The harvest index for many crops can be improved through introduction of dwarf stature to increase lodging resistance, combined with early maturity. The inbred line Shen5003 has been widely used in maize breeding in China as a key donor line for the dwarf trait. Also, one major quantitative trait locus (QTL) controlling plant height has been identified in bin 5.05–5.06, across several maize bi-parental populations. With the progress of publicly available maize genome sequence, the objective of this work was to identify the candidate genes that affect plant height among Chinese maize inbred lines with genome wide association studies (GWAS).
Methods and Findings
A total of 284 maize inbred lines were genotyped using over 55,000 evenly spaced SNPs, from which a set of 41,101 SNPs were filtered with stringent quality control for further data analysis. With the population structure controlled in a mixed linear model (MLM) implemented with the software TASSEL, we carried out a genome-wide association study (GWAS) for plant height. A total of 204 SNPs (P≤0.0001) and 105 genomic loci harboring coding regions were identified. Four loci containing genes associated with gibberellin (GA), auxin, and epigenetic pathways may be involved in natural variation that led to a dwarf phenotype in elite maize inbred lines. Among them, a favorable allele for dwarfing on chromosome 5 (SNP PZE-105115518) was also identified in six Shen5003 derivatives.
The fact that a large number of previously identified dwarf genes are missing from our study highlights the discovery of the consistently significant association of the gene harboring the SNP PZE-105115518 with plant height (P = 8.91e-10) and its confirmation in the Shen5003 introgression lines. Results from this study suggest that, in the maize breeding schema in China, specific alleles were selected, that have played important roles in maize production.
Citation: Weng J, Xie C, Hao Z, Wang J, Liu C, Li M, et al. (2011) Genome-Wide Association Study Identifies Candidate Genes That Affect Plant Height in Chinese Elite Maize (Zea mays L.) Inbred Lines. PLoS ONE6(12): e29229. https://doi.org/10.1371/journal.pone.0029229
Editor: Shin-Han Shiu, Michigan State University, United States of America
Received: May 20, 2011; Accepted: November 22, 2011; Published: December 28, 2011
Copyright: © 2011 Weng et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was supported by National Basic Research Program (2011CB100100) from the Ministry of Science and Technology of China and National Natural Science Foundation of China (30871535). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Attaining high and stable yield has been one of the major goals in the production of crops, including maize. Semi-dwarfism, an important agronomic trait that contributes to crop yield, improves harvest index and nitrogen response, and increases lodging resistance . For example, semi-dwarf genes, such as sd1 in rice and RHT in wheat, led to the first Green Revolution in the 1960s . Similarly, dwarf genes provided by the elite maize inbred line Shen5003 have been successfully used to develop other Chinese maize inbreds with reduced plant height . One Shen5003 derivative, Zheng58, for example, is the parent of the maize hybrid ZhengDan958, which has been grown in an extensive area totaling ∼35 Mha in China.
Plant height is a complicated quantitative trait that is controlled by a large number of genes. Defects in either signaling or biosynthesis of GA or brassinosteroids (BR) may give rise to a typical dwarf phenotype. Semi-dwarfism in rice, controlled by sd1, results from a deficiency of GA20-oxidase activity in the GA biosynthetic pathway . In wheat, semi-dwarfism can be controlled by the reducing-height (RHT) gene, which encodes a DELLA protein involved in the GA signaling pathway . Recent genetic analysis and molecular characterization of dwarf mutants in maize revealed that mutations in dwarf genes including d1, d2, d3, d5, d8, d9, An1, DWF1 and DWF4 , , , , , , , , , lead to a dwarf phenotype. In addition, auxin synthesis and signalling, and epigenetic networks, also play important roles in stem elongation. For example, a multidrug resistant-like (MDR-like) ABC transporter involved in polar auxin transport leads to reduced plant height in the maize br2 mutant . An epigenetic effect is illustrated by Epi-d1, a spontaneous rice mutant that displays a metastable dwarf phenotype caused by silencing of the DWARF1 (D1) gene, which encodes a GTP-binding protein involved in giberellin signalling . However, due to the yield loss caused by various reproductive abnormalities associated with these genes or mutants, they have been difficult to use in breeding for dwarf stature in maize .
To date, using different mapping populations, more than 219 QTLs for plant height have been reported across ten maize chromosomes (2010 December update to Gramene database). One major locus consistently detected between bins 5.05 and 5.06 reduces height of maize inbreds from different bi-parental populations developed from Shen5003 and derived inbreds , , , and also has the same effect in the authors' previous study . Among important loci, ph3, which is located on chromosome 5, accounts for 38.6% of the total phenotypic variation . This locus should play an important role in future breeding for dwarfness in maize in China.
Several approaches such as map-based cloning and association mapping have been successfully used to identify candidate genes. Teosinte branched1 (Tb1), which explains a large proportion of phenotypic variation in maize plant architecture, was first cloned by time-consuming positional cloning in maize . Association mapping proceeds by assessing genetic and phenotypic variation in a population to establish newly discovered linkages between markers and particular phenotypes . With the development of SNP assays and associated statistical methods, GWAS has also been used to scan for novel loci influencing human diseases , and is a useful adjunct to classical genetic mapping of quantitative traits in plants . In Arabidopsis, Aranzana et al. were able to identify genes controlling natural variation in flowering time and pathogen resistance using genome-wide polymorphisms among 95 accessions . Also, the power of GWAS for analyzing 107 phenotypes in Arabidopsis was demonstrated using 250,000 SNPs . In rice, approximately 3.6 million SNPs were used to confirm six loci harboring previously identified genes for color, amylose content, and grain shape among sequences of 517 landraces . Forty-eight genomic regions correlated to aluminum tolerance were identified by GWAS in a diverse collection of 383 rice accessions . In barley, the HvbHL H1 gene was fine-mapped by GWAS using a 1536-feature SNP array based on expressed sequence tags . As an outcrossing species, abundant diversity and rapid linkage disequilibrium (LD) decay make maize an ideal crop for association mapping. Using a nested association mapping (NAM) panel, 1.6 million HapMap SNPs  have been used to test each SNP for effects on several traits, including flowering time , southern leaf blight , and leaf architecture . Association analysis may also be used to identify traits controlled by selection-candidate genes . Two types of maize SNP arrays have been constructed and commercialized for GWAS using populations of elite maize inbreds. One is the Illumina oligo pool assay that was developed from 582 candidate genes with 1536 SNPs, mainly for drought-related loci , . Another is the MaizeSNP50 BeadChip with over 55,000 evenly spaced markers, covering two-thirds of predicted genes (19,350) in the maize genome with 1–17 SNPs/gene (Illumina, Inc.). Because LD decays more rapidly in outcrossing species than in selfing species, more markers were needed to scan the whole genome for associated loci in maize: 50,000 markers were required for elite maize lines, but 750,000 markers were required for diverse maize landraces .
