosphate synthase • STS, sequence tagged site • SUS, sucrose synthase • TNN, total node number • TSW, 1000-seed weight


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
PEA IS GROWN throughout the world for diverse uses as food and feed. Field pea is the world's third most significant grain legume after soybean and common bean. The aims when developing pea cultivars include obtaining high yields, 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 that, from a crop physiology perspective, is the culmination of a series of processes (phenological and canopy development, radiation interception, biomass production and partitioning) that are driven by environmental influences (Charles-Edwards, 1982). A genotype's ultimate performance is determined by how it integrates genotype and environmental influences. The end result is seed yield, which has often been described as the product of its components: number of plants per unit area, number of seeds per unit area (number of pods per plant, number of seeds per pod), and mean seed weight (Moot and McNeil, 1995). These yield components show interdependence or plasticity (Wilson, 1987). For example, compensation is observed between the number of pods per plant and number of seeds per pod (Moot and McNeil, 1995), or between seed number and seed weight (Sarawat et al., 1994). Since selecting for a particular yield component may be ineffective as a means of increasing yield per se, alternate approaches have been proposed. Hedley and Ambrose (1985) suggested selecting for plants that produce a high, stable plant HI, while Wilson (1987) suggested that selecting weakly competitive plants may increase yield stability, based on evidence that weak competitors may be more successful within a pea crop than vigorous genotypes that perform well as single plants. Bourion et al. (2002) suggested that progress in seed yield and yield stability may be achieved using lines with different developmental and architectural features. For the genotypes that they examined, this would involve focusing primarily on NFF and on leaf appearance rate.

With the development of molecular linkage maps and statistical approaches for mapping QTL, it is possible to identify and tag genetic loci controlling genetically complex traits. Quantitative trait loci mapping provides information on the minimum number of genetic loci that may be involved in determining a trait phenotype, the genetic map locations of those loci and estimates of the effects of the loci on trait values—information that is potentially very useful to the plant breeder. Molecular markers associated with QTL may also prove to be useful for marker-assisted plant breeding. In addition, QTL mapping offers the opportunity to use map-based cloning approaches to identify the genes underlying the genetic loci (Peters et al., 2003). Quantitative trait loci mapping studies examining complex and interrelated traits often identify coincident QTL. In rice (Oryza sativa L.), for example, QTL determining yield and yield components (Thomson et al., 2003) or agronomic traits (Mei et al., 2003) cluster in a number of genomic locations. In common bean (Phaseolus vulgaris L.; Tar'an et al., 2002), QTL for agronomic traits occur in clusters (e.g., days to flowering and days to maturity; or total nodes, days to maturity, and yield). In pea, QTL coincide for relative plant maturity and field resistance to Ascochyta blight, two traits that show phenotypic correlations (Timmerman-Vaughan et al., 2002). Quantitative trait loci mapping, therefore, also provides information on the genetic basis of phenotypic correlations.

Pea is an important model system for legume genetics. Single gene characteristics potentially related to yield, including the genetics of flowering, photoperiodism, and plant architecture have been studied using classical genetic approaches (reviewed in Weller et al., 1997; Beveridge et al., 2003). Quantitative trait loci mapping provides the opportunity to extend the results of classical genetic studies. Quantitative trait loci mapping studies in pea have examined seed weight and color (Timmerman-Vaughan et al., 1996; McCallum et al., 1997), disease resistance, and development or plant architecture (Dirlewanger et al., 1994; Tar'an et al., 2003), branching (Rameau et al., 1998) and disease resistance (Pilet-Nayel et al., 2002; Timmerman-Vaughan et al., 2002). This paper reports the results of a QTL mapping study of yield, yield components, and developmental traits in a population derived from a cross between two adapted field pea genotypes, the marrowfat pea cultivar Primo and the blue pea breeding line OSU442-15.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
Population Development
The Primo x OSU442-15 (Px4) F2 population used for linkage and QTL mapping was developed as described by Timmerman-Vaughan et al. (1996). Two hundred and twenty seven F2:3 families were selected for field trialing in 1997–1998 based on the availability of >80 seeds from single F2 plants. The F2:4 seed harvested from the 1997–1998 trial was sown in subsequent field trials conducted in 1998–1999 and 2002–2003.