So far, two kinds of genome-wide association panels have been constructed for association mapping studies in maize. The NAM panel, which comprises 5000 recombinant inbred lines derived from crossing the B73 reference line to 25 diverse inbred lines, was constructed to understand the molecular basis of phenotypic variation with great power . Another panel is comprised of elite maize inbred lines, including 632 inbred lines representing global maize diversity , panels representative of American and European diversity , and 375 lines of mainly Chinese germplasm (data unpublished). The population structure of 187 inbred lines commonly used in Chinese maize breeding programs provided the basis for association mapping in this study .
There is, however, little GWAS information for SNP markers associated with dwarf loci in panels of elite inbred Chinese maize lines and germplasm, particularly for the widely used inbred line Shen5003. Therefore, the objectives of this study were to: 1) carry out a GWAS for plant height including major dwarf loci using 41,101 well-selected SNPs covering the entire maize genome; 2) identify loci for plant height in Chinese elite maize inbred lines under three environmental conditions; and 3) explore the genetics of major alleles at dwarf loci on chromosome 5 in Shen5003 and six of its derivatives. Finally, approaches for identifying candidate genes using GWAS in maize, in comparison with selfing species, are discussed.
SNP calling results
Using the MaizeSNP50 BeadChip, a total of 50,195 SNPs (91.1%) were successfully called with less than 20% missing data across the 288 inbreds. Of these 50,195 SNPs, 2784 were monomorphic in all lines. The source sequences of the SNP regions were used for basic local alignment search tool (BLAST) comparisons with the MaizeGDB (B73 RefGen_v1) to update coordinate information. A total of 231 of the 50,195 successful assays did not have a BLAST match below the threshold of 1e−4 and 2945 markers harbored multiple loci on maize chromosomes. Thus, a subset of 44,235 SNPs (80%) with known physical position was generated in this study.
Seven inbred lines with a high heterozygosity rate (HR) (>10%) across 44,235 markers were discarded. The remaining 281 lines with an average heterozygosity of 0.5% were used for further analysis. For these samples, the reproducibility between replicates across all data points was about 99.9%, which was consistent with high-quality data for the maize B73 line (Illumina, Inc.). Subsequently, the error rate (ER) for the genotyping was reduced from 0.0873% to 0.0353% by repeating assays in four line and discarding SNPs with multiple loci (Table 1).
Linkage disequilibrium in association panels
All 44,235 of the SNPs that were mapped in silico to the maize physical map across ten chromosomes were used to determine the extent of LD in the association panel. As shown in Table 2, the LD estimate is about 27.7 kb across all chromosomes. A shorter LD (25.4 kb) estimate is observed on chromosome 6, but the longest (29.2 kb) was observed on chromosome 3.
Genetic feature identified for plant height
The mean heritability for plant height across three environmental conditions was estimated as 70.63%, suggesting that this trait would respond well to artificial phenotypic selection . Table 3 shows that there was greater phenotypic variation but a less significant genotype-by-environment effect in environment II. Phenotypic correlations for plant height among the three locations were all significant (P<0.01).
SNP markers and genes associated with plant height
Approximately every fifth or sixth SNP was selected according to physical position from a collection of 5000 SNP markers and used to estimate population structure and relative kinship (Figure S1), and the power of these SNPs was similar to 500 SSRs . Across three environments, taking the population structure into consideration, the mean value for −log10(P) was markedly decreased by about half (0.552551/0.973008) within a total of 41,101 SNPs, and the peaks for SNPs on chromosomes were reduced with MLM (Figure 1, S2). These data demonstrate that the excess of tenuous associations arising mostly from confounding effects of population structure were eliminated.
Four blocks (red boxes) harboring the candidate genes controlling plant height were identified across three environments (MAF≥0.05).
A total of 204 SNPs across 10 chromosomes were significantly associated with plant height (P ≤0.0001), among which SNP PZE-105115518 was the most significant (P = 8.91e-10). These 204 SNPs covered 105 genomic blocks harboring coding regions from which more than 225 cDNAs were identified. The data showed that 78.9% of the SNPs were surrounded by cDNAs and that eight SNPs were located in the fragments of genes predicted from the B73 reference sequence. Several hot spots, including bins 1.11, 2.09, 3.02, 4.05, 5.05, 5.06, 6.03, 6.04, 6.05, 9.07, and 10.03 were identified as containing more than three SNPs in the predicted genes (Table S1), probably due to the extended LD decay revealed by high-density markers. Notably, peak signals for GWAS loci closely linked to genes for a particular trait were significantly over-represented. Given our emphasis on the factors influencing plant height, four loci harboring candidate genes for plant height were located in bins 1.11, 5.05, 6.04, and 9.07. The gene (GenInfo Identifier, GI: 194692741) for SNP SYN21642 in bin 1.11 (−log10(P) = 5.26) was similar to AUXIN RESISTANT 1 in Arabidopsis . The SNP PZE-105115518 in bin 5.05 (−log10(P) = 9.05) is located within a gene (GI: 100273912) annotated as a hypothetical protein, related to an Arabidopsis DNA glycosylase, loss of function of which causes a semi-dwarf phenotype . SNP PZE-106064587 in bin 6.04 (−log10(P) = 5.38), is located upstream of a gene (GI: 194701035) similar to Gibberellin 20 oxidase. One homolog of the candidate gene (GI: 195636889), upstream of SNP PZE-109106743 (−log10(P) = 4.83), was annotated as a tetratricopeptide repeat (TPR) domain protein involved in cell division cycle in rice. Together, these prediction data showed that GA, auxin, and epigenetic pathways may all be involved in natural variation that led to a dwarf phenotype in elite maize inbred lines.
Introgressed fragments from Shen5003 on chromosome 5
Based on the pedigree of inbred lines, plant heights in six derivatives from Shen5003 were modified with this dwarf germplasm (Table 4). The introgressed fragments from Shen5003 could reduce plant height in these six derivatives from 10.5% to 25.77%, compared to the tall parent across all three environments tested (Table 4). Also, both the entire maize genome and chromosome 5 contain more alleles from Shen5003 than average in these derivatives (Figure 2). That is to say, several important traits, such as dwarfism, controlled by the introgressed fragments of this chromosome can be selected during the maize breeding process. Figure 3 shows that primarily four fragments between bins 5.05 and 5.06 from Shen5003 were transferred to the six derivatives. Most importantly, the introgressed fragment IV contains the SNP PZE-105115518, which is significantly associated with plant height (P = 8.91e-10) (Figure 1). No other genes with known dwarfing function were detected in these four fragments.
A, chromosome 5; B, Four fragments at bin 5.05–5.06 harboring dwarf loci from Shen5003. Red bars indicate the fragments or SNPs from Shen5003, gap is chromatin from other lines.
Genome wide association study
Maize is an ideal candidate system for the application of GWAS with panels of inbred lines, because it has rapid LD decay as an outcrossing species and a high-quality reference genome sequence for generating enough informative markers . In rice, the GW2 gene, controlling grain width (GW) and weight, was undetectable via significant associations in the panel of 517 rice landraces tested . However, two homologs of this gene, ZmGW2-CHR4 and ZmGW2-CHR5, were significantly associated with kernel width (KW), or one of three other yield-related traits (kernel length (KL), kernel thickness (KT), and one-hundred kernel weight (HKW)) . Previous studies have suggested that most SNPs associated with phenotype would be located very close to the causative genetic variant . In our association panel, a total of 204 SNPs (P≤0.0001) and 105 genomic regions related to plant height were pinpointed with 41,101 SNPs (minor allele frequency, MAF≥0.05) across three environments. According to previous studies, only four genomic blocks harbored the known candidate genes controlling plant height (Figure 1, Table S1), the heritability of which is 70.63% in this panel. In rice, association signals for six traits, controlled by major loci were located close to known genes that were previously identified . Similar findings were also identified in Arabidopsis thaliana , . However, association signals for loci controlling heading date (flowering time) were not revealed on a whole-genome scale , . These data suggest that the heritability of traits and locus-specific effects contributing to phenotypic variance should be the key factors in GWAS with maize inbred panels. In addition, for complex traits such as flowering time, a nested association mapping population has been constructed to identify evidence of numerous minor single-locus effects .
Genetic architecture of plant height
Based on extensive results from dwarf mutants, GA and BR are major factors determining plant height . Dwarfing effects of mutants in the GA biosynthesis and signaling pathways have been important for allowing manipulation of plant height during the first Green Revolution . For example, D1, the gibberellin-responding dwarf gene of maize, controls the three biosynthetic steps: from GA20 to GA1, from GA20 to GA5, and from GA5 to GA3 . The GA-insensitive dwarf genes D8 in maize and RHT in wheat are both DELLA proteins with GTP-binding roles in GA signalling . In our study, a total of 204 SNPs covering 105 genomic regions were significantly associated with plant height (P≤0.0001), and four known loci harboring candidate genes for plant height were located in bins 1.11, 5.05, 6.04, and 9.07, based on previous studies. However, only one SNP, PZE-106064587 in bin 6.04, was identified as closely linked to a predicted gene for a protein similar to gibberellin 20 oxidase. In addition, auxin signalling and epigenetic networks also play important roles in internode elongation. In the maize br2 mutant, height reduction resulted from the loss of function of a P-glycoprotein that modulates polar auxin transport in the maize stalk . Our results show a significant association between plant height and SNP SYN21642 in bin 1.11 across environments, located close to a gene similar to AUXIN RESISTANT 1 (AXR1). In the case of Arabidopsis axr1, the decrease in plant height is the consequence of a defect in auxin action . Also, joint mapping suggested that there is a stable QTL for plant height between bin 1.10 and bin 1.11 , , , . Moreover, SNP PZE-105115518 is located within a gene annotated as hypothetical protein related to the Arabidopsis DNA glycosylase, loss of function of which results a semi-dwarf phenotype . This suggests that this gene in bin 5.05 may also participate in control of plant stature through an epigenetic pathway and would be a good candidate for manipulation of plant height in maize. These data indicate that the genetic network for dwarfness is composed of multiple pathways and that these SNPs reflect the genetic variation existing in our association panel. The new loci identified here are favorable candidates for subsequent studies to further our understanding of the genetic control of plant architecture including height in maize.
Dwarf loci as revealed by the association panel
Breeding for semi-dwarfness was a key objective for high yield in the first Green Revolution , when it was discovered in maize, yield potential may be increased by reducing plant height and selecting for erect leaves . However, most maize dwarf mutants are short, compact plants with shortened internodes, short wide leaves, and short erect tassels . In this study, the effects of above-mentioned dwarf loci, such as d1, d2, d3, d5, d8, d9, An1, DWF1, and DWF4 , , , , , , , ,  were not detectable across our association panel. Furthermore, previous studies had indicated difficulties with anther extrusion or low yield for mutants in such key genes controlling plant height . It had been suggested that mutagenesis of the maize gene D8 would affect quantitative variation in maize plant height , however, no significant association between D8 polymorphisms and plant height were found in the DELLA region, although this region was completely conserved across the tested lines . Indeed, some mutations in key genes for plant height often cause extreme phenotypes in addition to reducing plant height and would be difficult to select by maize breeding and to exploit commercially.
To date, more than 219 quantitative trait loci for plant height have been detected (2010 December update to Gramene database), and a total of 204 SNPs were significantly associated with plant height (P≤0.0001) in this study. However, there are cases in which QTLs discovered in bi-parental mapping populations are not detected by GWAS. One reason for this is that the moderate number of founders has limited resolution to dissect low-frequency or subpopulation-specific alleles. For instance in rice, the peak signal for grain width was tied closely to a previously identified domestication gene, GW5 , whereas no significant association was found for the sub-population-specific GW2 gene  due to the lack of a wide range of alleles in theassociation panel . Interestingly, two maize homologs of the GW2 gene, ZmGW2-CHR4, and ZmGW2-CHR5, have been significantly associated with kernel-related traits . Meanwhile, the GWAS that used 383 diverse rice accessions discovered 48 regions associated with Al tolerance, most of which were sub-population specific . Other reasons for lack of detection of some alleles may be that LD may diminish rapidly with distance in some chromosomal regions, or that a set of 41,101 SNPs is unlikely to capture all of the haplotypes present in diverse maize inbred lines. Figure 4 indicates that more than 50% of the gaps between two adjacent SNPs are less than 10 kb, and the average marker density is one marker every 50 kb (Table S2). In commercial inbred lines, LD decay may be slower and distance between markers may extend to more than 100 kb , while the average LD decay in our association panel is about 27.7 kb across ten chromosomes. Also, previous studies showed that the LD decay distance ranged from 5 to 10 kb among chromosomes in a set of global maize inbred lines , to 250 bp within the PSY2 gene in maize , which indicated that a higher density of markers should be used to adequately cover the whole genome in GWAS for certain populations. Finally, incorrect prediction, or imputation, of unobserved genotypes among SNPs in an association panel could reduce the accuracy of association estimates . With the panel used in this study, there was a 0.0873% error rate for 55,126 SNP markers, while there was 0.0353% error rate in the final dataset (Table 1). That is to say, SNP accuracy was high for association mapping in this study. Despite these limitations, four candidate genes prevalent in maize breeding have been identified and validations by joint mapping are in progress.