DNA Marker Methods
DNA was extracted from young leaves pooled from ≥5 F3 plants to reconstitute the F2 genotypes. DNA extractions, RAPD (random amplified polymorphic DNA), RFLPs (restriction fragment length polymorphisms), and AFLPs (amplified fragment length polymorphisms) were performed as described previously (Timmerman-Vaughan et al., 2002). The AFLPs are labeled according to the primers used for selective amplification and the size of the PCR product. The RFLP probes c44 and c267 were obtained from Noel Ellis (John Innes Institute, Norwich, UK). R-gene analog probes were described by Timmerman-Vaughan et al. (2000), and other RFLP probes were described by Gilpin et al. (1997). New polymorphic STS (sequence tagged site) loci were developed as described by Frew et al. (2002). The STS PCR amplifications were performed using the following primers (sAFP = allele-specific PCR marker derived from an AFLP): sAFP9-6 (5' GCT TGT GGC TTG AAT TGT ATT TGC 3'; 5' GCT TCA CCT GCT ACA CAA CAC CAG 3'), sAFP16-3 (5' GCT CAT ACC AAA AGG ATC TTG AAC 3'; 5' CAG GGA AAA TGG AGA ACA GAA A 3'), SPS2 (sucrose phosphate synthase, Genbank Z56278; 5' TGG CAT ATC CTA AAC ACC ACA 3'; 5' GGT AAG ACC AAA CGG CTC AA 3'), SUS3 (sucrose synthase, rug4 locus, Genbank AJ012080; 5' CAA ACC AGA TTT GAT TGT TGG A 3'; 5' TTC TCA AGT GCA TGA GCA AT 3') and agpL2 (ADP glucose pyrophosphorylase subunit L2, Genbank Y08728; 5' CAA GTG GAT ACA ACT GTT CTT GGT 3'; 5' CTC CTA TTC CTA CTG GAA CTC TCC 3'). Polymorphisms were revealed by SSCP (single strand conformational polymorphism) gel electrophoresis (McCallum et al., 2001), except that the agpL2 polymorphism was revealed by RsaI digestion of the PCR product.

Field Trials
Field trials to measure seed yield and related traits were conducted during the summers of 1997–1998, 1998–1999, and 2002–2003 in high performance irrigated fields near Rakaia, Canterbury, New Zealand (1997–1998 and 1998–1999), and near Lincoln, Canterbury, New Zealand (2002–2003). Each field trial contained two replicates of the 227 F2–derived families being used for QTL mapping, as well as a number of replicate check line plots (OSU442-15 and Primo). The check plots were distributed to give coverage across each whole trial using a modified Latin-square design.

The 1997–1998 trial consisted of small plots of 40 plants sown in twin rows (20 plants row–1) with a between-row spacing of 25 cm and a within-row spacing of 10 cm between plants. The 1998–1999 and 2002–2003 trials consisted of 0.75- by 3.0-m plots sown at 100 seeds m–2. Trial crops were managed for high performance using standard commercial management practices. Pre- and postemergent herbicides were applied, and crops were sprayed once (in December) to control powdery mildew after first lesions were seen. Base fertilizer (N–P–K–S = 19.5:10.0:0.0:12.5) was applied to the 1997–1998 trial at 150 kg ha–1. Soil tests indicated that fertility was high at the 1998–1999 and 2002–2003 sites so no fertilizer was applied. Irrigation was managed using measurements of soil moisture content with neutron probes following standard practices, with the aim of maintaining moisture content above the critical deficit level for the Templeton silt loam soil type.

The traits measured from the 1997–1998 trial were yield per plot (YLD98) and 100-seed weight. Thousand-seed weight (TSW98) was derived from 100-seed weight, and seed number per plot (NUM98) was derived from the raw yield and TSW values. The seed yield-related traits measured from the 1998–1999 and 2002–2003 trials were yield m–2 (YLD99, YLD03) and TSW (TSW99, TSW03). Seed number m–2 (NUM99, NUM03) was derived from the raw yield and seed weight values. In addition, HI [(seed dry weight/total plant dry weight) x 100] was measured from the 1998–1999 trial (HI99) by taking a 0.25 m2 quadrat from within each plot after grain filling was complete but before harvest. Plant development traits were also measured from the 1998–1999 and 2002–2003 trials. First flowering node (NFF99, NFF03) and total node number (TNN99, TNN03) are mean values determined by averaging node counts for five randomly selected plants taken from within the center three rows of each plot. The mean NFN (NFN99, NFN03) was derived using the formula NFN = TNN – (NFF + 1).

Statistical Analysis of Field Data
Field data were analyzed to estimate F2–derived family line means after adjustments for any spatial patterns within the trial. Possible spatial trends were first explored graphically, including looking at average row or column patterns and plotting all data using a color scale for each plot in its field layout. Adjustments were made by including factors for these trends in the analyses and/or by including correlations to model between-plot associations. Models were fitted using residual maximum likelihood (REML) as implemented in GenStat (GenStat Committee, 2002 and previous releases). The approach used followed closely that described in Gilmour et al. (1997). In the analysis, differences between the parents and the means of the trial lines were included as fixed effects, and differences between the trial lines as a random effect. Systematic field trends were included as fixed effects, and more general column, row, or replicate patterns as random effects. In addition, and using the same approach, a null model was fitted which did not include any adjustments for field trends. Spatial adjustment of field trends using this statistical approach has been discussed in more detail elsewhere (Gilmour et al., 1997; Verbyla et al., 1999; Timmerman-Vaughan et al., 2002).

Trait correlations were calculated using the adjusted mean values. Narrow-sense heritabilities (h2) were calculated for the entry means using the results of the null analysis and the analysis with adjustment for field trends using the formula:

where {sigma}g2 = genotypic variance, {sigma}p2 = phenotypic variance, {sigma}e2 = environmental variance, {sigma}2 = plot variance, r = replication of F2–derived families (i.e., 2). {sigma}g2 and {sigma}2 were estimated as an integral part of the analyses done.

Linkage and QTL Mapping
The Px4 molecular linkage map was constructed using MAPMAKER/EXP v. 3.0 (Lincoln et al., 1992) as described by Gilpin et al. (1997) and Timmerman-Vaughan et al. (2002). Markers were assigned to LGs using the two point and group commands with a threshold of LOD ≥ 5.0. The LGs were constructed using primarily the compare and build commands, and the final orders were tested using the ripple command, with a threshold of LOD ≥ 2.0.

Composite interval mapping (CIM) was applied using Windows QTL Cartographer v. 2.0 software (Wang et al., 2003). Composite interval mapping was conducted using the default settings (e.g., Model 6, five cofactors selected automatically by forward regression with a 10-cM window). Permutation tests (Churchill and Doerge, 1994) were used to determine chromosome-wise significance thresholds ({alpha} = 0.05) for QTL detection using QTL Cartographer with 1000 shuffles of the trait data. Quantitative trait loci peaks with significant likelihood ratio test statistics under H3/H0 have been reported. These two hypotheses are:

where a = additive effect and d = dominance effect.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
SUMMARY
 REFERENCES
 
Phenotype Analysis
Analysis showed strong spatial patterns in the field trials for yield and yield components (seed weight, seed number). For example, there were obvious fluctuations in yield across the width of the 2002–2003 field trial associated with the position of the irrigator nozzles and wheels, and there was a general decline in yield from left to right in the field (Fig. 1A). Adjusting for these trends with REML strongly attenuated this pattern (Fig. 1B). Means for F2–derived family lines after adjustment for spatial patterns differed from the raw means by up to 37%, indicating substantial field trends that could have biased further analysis without adjustment. In most cases, however, adjustment was rather less than 37%. Since obvious spatial patterns were observed in field trials, QTL detected using adjusted means are presented. To determine the effect of adjustment on QTL detection, CIM was performed using both raw and adjusted means. Very similar results were obtained (in terms of number of QTL detected, peak locations, and significance) in spite of the magnitude of some adjustments (data not presented). This similarity is probably due to the replication of the F2–derived family lines in field trials.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 1. Example of spatial trends in seed yield observed within the 2002 to 2003 field trial. The row means of the residuals (residual = value of the plot estimated by the analysis minus the raw plot value) are shown for the analysis with no spatial adjustment (top panel) and with adjustment for the trend across the field and for irrigator patterns (bottom panel).