To summarize, GWAS identified more phenotype–genotype associations and provided higher resolution with high-density SNP markers, but QTL mapping identified sub-population-specific alleles not detected by GWAS.
Dwarf maize germplasm in China
Compared with other maize cropping system around the world, the summer maize zone in China, which accounts for 35.6% of the national maize-growing areas and 35% of national total output , is unusual for its annual double-cropping rotation with wheat. Usually, the number of growing days for varieties in this zone is less than 110 days. Therefore, the traits associated with plant height, such as short vegetative growth, high harvest index, early maturity, and lodging resistance may explain the use of varieties derived from one key dwarf donor line that has always existed in this specific growing region. The favorable dwarf allele harbored by Shen5003, which was derived from the American hybrid cultivar 3174, has been widely utilized in China, particularly in the Yellow and Huai River summer maize zones. And Ye478, the variety with a genetic background most similar to that of Shen5003 (Figure 2), has been used to develop more than 30 inbred lines and about 58 hybrids in China. Zheng58, one of these derived lines, is one of the parents of the commercial hybrid ZhengDan958, the most widely grown cultivar in China. The major QTL allele from Shen5003 between bins 5.05 and 5.06 that reduces plant height has been consistently detected across all three environments  in the Ye478×Dan340 F2∶3 population, in addition to other bi-parental populations , , . SNP PZE-105115518, identified in bin 5.05 in the introgressed fragment from Shen5003, bears significant association with plant height. Altogether, our work shows that a favorable allele for the dwarf trait on chromosome 5 of great agronomic importance to maize breeding and production may have been previously selected from hybrid 3174 through Shen5003 during maize breeding in China.
Materials and Methods
All research was conducted using existing databases and permission for use of farm was granted when necessary by the Ministry of Agriculture of the People's Republic of China. In 2008, a subset of 190 of the 284 diverse inbred lines used for the association panel was grown across three locations, I, Gongzhuling, Jilin Province (43°30′N, 124°48′E), II, Shunyi, Beijing (40°07′N, 116°39′E) and III, Sanya, Hainan Province (18°45′N, 109°30′E). These locations were under the control of the Ministry of Agriculture of the People's Republic of China, and the required permission was obtained before these sites were used for planting.
Plant materials and phenotyping of plant height
For genotyping and SNP-calling, a total of 284 diverse inbred lines (including the lines published in our previous report  with partial results unpublished) were used, which represented six subpopulations identified by population structure analysis: BSSS (American BSSS including Reid), PA (group A germplasm derived from modern U.S. hybrids in China), PB (group B germplasm derived from modern US hybrids in China), Lan (Lancaster Surecrop), LRC (derivative lines from Lvda Red Cob, a Chinese landrace), and SPT (derivative lines from Si-ping-tou, a Chinese landrace). Four lines (Qi319, Huangzaosi, Ye478, and D340), as quality controls, were used in six independent BeadChip panels.
A subset of 190 of the 284 diverse inbred lines in the panel was grown across three locations. A randomized complete-block design with three replications was employed, in which each line was planted in a plot of 20 plants in a 4.5-m long row with 0.6-m spacing between rows. Normal agronomic practices for maize were used in field management. At maturity, data for plant height from the soil surface to the base of anthers was collected, and the mean value for ten plants from each line was used for further statistical analysis.
Broad-sense heritability (h2) for plant height across three environments was computed according to Knapp (1985) . The heritability was calculated as follows: h2 = σg2/(σg2+σgl2/n+σe2/nr), where σg2 is the genetic variance, σgl2 is genotype-by-environment interaction, σe2 is the error variance, r is the number of replications, and n is the number of locations. The estimates for σg2, σgy2, and σe2 were obtained by analysis of variance (ANOVA) using the general linear model procedure of the statistical software SPSS 12.0 (IBM SPSS Inc.). Correlation coefficients were obtained with the program PROC CORR of SAS 8.0 (SAS Inc., Cary, NC).
SNP genotyping was performed using the MaizeSNP50 BeadChip produced by Emei Tongde (Beijing). The SNP content featured on this chip, using 56,110 evenly spaced markers to cover the whole maize genome based on the B73 reference sequence, was selected from several public and private sources with 984 negative controls. These markers have been successfully validated in more than 30 diverse lines. DNA from 284 lines was extracted by a modified CTAB procedure according to Murray and Thompson (1980) , and the DNA quality for each sample was checked carefully before genotyping by gel-electrophoresis and spectrophotometer (Nanodrop 2000, Thermo Scientific). The DNA from each of four controls was divided into two samples, respectively. The SNP genotyping of 288 samples in six independent BeadChip panels was analyzed by the Infinium® HD assay ultra-protocol guide (Illumina, Inc.).
The selected 44,235 SNPs showed a wide distribution in minor allele frequencies, ranging from nearly monomorphic (MAF<0.5%) to equal allele frequency (MAF≈50%) across the 277 inbred lines (Figure S3). About 7.08% of SNPs have MAFs below 5%, with a roughly equal chance of being on any chromosome (Figure S4). As SNPs with low MAF usually produced unstable results in our preliminary data analysis, these markers (MAF<5%) were excluded from the association analysis, leaving 41,101 SNPs. Within this set of high-quality SNPs, the average marker density was one marker every 50 kb and 53.48% of SNPs were within a 10 kb interval of the neighboring markers (Figure 4).