 
Parental means and narrow-sense heritabilities for the trait values after calculation of the null model and adjustment for spatial trends in the three field trials are presented in Table 1. Very high heritabilities (h2 = 0.90–0.99) were obtained for TSW. Heritabilities for NUM and developmental traits (TNN, NFF, and NFN) were generally quite high for null model values, but seed number heritabilities in particular increased considerably after adjustment for spatial trends in the field. In contrast, heritabilities calculated for yield using the null model F2 family values were low, but increased substantially after adjustment. Transgressive segregation of progeny family means was observed for the developmental traits (data not shown).


View this table:
[in this window]
[in a new window]
  		Table 1. Parental mean values and narrow-sense heritabilities (h2) calculated for the entry means after the null analysis and after adjustment for field trends. Traits analysed were thousand-seed weight (TSW), seed number (NUM), seed yield (YLD), harvest index (HI), node of first flower (NFF), number of flowering nodes (NFN), and total node number (TNN). Standard errors (SE) of the calculated heritabilities are indicated.

 
Trait correlations for the progeny families were computed within environments and are presented in Table 2. 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 in the 1997–1998 and 2002–2003 environments. Negative correlation is commonly observed between seed weight and seed number among grain legumes, for example in cowpea (Kheradnam and Niknejad, 1970) and Vicia faba (Yassin, 1973). In pea, Pandey and Gritton (1975) observed significant phenotypic correlations between yield and seeds per plant, and negative correlations between pods per plant and seed weight.


View this table:
[in this window]
[in a new window]
  		Table 2. Correlation matrix for the adjusted means of traits within environments. Correlations > |0.130| are significant at P < 0.01.{dagger}

 
Linkage Map
The linkage map of the 227 Px4 F2–derived families used for QTL mapping was constructed using RFLP, STS, RAPD, and AFLP markers and covers 1369 cM (Haldane function) on 11 LGs. Of the 108 loci used for map construction, 48 show codominant inheritance. The mean distance between loci is 14.1 cM. The linkage maps for eight LGs are presented in Fig. 2. The LGs not drawn include two short regions of LGs II and IV, LG V, and a portion of LG VI containing anchor loci dehydrin and pID18 (Weeden et al., 1998). The LGs were identified based on mapping anchor loci from the consensus map for P. sativum (Weeden et al., 1998) or from other linkage maps (Timmerman-Vaughan et al., 2000). Comparison with the consensus pea linkage map (Weeden et al., 1998) indicates that some genomic regions are missing from the Px4 linkage maps used for QTL mapping. In particular, our maps appear to be missing the bottom end of LG I, the bottom end of LG V, and the top end of LG VII. Anchor loci for these regions were not polymorphic between the parental lines, Primo and OSU442-15. At 1369 cM, the pea linkage map is longer than might be expected, especially in comparison with the pea consensus map (Weeden et al., 1998). Other pea maps, however, have also been longer than the pea consensus map (Pilet-Nayel et al., 2002; Tar'an et al., 2003). A linkage map for 108 F2 progeny of the Px4 cross was published previously (Gilpin et al., 1997). Some LGs have been reassigned since that publication.



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 2. Quantitative trait loci (QTL) detected by composite interval mapping. Bars represent the 1 LOD confidence intervals for QTL peaks, and solid circles represent the QTL peak location. Linkage groups and QTL designations (e.g., num1.1) are indicated.

 
QTL for Yield and Yield Components
Seed weight QTL are associated with nine genomic regions, on LGs I (two QTL), IIIa, IV (two QTL), IVa, VI (two QTL), and VII. Most seed weight QTL were detected using data from at least two years' trials, or coincide with QTL peaks obtained for other traits (Fig. 2, Table 3). For most seed weight QTL, increased seed weight is associated with the Primo (large-seeded parent) alleles. The exception is tsw7.1, on LG VII, for which large seed weight is associated with marker alleles from the small-seeded parent, OSU442-15. Seed weight QTL account for 65% (TSW99), 66% (TSW98), and 46% (TSW03) of the observed phenotypic variation. Individual seed weight QTL account for 3 to 19% of the observed phenotypic variation (Table 3). A previous paper reported mapping QTL for seed weight using 108 F2 plants derived from the Px4 cross, and 51 RILs from the cross JI1794 x Slow (Timmerman-Vaughan et al., 1996). The results in the current paper have confirmed three QTL identified in the previous study: tsw1.1 associated with P445 (now assigned to LG I, Timmerman-Vaughan et al., 2000), tsw3.2 associated with LG IIIa markers near M27, and tsw4.2, which is associated with Y02-850 in the present study and was associated with A09_1250 in the previous study. The increased number of seed weight QTL identified in the present study is due in part to this Px4 linkage map being more complete than the previous map (Timmerman-Vaughan et al., 1996), but may also be due to the use of three environments and a larger mapping population.


View this table:
[in this window]
[in a new window]
  		Table 3. Summary statistics for quantitative trait loci detected using yield, yield component, and developmental traits.

 
On LG VI, the QTL tsw6.2 and tnn6.1 are found in the same genomic region as sbm1, a gene for resistance to pea seed-borne mosaic virus (PSbMV) (Timmerman et al., 1993). PSbMV infection can reduce seed weight and plant vigour. Primo (Sbm1/Sbm1, susceptible) and OSU442-15 (sbm1/sbm1, resistant) differ for PSbMV resistance. These LG VI QTL are unlikely to be the result of a pleiotropic effect of sbm1 gene action because the large seed weight and TNN progeny means are associated with Primo marker alleles.

Seed number QTL have been detected in association with nine genomic regions, on LGs I (three QTL), II, III, IIIa, IV, IVa, and VII. OSU442-15 contributes the large seed number allele at eight of the nine QTL regions, with the exception being the QTL on LG VII. Seed number QTL account for 43% (NUM03), 62% (NUM98), and 51% (NUM99) of the phenotypic variation. Individual QTL explain 5 to 27% of the observed variation (Table 3).

Harvest index was measured only on the 1998–1999 field trial. Harvest index QTL were detected in association with markers on LGs I, II, III and IIIa. Harvest index QTL account for 40% of the observed phenotypic variation and individual QTL accounted for 6 to 20% of that variation. The QTL detected for HI coincided with QTL for other traits.

Seed yield QTL are associated with genomic regions on LGs III, IV (two QTL), and VII (two QTL) (Fig. 2). Increased seed yield is associated with Primo marker alleles at all of these QTL except yld3.1 (Table 3). The QTL detected in each environment explain 15% (YLD98), 18% (YLD99), and 25% (YLD03) of phenotypic variation. In contrast with seed weight and seed number, relatively few yield QTL were identified from each environment (Table 3). The LG VII seed yield QTL (yld7.1 and yld7.2) were detected using data from two environments. In all cases, yield QTL coincided with at least one other QTL.

QTL for Developmental Traits
Eleven genomic regions are associated with QTL determining NFF, TNN, and NFN based on node counts made in the 1998–1999 and 2002–2003 field trials. These QTL are found on LGs I, II (three QTL), III, IIIa (two QTL), IV, IVa, VI, and VII (two QTL) (Fig. 2). Examination of the additive effect values for TNN, NFF, and NFN QTL indicates that both Primo and OSU442-15 carry alleles that increase these trait values. For example, the Primo allele increases NFF at QTL nff1.1 and nff3.1, while the Primo allele decreases NFF at QTL nff2.1 and nff4.1 (Table 3). Likewise, six genomic regions carry QTL for TNN. The Primo allele increases TNN at QTL tnn3.1, tnn3.2, and tnn6.1, and decreases TNN at QTL tnn2.1, tnn4.1, and tnn7.1 (Table 3). The contribution of high value alleles by both parental lines is likely to explain the transgressive segregation observed for these traits.

Coincidence of QTL
Quantitative trait loci for a number of traits coincide (Fig. 2), indicating either that a causal relationship exists, that single genes underlying the QTL have pleiotropic effects, or that the genomic regions associated with these QTL carry groups of linked genes that affect yield per se, yield components and/or developmental traits. The QTL for seed weight, number, yield and/or HI coincide in association with at least eight distinct genomic regions. The QTL for developmental traits also cluster with QTL for seed yield and yield components.