Genotype-phenotype association mapping
The linkage disequilibrium measurement parameter r2 (r2≥0.1) was used to estimate LD between all SNPs with less than 20% missing data on each chromosome via the software PLINK . The alleles for all SNPs were calculated using PowerMarker 3.25 . The population structure and kinship information for 277 lines were estimated with a mixed linear model using the software STRUCTURE version 2.3  and SPAGeDi  with 5000 SNPs (MAF≥0.2), respectively. STRUCTURE was run to test K = 6 according to our previous study  three times with a burn-in period of 500,000 and 500,000 replications. The general linear model (GLM) and mixed linear model implemented in the program TASSEL version 3.0  were used for genome-wide association mapping with 41,101 SNPs (MAF≥0.05), from which SNPs with –log10(P)≥4 were selected for candidate gene analysis.
Candidate gene analysis
The sequence available for the maize line B73 provides a reference genome that can be used to analyze candidate genes . BLAST against MaizeGDB was performed with 120 bp source sequence for SNP probes (Illumina, Inc.). The 30 kb window (the average LD decay is about 27.7 kb) was selected to fall within the estimated window of LD decay in our association panel. The genes within this window size were identified through MaizeGDB according to the positions of the closest flanking SNPs (P≤0.0001) or supporting intervals (http://gbrowse.maizegdb.org/cgi-bin/gbrowse/maize_v1/). The function of cDNAs were predicted using the blastx program at the National Center for Biotechnology Information (NCBI) database (http://blast.ncbi.nlm.nih.gov/).
Population structure across 277 lines with 5000 SNPs (MAF≥0.2).
Genome-wide association studies on plant height with general linear model across three environments (MAF≥0.05).
Minor allelic frequency for 44,235 SNPs in maizeSNP50.
SNP distribution across the maize genome (MAF≥0.05).
The SNPs, harboring cDNA, were detected with MLM in each environment (−log10( P )≥4, MAF≥0.05).
The authors would like to give great thanks to Dr. Yunbi Xu from the International Maize and Wheat Improvement Center (CIMMYT) for his invaluable suggestions and revisions to this manuscript.
Conceived and designed the experiments: J. Weng CX XL SZ. Performed the experiments: CX J. Weng J. Wang ZH ML. Analyzed the data: ZH J. Weng CX CL. Contributed reagents/materials/analysis tools: CX ZH DZ LB. Wrote the paper: J. Weng XL.
- 1. Khush GS (2001) Green revolution: the way forward. Nat Rev Genet 2: 815–822.GS Khush2001Green revolution: the way forward.Nat Rev Genet2815822
- 2. Peng J, Richards DE, Hartley NM, Murphy GP, Devos KM, et al. (1999) ‘Green revolution’ genes encode mutant gibberellin response modulators. Nature 400: 256–261.J. PengDE RichardsNM HartleyGP MurphyKM Devos1999‘Green revolution’ genes encode mutant gibberellin response modulators.Nature400256261
- 3. Li Y, Wang TY (2010) Germplasm base of maize in China and formation of foundation parents. Journal of Maize Sciences 18(5): 1–8.Y. LiTY Wang2010Germplasm base of maize in China and formation of foundation parents.Journal of Maize Sciences18518
- 4. Monna L, Kitazawa N, Yoshino R, Suzuki J, Masuda H, et al. (2002) Positional cloning of rice semidwarfing gene, sd-1: “rice green revolution gene” encodes a mutant enzyme involved in gibberellin synthesis. DNA Research 9: 11–17.L. MonnaN. KitazawaR. YoshinoJ. SuzukiH. Masuda2002Positional cloning of rice semidwarfing gene, sd-1: “rice green revolution gene” encodes a mutant enzyme involved in gibberellin synthesis.DNA Research91117
- 5. Lawit SJ, Wych HM, Xu D, Kundu S, Tomes DT (2010) Maize DELLA proteins dwarf plant8 and dwarf plant9 as modulators of plant development. Plant and Cell Physiol 51: 1854–1868.SJ LawitHM WychD. XuS. KunduDT Tomes2010Maize DELLA proteins dwarf plant8 and dwarf plant9 as modulators of plant development.Plant and Cell Physiol5118541868
- 6. Liu TS, Zhang JP, Wang MY, Wang ZY, Li GF, et al. (2007) Expression and functional analysis of ZmDWF4, an ortholog of Arabidopsis DWF4 from maize (Zea mays L.). Plant Cell Reports 26: 2091–2099.TS LiuJP ZhangMY WangZY WangGF Li2007Expression and functional analysis of ZmDWF4, an ortholog of Arabidopsis DWF4 from maize (Zea mays L.).Plant Cell Reports2620912099
- 7. Spray CR, Kobayashi M, Suzuki Y, Phinney BO, Gaskin P, et al. (1996) The dwarf-1 (dt) mutant of Zea mays blocks three steps in the gibberellin-biosynthetic pathway. Proc Natl Acad Sci USA 93: 10515–10518.CR SprayM. KobayashiY. SuzukiBO PhinneyP. Gaskin1996The dwarf-1 (dt) mutant of Zea mays blocks three steps in the gibberellin-biosynthetic pathway.Proc Natl Acad Sci USA931051510518
- 8. Fujioka S, Yamane H, Spray CR, Gaskin P, Macmillan J, et al. (1988) Qualitative and quantitative analyses of gibberellins in vegetative shoots of normal, dwarf-1, dwarf-2, dwarf-3, and dwarf-5 seedlings of Zea mays L. Plant Physiol 88: 1367–1372.S. FujiokaH. YamaneCR SprayP. GaskinJ. Macmillan1988Qualitative and quantitative analyses of gibberellins in vegetative shoots of normal, dwarf-1, dwarf-2, dwarf-3, and dwarf-5 seedlings of Zea mays L.Plant Physiol8813671372
- 9. Winkler RG, Helentjaris T (1995) The maize dwarf3 gene encodes a cytochrome P450-mediated early step in gibberellin biosynthesis. The Plant Cell 7: 1307–1317.RG WinklerT. Helentjaris1995The maize dwarf3 gene encodes a cytochrome P450-mediated early step in gibberellin biosynthesis.The Plant Cell713071317
- 10. Ogawa M, Kusano T, Koizumi N, Katsumi M, Sano H (1999) Gibberellin-responsive genes: high level of transcript accumulation in leaf sheath meristematic tissue from Zea mays L. Plant Mol Biol 40: 645–657.M. OgawaT. KusanoN. KoizumiM. KatsumiH. Sano1999Gibberellin-responsive genes: high level of transcript accumulation in leaf sheath meristematic tissue from Zea mays L.Plant Mol Biol40645657
- 11. Harberd NP, Freeling M (1989) Genetics of dominant gibberellin-insensitive dwarfism in maize. Genetics 121: 827–838.NP HarberdM. Freeling1989Genetics of dominant gibberellin-insensitive dwarfism in maize.Genetics121827838
- 12. Bensen RJ, Johal GS, Crane VC, Tossberg JT, Schnable PS, et al. (1995) Cloning and characterization of the maize An1 gene. The Plant Cell 7: 75–84.RJ BensenGS JohalVC CraneJT TossbergPS Schnable1995Cloning and characterization of the maize An1 gene.The Plant Cell77584
- 13. Tao YZ, Zheng J, Xu ZM, Zhang XH, Zhang K, et al. (2004) Functional analysis of ZmDWF1, a maize homolog of the Arabidopsis brassinosteroids biosynthetic DWF1/DIM gene. Plant Sci 167: 743–751.YZ TaoJ. ZhengZM XuXH ZhangK. Zhang2004Functional analysis of ZmDWF1, a maize homolog of the Arabidopsis brassinosteroids biosynthetic DWF1/DIM gene.Plant Sci167743751
- 14. Multani DS, Briggs SP, Chamberlin MA, Blakeslee JJ, Murphy AS, et al. (2003) Loss of an MDR transporter in compact stalks of maize br2 and sorghum dw3 mutants. Science 302: 81–84.DS MultaniSP BriggsMA ChamberlinJJ BlakesleeAS Murphy2003Loss of an MDR transporter in compact stalks of maize br2 and sorghum dw3 mutants.Science3028184
- 15. Doebley J, Stec A, Hubbard L (1997) The evolution of apical dominance in maize. Nature 386: 485–488.J. DoebleyA. StecL. Hubbard1997The evolution of apical dominance in maize.Nature386485488
- 16. Johnson EC, Fischer KS, Edmeades GO, Palmer FE (1986) Recurrent selection for reduced plant height in lowland tropical maize. Crop Sci 26: 253–260.EC JohnsonKS FischerGO EdmeadesFE Palmer1986Recurrent selection for reduced plant height in lowland tropical maize.Crop Sci26253260
- 17. Cao YG, Wang GY, Wang SC, Wei YL, Lu J, et al. (2000) Construction of a genetic map and location of quantitative trait loci for dwarf trait in maize by RFLP markers. Chinese Science Bulletin 45(3): 247–250.YG CaoGY WangSC WangYL WeiJ. Lu2000Construction of a genetic map and location of quantitative trait loci for dwarf trait in maize by RFLP markers.Chinese Science Bulletin453247250
- 18. Yang JP, Rong TZ, Xiang DQ, Tang HT, Huang LJ, et al. (2005) QTL mapping of quantitative traits in maize. Acat Agronomica Sinica 31(2): 188–196.JP YangTZ RongDQ XiangHT TangLJ Huang2005QTL mapping of quantitative traits in maize.Acat Agronomica Sinica312188196
- 19. Zhang ZM, Zhao MJ, Rong TZ, Pan GT (2007) SSR lingage map construction and QTL identification for plant height and ear height in maize (Zea mays L.). Acta Agronomica Sinica 33(2): 341–344.ZM ZhangMJ ZhaoTZ RongGT Pan2007SSR lingage map construction and QTL identification for plant height and ear height in maize (Zea mays L.).Acta Agronomica Sinica332341344
- 20. Yang XJ, Lu M, Zhang SH, Zhou F, Qu YY, et al. (2008) QTL mapping of plant height and ear position in maize (Zea mays L.). Hereditas 30(11): 1477–1486.XJ YangM. LuSH ZhangF. ZhouYY Qu2008QTL mapping of plant height and ear position in maize (Zea mays L.).Hereditas301114771486
- 21. Clark RM, Wagler TN, Quijada P, Doebley J (2006) A distant upstream enhancer at the maize domestication gene tb1 has pleiotropic effects on plant and inflorescent architecture. Nature Genet 38: 594–597.RM ClarkTN WaglerP. QuijadaJ. Doebley2006A distant upstream enhancer at the maize domestication gene tb1 has pleiotropic effects on plant and inflorescent architecture.Nature Genet38594597
- 22. Flint-Garcia SA, Thornsberry JM, Buckler ES (2003) Structure of linkage disequilibrium in plants. Ann Rev Plant Bio 54: 357–374.SA Flint-GarciaJM ThornsberryES Buckler2003Structure of linkage disequilibrium in plants.Ann Rev Plant Bio54357374
- 23. Klein RJ, Zeiss C, Chew EY, Tsai J-Y, Sackler RS, et al. (2005) Complement factor H polymorphism in age-related macular degeneration. Science 308: 385–389.RJ KleinC. ZeissEY ChewJ-Y TsaiRS Sackler2005Complement factor H polymorphism in age-related macular degeneration.Science308385389
- 24. Huang XH, Wei XH, Sang T, Zhao Q, Feng Q, et al. (2010) Genome-wide association studies of 14 agronomic traits in rice landraces. Nature Genet 42: 961–967.XH HuangXH WeiT. SangQ. ZhaoQ. Feng2010Genome-wide association studies of 14 agronomic traits in rice landraces.Nature Genet42961967
- 25. Aranzana MJ, Kim S, Zhao K, Bakker E, Horton M, et al. (2005) Genome-wide association mapping in Arabidopsis identifies previously known flowering time and pathogen resistance genes. PLoS Genet 1: e60.MJ AranzanaS. KimK. ZhaoE. BakkerM. Horton2005Genome-wide association mapping in Arabidopsis identifies previously known flowering time and pathogen resistance genes.PLoS Genet1e60
- 26. Atwell S, Huang YS, Vilhjalmsson BJ, Willems G, Horton M, et al. (2010) Genome-wide association study of 107 phenotypes in Arabidopsis thaliana inbred lines. Nature 465: 627–631.S. AtwellYS HuangBJ VilhjalmssonG. WillemsM. Horton2010Genome-wide association study of 107 phenotypes in Arabidopsis thaliana inbred lines.Nature465627631
- 27. Famoso AN, Zhao K, Clark RT, Tung C-W, Wright MH, et al. (2011) Genetic architecture of aluminum tolerance in rice (Oryza sativa) determined through genome-wide association analysis and QTL mapping. PLoS Genet 7: e1002221.AN FamosoK. ZhaoRT ClarkC-W TungMH Wright2011Genetic architecture of aluminum tolerance in rice (Oryza sativa) determined through genome-wide association analysis and QTL mapping.PLoS Genet7e1002221
- 28. Cockram J, White J, Zuluaga DL, Smith D, Comadran J, et al. (2010) Genome-wide association mapping to candidate polymorphism resolution in the unsequenced barley genome. Pro Natl Acad Sci USA 107: 21611–21616.J. CockramJ. WhiteDL ZuluagaD. SmithJ. Comadran2010Genome-wide association mapping to candidate polymorphism resolution in the unsequenced barley genome.Pro Natl Acad Sci USA1072161121616
- 29. Gore MA, Chia J-M, Elshire RJ, Sun Q, Ersoz ES, et al. (2009) A first-generation haplotype map of maize. Science 326: 1115–1117.MA GoreJ-M ChiaRJ ElshireQ. SunES Ersoz2009A first-generation haplotype map of maize.Science32611151117
- 30. Buckler ES, Holland JB, Bradbury PJ, Acharya CB, Brown PJ, et al. (2009) The genetic architecture of maize flowering time. Science 325: 714–718.ES BucklerJB HollandPJ BradburyCB AcharyaPJ Brown2009The genetic architecture of maize flowering time.Science325714718
- 31. Kump KL, Bradbury PJ, Wisser RJ, Buckler ES, Belcher AR, et al. (2011) Genome-wide association study of quantitative resistance to southern leaf blight in the maize nested association mapping population. Nature Genet 43: 163–168.KL KumpPJ BradburyRJ WisserES BucklerAR Belcher2011Genome-wide association study of quantitative resistance to southern leaf blight in the maize nested association mapping population.Nature Genet43163168
- 32. Tian F, Bradbury PJ, Brown PJ, Hung H, Sun Q, et al. (2011) Genome-wide association study of leaf architecture in the maize nested association mapping population. Nature Genet 43: 159–162.F. TianPJ BradburyPJ BrownH. HungQ. Sun2011Genome-wide association study of leaf architecture in the maize nested association mapping population.Nature Genet43159162
- 33. Weber AL, Zhao Q, McMullen MD, Doebley JF (2009) Using association mapping in teosinte to investigate the function of maize selection-candidate genes. PLoS ONE 4: e8227.AL WeberQ. ZhaoMD McMullenJF Doebley2009Using association mapping in teosinte to investigate the function of maize selection-candidate genes.PLoS ONE4e8227
- 34. Yan JB, Shah T, Warburton ML, Buckler ES, McMullen MD, et al. (2009) Genetic characterization and linkage disequilibrium estimation of a global maize collection using SNP markers. PLoS ONE 4: e8451.JB YanT. ShahML WarburtonES BucklerMD McMullen2009Genetic characterization and linkage disequilibrium estimation of a global maize collection using SNP markers.PLoS ONE4e8451
- 35. Setter TL, Yan JB, Warburton M, Ribaut JM, Xu YB, et al. (2011) Genetic association mapping identifies single nucleotide polymorphisms in genes that affect abscisic acid levels in maize floral tissues during drought. J Exp Bot 62(2): 701–716.TL SetterJB YanM. WarburtonJM RibautYB Xu2011Genetic association mapping identifies single nucleotide polymorphisms in genes that affect abscisic acid levels in maize floral tissues during drought.J Exp Bot622701716
- 36. McMullen MD, Kresovich S, Villeda HS, Bradbury P, Li H, et al. (2009) Genetic properties of the maize nested association mapping population. Science 325: 737–740.MD McMullenS. KresovichHS VilledaP. BradburyH. Li2009Genetic properties of the maize nested association mapping population.Science325737740
- 37. Reif J, Hamrit S, Heckenberger M, Schipprack W, Maurer H, et al. (2005) Trends in genetic diversity among European maize cultivars and their parental components during the past 50 years. Theor Appl Genet 111: 838–845.J. ReifS. HamritM. HeckenbergerW. SchipprackH. Maurer2005Trends in genetic diversity among European maize cultivars and their parental components during the past 50 years.Theor Appl Genet111838845
- 38. Xie CX, Zhang SH, Li MS, Li XH, Hao ZF, et al. (2007) Inferring genome ancestry and estimating molecular relatedness among 187 Chinese maize inbred lines. Journal of Genetics and Genomics 34(8): 738–748.CX XieSH ZhangMS LiXH LiZF Hao2007Inferring genome ancestry and estimating molecular relatedness among 187 Chinese maize inbred lines.Journal of Genetics and Genomics348738748
- 39. Huang H, Harding J, Byrne T, Huang N (1990) Quantitative analysis of correlations among flower traits in Gerbera hybrida Compositae. Theor Appl Genet 80: 559–563.H. HuangJ. HardingT. ByrneN. Huang1990Quantitative analysis of correlations among flower traits in Gerbera hybrida Compositae.Theor Appl Genet80559563
- 40. Yu JM, Zhang ZW, Zhu CS, Tabanao DA, Pressoir G, et al. (2009) Simulation appraisal of the adequacy of number of background markers for relationship estimation in association mapping. The Plant Genome 2: 63–77.JM YuZW ZhangCS ZhuDA TabanaoG. Pressoir2009Simulation appraisal of the adequacy of number of background markers for relationship estimation in association mapping.The Plant Genome26377
- 41. Lincoln C, Britton JH, Estelle M (1990) Growth and development of the axr1 mutants of Arabidopsis. The Plant Cell 2: 1071–1080.C. LincolnJH BrittonM. Estelle1990Growth and development of the axr1 mutants of Arabidopsis.The Plant Cell210711080
- 42. Gong ZZ, Morales-Ruiz T, Ariza RR, Rold-Arjona T, David L, et al. (2002) ROS1, a repressor of transcriptional gene silencing in Arabidopsis, encodes a DNA glycosylase/lyase. Cell 111: 803–814.ZZ GongT. Morales-RuizRR ArizaT. Rold-ArjonaL. David2002ROS1, a repressor of transcriptional gene silencing in Arabidopsis, encodes a DNA glycosylase/lyase.Cell111803814
- 43. Schnable PS, Ware D, Fulton RS, Stein JC, Wei F, et al. (2009) The B73 maize genome: complexity, diversity, and dynamics. Science 326: 1112–1115.PS SchnableD. WareRS FultonJC SteinF. Wei2009The B73 maize genome: complexity, diversity, and dynamics.Science32611121115
- 44. Li Q, Li L, Yang XH, Warburton M, Bai GH, et al. (2010) Relationship, evolutionary fate and function of two maize co-orthologs of rice GW2 associated with kernel size and weight. BMC Plant Biology 10: 143.Q. LiL. LiXH YangM. WarburtonGH Bai2010Relationship, evolutionary fate and function of two maize co-orthologs of rice GW2 associated with kernel size and weight.BMC Plant Biology10143
- 45. Gaut BS, Long AD (2003) The lowdown on linkage disequilibrium. The Plant Cell 15: 1502–1506.BS GautAD Long2003The lowdown on linkage disequilibrium.The Plant Cell1515021506
- 46. Wang YH, Li JY (2008) Molecular basis of plant architecture. Ann Rev Plant Bio 59: 253–279.YH WangJY Li2008Molecular basis of plant architecture.Ann Rev Plant Bio59253279
- 47. Beavis WD, Smith OS, Grant D, Fincher R (1994) Identification of quantitative trait loci using a small sample of topcrossed and F4 progeny from maize. Crop Sci 34: 882–896.WD BeavisOS SmithD. GrantR. Fincher1994Identification of quantitative trait loci using a small sample of topcrossed and F4 progeny from maize.Crop Sci34882896
- 48. Lubberstedt T, Melchinger AE, Schon CC, Utz HF, Klein D (1997) QTL mapping in testcrosses of European flint lines of maize: 1. Comparison of different testers for forage yield traits. Crop Sci 37: 921–931.T. LubberstedtAE MelchingerCC SchonHF UtzD. Klein1997QTL mapping in testcrosses of European flint lines of maize: 1. Comparison of different testers for forage yield traits.Crop Sci37921931
- 49. Khairallah MM, Bohn M, Jiang C, Deutsch JA, Jewell DC, et al. (1998) Molecular mapping of QTL for southwestern corn borer resistance, plant height and flowering in tropical maize. Plant Breeding 117: 309–318.MM KhairallahM. BohnC. JiangJA DeutschDC Jewell1998Molecular mapping of QTL for southwestern corn borer resistance, plant height and flowering in tropical maize.Plant Breeding117309318
- 50. Neuffer MG, Coe E, Wessler SR (1997) Mutants of maize. Cold Spring Harbor, , NY: Cold Spring Harbor Laboratory Press. MG NeufferE. CoeSR Wessler1997Mutants of maizeCold Spring Harbor, , NYCold Spring Harbor Laboratory Press
- 51. Izawa T, Takahashi Y, Yano M (2003) Comparative biology comes into bloom: genomic and genetic comparison of flowering pathways in rice and Arabidopsis. Curr Opin Plant Bio 6: 113–120.T. IzawaY. TakahashiM. Yano2003Comparative biology comes into bloom: genomic and genetic comparison of flowering pathways in rice and Arabidopsis.Curr Opin Plant Bio6113120
- 52. Weng JF, Gu SH, Wan XY, Gao H, Guo T, et al. (2008) Isolation and initial characterization of GW5, a major QTL associated with rice grain width and weight. Cell Research 18: 1199–1209.JF WengSH GuXY WanH. GaoT. Guo2008Isolation and initial characterization of GW5, a major QTL associated with rice grain width and weight.Cell Research1811991209
- 53. Song XJ, Huang W, Shi M, Zhu MZ, Lin HX (2007) A QTL for rice grain width and weight encodes a previously unknown RING-type E3 ubiquitin ligase. Nature Genet 39: 623–630.XJ SongW. HuangM. ShiMZ ZhuHX Lin2007A QTL for rice grain width and weight encodes a previously unknown RING-type E3 ubiquitin ligase.Nature Genet39623630
- 54. Ching A, Caldwell K, Jung M, Dolan M, Smith O, et al. (2002) SNP frequency, haplotype structure and linkage disequilibrium in elite maize inbred lines. BMC Genetics 3: 19.A. ChingK. CaldwellM. JungM. DolanO. Smith2002SNP frequency, haplotype structure and linkage disequilibrium in elite maize inbred lines.BMC Genetics319
- 55. Palaisa KA, Morgante M, Williams M, Rafalski A (2003) Contrasting effects of selection on sequence diversity and linkage disequilibrium at two phytoene synthase loci. The Plant Cell 15: 1795–1806.KA PalaisaM. MorganteM. WilliamsA. Rafalski2003Contrasting effects of selection on sequence diversity and linkage disequilibrium at two phytoene synthase loci.The Plant Cell1517951806
- 56. Bennetzen JL, Hake SC, Li JS (2009) Production, breeding and process of maize in China. Handbook of Maize. Its Biology: Springer New York. pp. 563–576.JL BennetzenSC HakeJS Li2009Production, breeding and process of maize in China. Handbook of MaizeIts BiologySpringer New York563576
- 57. Knapp SJ, Stroup WW, Ross WM (1985) Exact confidence intervals for heritability on a progeny mean basis. Crop Sci 25: 192–194.SJ KnappWW StroupWM Ross1985Exact confidence intervals for heritability on a progeny mean basis.Crop Sci25192194
- 58. Murray MG, Thompson WF (1980) Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research 8: 4321–4326.MG MurrayWF Thompson1980Rapid isolation of high molecular weight plant DNA.Nucleic Acids Research843214326
- 59. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira M, et al. (2007) PLINK: a tool set for whole-genome association and population-based linkage analyses. American Journal of Human Genetics 81: 559–575.S. PurcellB. NealeK. Todd-BrownL. ThomasM. Ferreira2007PLINK: a tool set for whole-genome association and population-based linkage analyses.American Journal of Human Genetics81559575
- 60. Liu KJ, Muse SV (2005) PowerMarker: an integrated analysis environment for genetic marker analysis. Bioinformatics 21: 2128–2129.KJ LiuSV Muse2005PowerMarker: an integrated analysis environment for genetic marker analysis.Bioinformatics2121282129
- 61. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945–959.JK PritchardM. StephensP. Donnelly2000Inference of population structure using multilocus genotype data.Genetics155945959
- 62. Hardy OJ, Vekemans X (2002) SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes 2: 618–620.OJ HardyX. Vekemans2002SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels.Molecular Ecology Notes2618620
- 63. Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, et al. (2007) TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics 23: 2633–2635.PJ BradburyZ. ZhangDE KroonTM CasstevensY. Ramdoss2007TASSEL: software for association mapping of complex traits in diverse samples.Bioinformatics2326332635