Quantitative trait loci for seed weight and number coincide in association with five genomic regions (Fig. 2). In all cases, the additive effects associated with linked marker alleles have inverse effects on the seed weight vs. seed number means (Table 3). This is consistent with the negative correlation observed between the seed weight and number (Table 2) and provides a genetic basis for the compensation that is observed between these traits.

Likewise, seed yield QTL coincide with seed weight, number, and/or HI QTL. Although seed yield and number are highly correlated, QTL for these traits coincided in only two genomic regions, on LGs III (yld3.1, hi3.1, and num3.1) and VII (yld7.1 and num7.1). A strong positive correlation was observed between seed yield and seed number (Table 2). The additive effects associated with markers linked to QTL yld3.1 and num3.1 as well as to yld7.1 and num7.1 are in the same direction (Table 3), which is consistent with the observed trait correlations. Quantitative trait loci for seed yield and seed weight coincide on LGs IV (yld4.2 and tsw4.1) and VII (yld7.1 and tsw7.1). Seed yield and seed weight are more weakly negatively correlated (Table 2), and this is reflected in the directions of the additive effects observed for the coincident QTL, which are in the same direction for yld4.2 and tsw4.1 but are in opposite directions for yld7.1 and tsw7.1.

Quantitative trait loci detected by NFF, TNN, and/or NFN also map to genomic regions that carry QTL for yield, seed weight, and number, in particular on LGs I, II (two regions), III, IIIa, IV, IVa, and VI. The coincidence of QTL for these developmental traits and yield-related QTL confirms the role of plant development in yield determination, including the duration of flowering (Bourion et al., 2002).

The resolution of a QTL analysis is usually not adequate to distinguish whether coincident QTL have a causal relationship, are due to pleiotropic effects of single genes, or are due to the occurrence of linked genes. Coincidence of QTL for related traits may suggest that the genetic loci involved have an overall effect on aspects of plant growth and/or development related to seed yield or regulate specific processes, for example, the trade-off between seed weight and number. Added support is given to the hypothesis that trait variation attributed to coincident QTL is due to a causal relationship when the traits are strongly correlated and the allelic effects of coincident QTL are in the same direction as trait correlations.

Candidate Genes
A primary aim of developing the consensus map of the pea genome has been to integrate classical genetic and molecular linkage maps (Weeden et al., 1998). Since the Px4 linkage map is anchored to the consensus pea map, the locations of QTL identified in this study can be compared with the locations of loci identified by classical genetic methods that are likely to be involved in fundamental processes relevant to seed yield. As a consequence, hypotheses may be developed regarding the role of natural allelic variation at classical genetic loci in quantitative trait variation. The linkage of QTL with genes involved in determination of photoperiod response and flowering time, and with genes encoding proteins involved in sucrose and starch metabolism is discussed.

Flowering time in pea has been characterized extensively (Weller et al., 1997). Major genetic loci involved in the control of flowering time include Lf, E, Hr, Sn, Ppd, and Dne. Other quantitative gene systems may also be involved in controlling flowering time (Murfet, 1990). Three of the QTL for developmental traits are associated with genomic regions where genetic loci affecting flowering have also been mapped. For example, Ppd (photoperiod) and Lf (late flowering) both map on LG II (Arumingtyas and Murfet, 1994; Murfet, 1971). Ppd is associated with the same genomic regions as QTL num2.1 and tnn2.1. Lf is associated with the genomic region containing nff2.1. On LG III, a cluster of QTL (tnn3.1, nfn3.1, nff3.1, yld3.1, num3.1, and hi3.1) occurs in the chromosomal region that carries Dne (daylength neutral). Therefore, the QTL in these three regions may detect allelic variation at Ppd, Lf, and Dne. The remaining QTL for developmental traits may detect the quantitative gene systems referred to by Murfet (1990) that have not been characterized previously using single-gene approaches. In addition to the above, yld7.2 is associated with the region of LG VII that contains Sn (Weeden et al., 1998). Kelly and Spanswick (1997) used nearly isogenic lines to show that Sn locus, which is regulated by photoperiod, delays senescence and slows flower and fruit growth, reduces seed growth and net storage accumulation, but does not condition a difference in leaf photosynthetic rates or availability. They hypothesized that Sn may control assimilate availability to the seed. If the QTL yld7.2 does indeed detect natural allelic variation at Sn, then the physiological effect of this genetic locus under the long-day conditions of our field trials would be to delay senescence and to extend flowering and fruiting. However, no effect on TNN or NFN was associated with this genomic region.

A number of sequences involved in sucrose and starch metabolism have been mapped in the Px4 population. These include granule bound starch synthase I (GBSSI, Genbank accession X88789; PID5 in Gilpin et al., 1997), which is linked to P482 on LG II; SPS2 (Genbank accession AF322116), sucrose synthase (SUS3, rug4 locus, Genbank accession AJ102080), and ADP glucose pyrophosphorylase subunit L2 (agpL2, Genbank accession Y08728) on LG III; starch branching enzyme I (r locus, Genbank accession X80009) on LG V; and granule bound starch synthase II (GBSSII, Genbank accession X88790; PID18 in Gilpin et al., 1997) on the segment of LG VI that is not shown. Two QTL are associated with markers linked to these sequences: num2.1 with P482, which is linked to GBSSI; and yld4.1 with SUS3.


   SUMMARY
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
REFERENCES
 
Molecular linkage mapping provides opportunities for plant breeders to practice marker-assisted selection (MAS) as they develop new cultivars. As a result of this study, molecular markers associated with QTL involved in seed yield determination have been identified. However, implementation of MAS for seed yield in pea will undoubtedly be difficult. Clearly, seed yield is genetically and physiologically complex, involving multiple genomic regions and coincident QTL. Furthermore, implementing MAS for quantitatively inherited traits is difficult due to a number of factors associated with QTL mapping results, including the accuracy and precision of QTL estimates, as discussed by Schön et al. (2004). In the short term, however, QTL characterization provides breeders with knowledge of the genomic regions involved in yield determination. This awareness may help breeders understand or predict possible impacts of MAS strategies on seed yield, for example, when selecting for traits such as improved disease resistance.


   ACKNOWLEDGMENTS
 
Thanks to Myles Rea for assistance with the 2002–2003 field trial, and to Jeanne Jacobs, Huub Kerchoffs, and Tracy Williams for critical reading of the manuscript. The research was funded by the New Zealand Foundation for Research, Science, and Technology.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 
The research was funded by the New Zealand Foundation for Research, Science and Technology.

Received for publication July 13, 2004.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY
 REFERENCES
 




This article has been cited by other articles:


Home page
 Am. J. Bot.Home page
J. C. Burger, M. A. Chapman, and J. M. Burke
Molecular insights into the evolution of crop plants
Am. J. Botany, February 1, 2008; 95(2): 113 - 122.
[Abstract] [Full Text] [PDF]

Home page
 ANN BOT (LOND)Home page
N. F. Weeden
Genetic Changes Accompanying the Domestication of Pisum sativum: Is there a Common Genetic Basis to the 'Domestication Syndrome' for Legumes?
Ann. Bot., October 1, 2007; 100(5): 1017 - 1025.
[Abstract] [Full Text] [PDF]

Home page
 Plant Physiol.Home page
J. Burstin, P. Marget, M. Huart, A. Moessner, B. Mangin, C. Duchene, B. Desprez, N. Munier-Jolain, and G. Duc
Developmental Genes Have Pleiotropic Effects on Plant Morphology and Source Capacity, Eventually Impacting on Seed Protein Content and Productivity in Pea
Plant Physiology, June 1, 2007; 144(2): 768 - 781.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (8)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Timmerman-Vaughan, G. M.
Right arrow Articles by Wilson, D.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Timmerman-Vaughan, G. M.
Right arrow Articles by Wilson, D.
Agricola
Right arrow Articles by Timmerman-Vaughan, G. M.
Right arrow Articles by Wilson, D.
Related Collections
Right arrow Other Legumes
Right arrow Crop Genetics

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome