Two in one sweep: aluminum tolerance and grain yield in P-limited soils are associated to the same genomic region in West African Sorghum
© Leiser et al. 2014
Received: 20 February 2014
Accepted: 22 July 2014
Published: 12 August 2014
Sorghum (Sorghum bicolor L. Moench) productivity is severely impeded by low phosphorus (P) and aluminum (Al) toxic soils in sub-Saharan Africa and especially West Africa (WA). Improving productivity of this staple crop under these harsh conditions is crucial to improve food security and farmer’s incomes in WA.
This is the first study to examine the genetics underlying sorghum adaptation to phosphorus limitation in a wide range of WA growing conditions. A set of 187 diverse sorghum genotypes were grown in 29 –P and + P field experiments from 2006-2012 in three WA countries. Sorghum grain yield performance under –P and + P conditions was highly correlated (r = 0.85***). Significant genotype-by-phosphorus interaction was detected but with small magnitude compared to the genotype variance component. We observed high genetic diversity within our panel, with rapid linkage disequilibrium decay, confirming recent sequence based studies in sorghum. Using genome wide association mapping based on 220 934 SNPs we identified one genomic region on chromosome 3 that was highly associated to grain yield production. A major Al-tolerance gene in sorghum, SbMATE, was collocated in this region and SbMATE specific SNPs showed very high associations to grain yield production, especially under –P conditions, explaining up to 16% of the genotypic variance.
The results suggest that SbMATE has a possible pleiotropic role in providing tolerance to two of the most serious abiotic stresses for sorghum in WA, Al toxicity and P deficiency. The identified SNPs can help accelerate breeding for increased sorghum productivity under unfavorable soil conditions and contribute to assuring food security in WA.
KeywordsSorghum Phosphorus Aluminum Breeding Genetics West Africa
Sorghum (Sorghum bicolor L. Moench) is a staple crop of the Savannah Zone of West- and Central Africa, where it is cultivated in low input cropping systems . Limited P availability and aluminum (Al) toxicity in the soil are serious and frequent constraints to sorghum growth and productivity across sub-Saharan Africa, particularly in West Africa (WA) –. The adaptation of sorghum to these conditions is crucial for food security, and increasingly for farmer’s income. Sorghum breeding specifically under P-limited conditions, carried out in WA since the last decade, was shown to be necessary and feasible to obtain superior genotypes for these conditions . However, the genetic basis underlying grain yield performance under P-limited conditions in sorghum is not yet known. In contrast, breeding for Al-tolerance in sorghum has received much more attention and the underlying genetic mechanism has been identified. A gene in the multidrug and toxic compound extrusion (MATE) family, SbMATE, underlies the Alt SB locus on chromosome 3, which is mainly responsible for Al-tolerance in sorghum via Al-induced root citrate release . Al-toxicity inhibits root growth and function, which leads to severe crop-yield losses . Release of citrate and other organic anions into the rhizosphere prevents root damage by chelating Al3+. At the same time, citrate can mobilize P that is bound to soil clays by ligand exchange, dissolution and occupation of sorption sites, thus increasing P availability to the plants –. Hence, exudation of organic anions can lead to a higher P acquisition rate ,. Recent work showed that over-expression of citrate synthesis and malate transporter genes in different species resulted in improved Al-tolerance and enhanced P uptake under low P conditions –, supporting previous hypothesis that Al-tolerance and P uptake can be regulated by similar mechanisms . However, in soybean and rape, citrate release is predominantly induced by Al and to a much lesser extent by low P conditions –. Evaluation of sorghum genotypes under P-limited and slightly Al-toxic soils, conditions common in WA, could shed light on the possible pleiotropic role of SbMATE in Al tolerance and P-efficiency and provide new possibilities for enhancing grain yield of sorghum under the prevalent low-input conditions in WA.
We provide the first evidence for a genetic and possible molecular link between Al-tolerance and P-efficiency, opening possibilities for the use of molecular breeding tools to facilitate development of cultivars with superior performance for poor WA soil conditions. We conducted an unprecedented evaluation of WA sorghum genotypes, with 187 West African genotypes assessed for grain yield in 15 low P (–P) and 14 P-fertilized (+P) field experiments in Mali, Niger and Senegal from 2006-2012. Although fertilized, some of the + P fields showed also slight P limitations, due to the prevailing soil conditions in WA. The genetic architecture of grain yield under –P and + P conditions was investigated using both genome wide association mapping with 220 934 SNPs, obtained by genotyping-by-sequencing (GBS), and gene specific markers associated with Al-tolerance within Alt SB . Significant SNPs close to (<40 kb) and within Alt SB were associated to grain yield under P-limited conditions, suggesting that Alt SB has a pleiotropic role in providing tolerance to two of the most important abiotic constraints to sorghum productivity in WA, Al toxicity and P deficiency.
Performance across –P and + P environments
Variance components (±se) and broad-sense heritability (h 2 ) for grain yield of 187 sorghum genotypes analyzed separately within 15 –P and 14 + P environments and combined across P treatments
σ 2 G
701*** ± 90
1388*** ± 183
σ 2 GxE
423*** ± 31
988*** ± 70
σ 2 G
983*** ± 123
σ 2 GxE
503*** ± 35
σ 2 GxP
47*** ± 13
σ 2 GxExP
101*** ± 17
Association results of four SNP markers within the Alt SB locus with grain yield in + P, –P and –P/+P grain yield ratio across 187 sorghum genotypes
Genetic markers and linkage disequilibrium
Relatedness and population structure
The diversity panel used in this study consisted of five major groups based on kinship and population structure (Figure 2, Additional files 6 and 7). The clustering was primarily associated with differences for morphological race but not for geographic origin. The Guinea and Durra race genotypes formed two very distinct clusters, whereas the Caudatums could be considered intermediate between these two clusters. Generally the inter-racial genotypes were intermediate to their major race clusters, showing different levels of admixture (Additional file 7). A distinct sub-group of Guinea genotypes consisting of the Guinea-margaritiferums clustered closer to the majority of Durra rather than Guinea genotypes, but could only be detected in an admixture analysis using seven, and not the optimum five clusters (Figure 2, Additional file 7).
GWAS for grain yield
Grain yield BLUPs estimated across 15 –P, 14 + P, combined across all 29 environments and the grain yield ratios (–P/+P) were tested for association to 220 934 SNPs using a mixed model as described in . Several models were compared and the models correcting for kinship (K) and two to three principal components (PCAs) fitted our data best, had the lowest inflation-factor (λ) and therefore the lowest false-discovery rate.
Regions on chromosome 6 and 7 also showed strong associations to grain yield (Figure 3). Nine SNPs in gene Sb07g023120 and one SNP (S7_57976035) in gene Sb07g023130 of chromosome 7 were highly associated (–log10(p) > 4.5). None of the highly associated SNPs on chromosome 6 were located within coding regions. Grain yield BLUPs estimated only within –P or + P environments tended to show similar associations, but these single P-level associations had less power in detecting significant SNPs (Additional file 9). S3_71178053 was the most significant SNP for grain yield under both –P and + P conditions, whereas we observed differences across the other chromosomes depending on the P-level. SNPs on chromosome 6 played a more important role under –P conditions, whereas chromosome 10 exhibited a region highly associated to grain yield under + P conditions (Additional file 9). Examination of –P/+P grain yield ratios as an indicator for specific adaptation to –P conditions revealed one single SNP on chromosome 5, although not located in a coding region, and a second SNP on chromosome 1, S1_54947742 (–log10(p) > 4.5), which caused a missense variant in gene Sb01g032090, leading to an amino acid change from methionine to isoleucine (Additional file 10: Table S2).
Candidate gene association
The large grain yield and plant height reductions and the delay in flowering in the –P compared to the + P conditions indicate important growth differences between the two P treatments, that should permit a genetic study of adaptation to P-limited conditions . Although the –P trials always had significantly lower grain yields than the + P trials, some P limitations in + P trials cannot be discounted, especially for three + P trials due to the high P-retention in these low pH soils and late application of P fertilizer, thus reducing the impact of P fertilizer application on plant development and final grain yield . Despite an average 52% grain yield reduction in –P relative to + P conditions, the sorghum genotypes exhibited a strong correspondence between –P and + P grain yield performance (r = 0.85***). This suggests that WA sorghums have a general good adaptation to P-limited conditions and corresponds with previously reported higher P uptake and use efficiency of sorghum compared to other crops ,–. The strong correlation between –P and + P conditions and the small GxP variance component suggest that selection under + P conditions could result in gains under –P conditions. Nevertheless, the slightly higher broad sense heritability in –P (Table 1) and the specific adaptation of some genotypes towards –P conditions (Figure 1, Additional file 3) suggests that direct selection in –P conditions is more efficient for targeting P-limited production systems, as was shown by .
The phenological and racial diversity of our panel was so wide that the entire set would not normally be cultivated in the same environments. The Durra-race genotypes and very early maturing Caudatum-race genotypes are cultivated in the Sahelian zone whereas the Guinea-race genotypes are grown in the Sudanian zone of WA ,. Our field trials were therefore designed to evaluate accessions in their respective zone of adaptation. The relatively small GxE-interaction and lack of clear mega-environment delineations based on zones (Additional file 1: Table S1, Additional file 2) may be due, in part, to the experimental design in which only eight Guinea, Caudatum and Guinea-Caudatum genotypes with acceptable adaptation to both zones were common across all environments, hence only these common genotypes actually displayed the GxE-interaction pattern.
Genetic diversity among the 187 WA sorghum genotypes was high, with considerable separation of the three races, but no separation due to geographic origin. Studies showing clear geographic differentiation are mostly based on worldwide collections –, whereas studies considering germplasm from within a smaller region find mostly racial differences . A group of Guinea-margaritiferum accessions was clearly separated from the other races, but clustering closer to the Durra- rather than other Guinea-race accessions, confirming the genetic distinctness of this Guinea-sub-race ,,. Although there were clear racial differentiations, we also observed a high level of genetic admixture (Additional file 7) in our diversity panel, suggesting that considerable gene flow has occurred in the WA region and thus providing a high genetic diversity exploitable in GWAS. Furthermore, the rapid LD decay, in agreement with recent studies based on GBS  and re-sequencing data  of diverse sorghum genotypes additionally points to the high genetic diversity of our panel and its usefulness for GWAS with a mapping resolution on gene-level.
The –P specific adaptation and general good performance of the Durra genotypes (Figure 1, Additional file 3) makes this germplasm group particularly promising for use in future sorghum breeding efforts. Their genetic distinctness from the Guinea pool suggest, that Durra by Guinea-race crossing could be explored to achieve high heterosis and to obtain broader adaptation  if acceptable grain and glume types can be obtained from these inter-racial matings.
Genomic regions for grain yield under P-limited conditions
This study, the first to our knowledge, identifies regions in the sorghum genome related to grain yield under P-limited soils and a genomic region with a possible pleiotropic effect for Al-tolerance and grain yield productivity under P-limited conditions. Recent work show that over-expression of citrate synthesis and malate transporter genes improve Al-tolerance and P uptake under low P conditions in several species –, supporting the hypothesis that Al-tolerance and P uptake under low P conditions can be regulated by similar processes . However citrate release is predominantly induced by Al and to a much lesser extent by low P conditions –. Al-saturation values between 10-28% were found in six out of 16 field trials analyzed for Al-saturation (Additional file 11: Table S3), values which are at or near the threshold for limiting sorghum root growth due to Al-toxicity. Thus, although the Al-stress may have been mild, these conditions could have induced SbMATE expression that increased citrate release and benefitted P availability and retention of viable roots , thus leading to an increased P uptake. Furthermore, Al saturation shows generally a high spatial variation  and has been reported to have a major influence on crop growth in WA , therefore an impact of Al toxicity on crop growth in the other environments cannot be ruled out. This might have also lead to the observed strong association of grain yield from both P-treatments to the same genomic region.
The strong effect of SbMATE specific SNPs and their rather tight linkage to the other significant SNPs within the 78 kb region identified by GWAS suggests that grain yield under P-limited conditions is possibly controlled by SbMATE and likely other genes in this region. S3_71178053, the SNP with the strongest association to grain yield caused a missense mutation in gene Sb03g043930, leading to an amino-acid change from threonine to alanine (Additional file 10: Table S2). Since Sb03g043930 is an uncharacterized protein, it is not clear how it might be involved in grain yield production or SbMATE regulation. Recent evidence shows that the SbMATE expression in sorghum is regulated by additional factors  and thus its expression is more complex than previously thought. The linked SNPs identified in our study could be valuable for studying these background factors. The C2H2-type zinc-finger transcription factors (STOP 1 and 2) regulate MATE expression in A. thaliana,, and orthologous genes to STOP1 have been found in several other crops . Other C2H2-type zinc finger genes are also known to play a role in P adaptation and are mostly locally regulated (e.g. roots), whereas C3HC4-type zinc-finger genes are systematically regulated  and act as transcription factors in early low P adaptation  and other abiotic stress adaptation . SNP S3_71101374, which was highly associated to grain yield performance (Figure 4) and in strong linkage to AltSB_24804, is located close to Sb03g043850, which is predicted to be a C3HC4-type zinc-finger gene (Additional file 10: Table S2), hence possibly involved in adaptation to P-limited conditions. The highly associated AltSB specific SNP AltSB_24804, located in the 3´ region of SbMATE, was shown to be associated to Al-tolerance in another study, although showing rather small effects . The other three SNPs within or closer to SbMATE, especially AltSB_6083, which showed specific responses to –P conditions, are known to be highly responsible for Al-tolerance in sorghum , pointing to a possible pleiotropic regulation of Al-tolerance and specific –P adaptation. However, we did not evaluate this set of genotypes specifically for Al-tolerance in soil-based or hydroponic systems, but previous studies showed a very strong effect of AltSB_6083 on phenotypic Al-tolerance, that support this hypothesis. Although the frequency of these positive –P specific SNPs was low (9%) in our WA panel, the frequency appears to be much higher in the Guinea-race germplasm, with 95% of the genotypes carrying the positive SbMATE alleles being either pure Guinea race or Guinea introgressed genotypes. The low MAF and confounding of population structure might have limited a stronger association of these SNPs to low P grain yield . A similar allele frequency (0.1) of these SNPs and a higher frequency in WA Guinea sorghums was also found in a worldwide sorghum collection , hence pointing to this germplasm group as an important source for Al-tolerance in sorghum .
The rather high association on chromosome 7 to grain yield under both –P and + P conditions may have been caused by an unidentified foliar disease resistance gene. The SNP with the best association to grain yield was located within an uncharacterized protein Sb07g023120 but was also in strong LD (R2 > 0.85) with SNPs in Sb07g023130. Sb07g023130 is similar to NADPH HC toxin reductase, which confers specific resistance reaction to Helminthosporium (Cochliobolus carbonum) in maize . Although Cochliobolus carbonum is not known as a major sorghum disease in WA, it is known to cause foliar diseases in sorghum  and the HC toxin reductase mechanism is also conserved in other grasses and acts as resistance response to Cochliobolus carbonum. Although foliar disease scoring was not done in our study to enable further investigation, the effect on chromosome 7 being expressed under both + P and –P conditions would be consistent with some type of biotic resistance.
Specific adaptation to P-limited conditions was assessed by the –P versus + P ratio of grain yield performance. Although no genomic region was found with strong associations to –P/+P grain yield ratio, SNPs with –log10(p) > 4.3 may indicate the existence of mechanisms involved in specific –P adaptation other than SbMATE. One of these, SNP S5_2179409 on chromosome 5 was located in an intron region of gene Sb05g001996. Sb05g001996 is predicted to be of the rhamnogalacturonan lyase family, hence possibly involved in cell wall degradation by disrupting rhamnogalacturonan . This might imply that Sb05g001996 is involved in specific low P adaptation by cell wall degradation and consequently in P translocation at later developmental stages. Somewhat similarly, S1_54947742 causes a missense variant in Sb01g032090, which is predicted to be a papain family cysteine protease. Papain family cysteine proteases are known to be involved in programmed cell death, either due to senescence and therefore nutrient cycling or as pathogen defense ,. Phosphorous translocation from shoot to grain is known to be of major importance in sorghum, and P harvest index is increased under P-limited conditions . The existence of significant genotypic differences for specific –P grain yield, measured as –P/+P grain yield ratio (unpublished data), but no identification of a genomic region with clear effects for this specific –P/+P grain yield ratio may result from the existence of multiple adaptive mechanisms contributing to this complex trait.
Implications for applied breeding
Our test sites represent a wide range of soils and climate conditions relevant to sorghum production in WA. It is thus remarkable that we found individual SNPs accounting for up to 16% of the genotypic variation for grain yield. The significant GxP interaction, higher response to direct selection under –P conditions  and the low frequency of Al-tolerance in WA sorghum, indicate the need for selection efforts specifically targeting P-limited and high Al-saturation production environments, which are of major importance for WA sorghum producers. Our results suggesting that the same genomic region or even the same gene, SbMATE, possibly contribute to adaptation to both conditions focuses attention on this region. Currently we are evaluating further diversity and bi-parental mapping sets for performance under low P conditions to validate these findings. Once validated, it should be possible to incorporate the desired loci for both traits into elite breeding material with minimum linkage drag using marker assisted selection. Furthermore, the gene specific markers within Alt SB will make allele mining possible in large germplasm collections, hence increase allelic frequency and Al-tolerance of the currently used germplasm. These markers are already available on the Integrated Breeding Platform of the Generation Challenge Program [https://www.integratedbreeding.net/]. All other markers, once validated, will be also accessible on this platform to help sorghum breeders select promising genotypes for two traits in one sweep.
We provide the first evidence for a genetic and possible molecular link between Al-tolerance, P-efficiency and grain yield potential, opening possibilities for the use of molecular breeding tools to facilitate development of cultivars with superior performance under the predominant poor soil conditions of WA. The significant GxP interaction, higher response to direct selection under –P conditions and the low frequency of Al-tolerance in WA sorghum, indicate the need for selection efforts specifically targeting P-limited and high Al-saturation production environments. Our results point to the possibility that adaptation to both conditions is pleiotropically controlled by the same genomic region or even the same gene, SbMATE. The identified SNPs can help accelerate breeding for increased sorghum productivity under the two most important abiotic constraints to sorghum productivity in WA, Al toxicity and P deficiency, and contribute to assuring food security in WA.
A total of 187 sorghum genotypes from six West and Central African countries, consisting of researcher-bred and landrace varieties representing eight racial groups (Guinea, Durra, Caudatum and 5 intermediate groups; Additional file 4) and differing degrees of photoperiod sensitivity and stem internode lengths, were grown in paired –P (0 kg P ha−1 application) and + P (20-40 kg P ha−1 application, Additional file 11: Table S3) experiments in each location-year combination. A total of 29 rain fed field experiments (Additional file 1: Table S1), consisting of 15 –P and 14 + P environments, were successfully conducted in three countries (Mali, Niger, Senegal; Additional file 12) at five locations over seven years (2006-2012). Due to field size constraints (max. 320 plots per experiment) and specific adaptation requirements for the Sahelian and Sudanian zones, three separate series of trials were conducted with subsets of the 187 genotypes (Additional file 1: Table S1). At least eight common genotypes considered to have acceptable adaptation to both the Sudanian and Sahelian zones were included in all experiments to enable a combined analysis across all sites, hence fit BLUPs for each genotype across all environments and estimate genotype-by-environment interaction. The set 1 genotypes, consisting of 36 landrace and 34 research-bred varieties originating from the Sudanian zone, were tested at two locations in the Sudanian zone of Mali from 2006-2010 in the first series of trials comprising eight –P and seven + P environments . Set 2 genotypes, consisting of 54 landraces from the Sahelian zone and 28 early maturing research-bred varieties, were evaluated at four Sahelian zone locations in Mali, Niger and Senegal from 2010-2012 in the second series of trials, resulting in five pairs of –P and + P environments. Set 3 genotypes, consisting of Sudanian zone adapated genotypes (56 researcher bred and 4 landrace varieties), was grown at a single Sudanian zone location in Mali in 2011 and 2012, providing two pairs of –P and + P environments. Four late maturing genotypes in set 1 were dropped from the analyses due to severe grain losses due to sorghum midge (Stenodiplosis sorghicola) infestation. A number of the trials conducted were not used in this analysis due to crop failure, heterogeneous growth due to edaphic constraints other than low P, such as soil crusting and variable depth of soil hard-pan and planting errors.
All experiments were laid out in an α-design with four replicates. Plots consisted of two three-meter rows with 75-80 cm distance between rows and 30-40 cm between hills within rows. Hills were thinned to two plants, resulting in a total of 6-9 plants m−2. A single border row planted with a common genotype of medium height and maturity separated each test plot to minimize neighbor effects. Soil samples were taken from most field experiments and analyzed for plant available soil P (Bray-1P) , pH-H2O and subset of samples were tested for aluminum saturation (% Al of CEC) (Additional file 11: Table S3). The –P fields had an average Bray-1P value of 6.3 mg P kg−1soil, whereas the + P fields averaged 16.5 mg P kg−1soil. The –P and + P sites had similar pH values, which averaged 5.6.
Phenotypic statistical analysis
Each single environment was separately analyzed for grain yield using a mixed model with genotypes as fixed and replications and incomplete blocks nested within replications as random. When this model had repeatability values below 0.75, several spatial models were fitted (see ) and the optimum model was selected based on its AIC value. Predicted values and standard errors were computed for each genotype in each environment and were used in a combined weighted two-stage analysis  with genotypes and environments as random within one fertility level, yielding best linear unbiased predictors (BLUPs) for each genotype. Genotypic variances differed within individual environments (Additional file 1: Table S1) and correlations among the various environments likewise differed (Additional file 2). Several variance and co-variance models (unstructured, diagonal, factorial) were fitted to model specific variance co-variance structures but they did not converge (e.g. unstructured model) or did not improve the model fit based on AIC estimates (e.g. diagonal, factorial). Therefore instead of these models a compound symmetry model was used for our analyses. For a higher computational efficiency, the variance components were first estimated using the AI algorithm within Genstat 15 and then used as starting values running the same model but using the Fisher-scoring option to estimate broad-sense heritability based on . Further, a combined analysis across P-treatments was conducted with P-treatment as fixed and genotypes and environments (location x year-combination) as random to estimate the genotype-by-P interaction. Similar as in the analysis within each fertility level, we fitted several variance co-variance models. Convergence problems and lack of superiority of model fit based on AIC, lead us to a compound symmetry model.
Grain yield ratios (-P/+P) were estimated, whereby grain yield BLUPs of the –P trials were divided by the grain yield BLUPs of the + P trials. Genotypes showing a higher grain yield ratio are expected to be specifically adapted to –P conditions.
For possible mega-environment delineation and GxE visualization a genotype-genotype-by-environment (GGE) analysis was conducted for grain yield across all –P and + P sites using only grain yield data from the 8 common genotypes . All data analyses were done with the Genstat 15 software.
Total genomic DNA was extracted from a single 20 day old plant of each line by using DNeasy Plant Mini Kit (QIAGEN). Extracted DNA was checked for quality and quantity by Nanodrop and gel analyses. GBS libraries were prepared and analyzed at the Institute for Genomic Diversity (IGD), according to , using the enzyme ApeKI for digestion and creating a library with 188 unique barcodes. GBS libraries were sequenced on the Illumina HiSeq2000. The GBS analysis pipeline Version 3.0.121, an extension to the Java program TASSEL , was used to call SNPs from the sequenced GBS library. The sequenced tags were aligned to the sorghum reference genome BTx623 v1.0  and only tags with at least 10x coverage were retained. Ambiguous or heterozygous sites were set as missing SNPs and finally imputed with all other missing SNPs using NPUTE  for each chromosome separately. Imputation accuracy was on average above 96%. In total 308 623 SNPs were retrieved. After filtering for 5% minor allele frequency (MAF) using TASSEL, 220 934 SNPs were retained and used for further analysis. Gene specific markers for the Alt SB locus were created by EMBRAPA and converted to the KASPar system at LGC-Genomics . All genotypes were analyzed for these markers.
Pairwise linkage disequilibrium (LD) was calculated for each chromosome separately in TASSEL using a sliding window size of 50 SNPs. Average LD decay across each chromosome was calculated and plotted in R. LD within a specific genomic region was calculated and visualized using LDheatmap in R . Population structure was estimated using principal component analysis (PCA) as implemented in SNPrelate  and Admixture . Kinship (K) among genotypes was calculated using the EMMA algorithm within GAPIT . Allele frequencies of each site were calculated in TASSEL.
The grain yield BLUPs estimated across –P, +P, combined across all –P and + P sites and the grain yield ratios (–P/+P) were used in a genome wide association study (GWAS). GWAS was carried out in R using the package GenABEL  running several models using the polygenic function correcting only for kinship or for kinship and population structure using either one, two, three, five or ten PCAs. The best model was chosen based on the lambda estimate of the quantile-quantile plots of the expected versus the observed p-values, thus having the lowest genome-wide inflation. For most traits a model with one to three PCAs fitted best. Multiple testing was corrected using a Bonferroni threshold of p < 0.05 (p-value/SNPs). Association mapping for the Alt SB specific markers was performed in TASSEL using a compressed mixed linear model correcting for kinship and three PCAs . Possible causal changes of associated SNPs were tested by the Variant effect Predictor of the www.gramene.org website using the sorghum genome v.1.0.
The vital support of the technical staff of the Institut de Economie Rural, Mali, the Institut Sénégalais de Recherches Agricoles, Senegal, the Institut National de la Recherche Agronomique du Niger, Niger, and the International Crops Research Institute for the Semi-Arid Tropics, Mali, for conducting the yield trials is greatly appreciated. We thank Sharon E. Mitchell at the Institute for Genomic Diversity at Cornell University for the production and provision of the GBS marker data. The financial support of the McKnight Foundation Collaborative Crop Research Program, the Generation Challenge Program and the German Federal Ministry for Economic Cooperation and Development (BMZ) is gratefully acknowledged. The work was undertaken as a part of the CGIAR Dryland Cereals Research Program.
- Vom Brocke K, Trouche G, Weltzien E, Barro-Kondombo CP, Gozé E, Chantereau J: Participatory variety development for sorghum in Burkina Faso: farmers’ selection and farmers’ criteria. Field Crops Res. 2010, 119: 183-194. 10.1016/j.fcr.2010.07.005.View ArticleGoogle Scholar
- Buerkert A, Bationo A, Piepho H-P: Efficient phosphorus application strategies for increased crop production in sub-Saharan West Africa. Field Crops Res. 2001, 72: 1-15. 10.1016/S0378-4290(01)00166-6.View ArticleGoogle Scholar
- Kochian LV, Hoekenga OA, Pineros MA: How do crop plants tolerate acid soils? Mechanism of aluminum tolerance and phosphorus efficiency. Annu Rev Plant Biol. 2004, 55: 459-493. 10.1146/annurev.arplant.55.031903.141655.View ArticlePubMedGoogle Scholar
- Kochian LV: Plant nutrition: rooting for more phosphorus. Nature. 2012, 488: 466-467. 10.1038/488466a.View ArticlePubMedGoogle Scholar
- Leiser WL, Rattunde HFW, Piepho H-P, Weltzien E, Diallo A, Melchinger AE, Parzies HK, Haussmann BIG: Selection strategy for sorghum targeting phosphorus-limited environments in West Africa: analysis of multi-environment experiments. Crop Sci. 2012, 52: 2517-2527. 10.2135/cropsci2012.02.0139.View ArticleGoogle Scholar
- Magalhaes JV, Liu J, Guimarães CT, Lana UGP, Alves VMC, Wang Y-H, Schaffert RE, Hoekenga OA, Piñeros MA, Shaff JE, Klein PE, Carneiro NP, Coelho CM, Trick HN, Kochian LV: A gene in the multidrug and toxic compound extrusion (MATE) family confers aluminum tolerance in sorghum. Nat Genet. 2007, 39: 1156-1161. 10.1038/ng2074.View ArticlePubMedGoogle Scholar
- Kochian LV: Cellular mechanisms of aluminum toxicity and resistance in plants. Annu Rev Plant Biol. 1995, 46: 237-260. 10.1146/annurev.pp.46.060195.001321.View ArticleGoogle Scholar
- Zhang FS, Ma J, Cao YP: Phosphorus deficiency enhances root exudation of low-molecular weight organic acids and utilization of sparingly soluble inorganic phosphates by radish (Raghanus satiuvs L.) and rape (Brassica napus L.) plants. Plant Soil. 1997, 196: 261-264. 10.1023/A:1004214410785.View ArticleGoogle Scholar
- Neumann G, Römheld V: Root excretion of carboxylic acids and protons in phosphorus-deficient plants. Plant Soil. 1999, 211: 121-130. 10.1023/A:1004380832118.View ArticleGoogle Scholar
- Ma JF, Ryan PR, Delhaize E: Aluminium tolerance in plants and the complexing role of organic acids. Trends Plant Sci. 2001, 6: 273-278. 10.1016/S1360-1385(01)01961-6.View ArticlePubMedGoogle Scholar
- Vance CP, Uhde-Stone C, Allan DL: Phosphorus acquisition and use: critical adaptations by plants for securing a nonrenewable resource. New Phytol. 2003, 157: 423-447. 10.1046/j.1469-8137.2003.00695.x.View ArticleGoogle Scholar
- Richardson AE, Lynch JP, Ryan PR, Delhaize E, Smith FA, Smith SE, Harvey PR, Ryan MH, Veneklaas EJ, Lambers H, Oberson A, Culvenor RA, Simpson RJ: Plant and microbial strategies to improve the phosphorus efficiency of agriculture. Plant Soil. 2011, 349: 121-156. 10.1007/s11104-011-0950-4.View ArticleGoogle Scholar
- Delhaize E, Taylor P, Hocking PJ, Simpson RJ, Ryan PR, Richardson AE: Transgenic barley (Hordeum vulgare L.) expressing the wheat aluminium resistance gene (TaALMT1) shows enhanced phosphorus nutrition and grain production when grown on an acid soil. Plant Biotechnol J. 2009, 7: 391-400. 10.1111/j.1467-7652.2009.00403.x.View ArticlePubMedGoogle Scholar
- Wang Y, Xu H, Kou J, Shi L, Zhang C, Xu F: Dual effects of transgenic Brassica napus overexpressing CS gene on tolerances to aluminum toxicity and phosphorus deficiency. Plant Soil. 2013, 362: 231-246. 10.1007/s11104-012-1289-1.View ArticleGoogle Scholar
- Liang C, Piñeros MA, Tian J, Yao Z, Sun L, Liu J, Shaff J, Coluccio A, Kochian LV, Liao H: Low pH, aluminum, and phosphorus coordinately regulate malate exudation through GmALMT1 to improve soybean adaptation to acid soils. Plant Physiol. 2013, 161: 1347-1361. 10.1104/pp.112.208934.PubMed CentralView ArticlePubMedGoogle Scholar
- Nian H, Ahn SJ, Yang ZM, Matsumoto H: Effect of phosphorus deficiency on aluminium-induced citrate exudation in soybean (Glycine max). Physiol Plant. 2003, 117: 229-236. 10.1034/j.1399-3054.2003.1170210.x.View ArticleGoogle Scholar
- Dong D, Peng X, Yan X: Organic acid exudation induced by phosphorus deficiency and/or aluminium toxicity in two contrasting soybean genotypes. Physiol Plant. 2004, 122: 190-199. 10.1111/j.1399-3054.2004.00373.x.View ArticleGoogle Scholar
- Ligaba A, Shen H, Shibata K, Yamamoto Y, Tanakamaru S, Matsumoto H: The role of phosphorus in aluminium-induced citrate and malate exudation from rape (Brassica napus). Physiol Plant. 2004, 120: 575-584. 10.1111/j.0031-9317.2004.0290.x.View ArticlePubMedGoogle Scholar
- Liao H, Wan H, Shaff J, Wang X, Yan X, Kochian LV: Phosphorus and aluminum interactions in soybean in relation to aluminum tolerance: exudation of specific organic acids from different regions of the intact root system. Plant Physiol. 2006, 141: 674-684. 10.1104/pp.105.076497.PubMed CentralView ArticlePubMedGoogle Scholar
- Elshire RJ, Glaubitz JC, Sun Q, Poland JA, Kawamoto K, Buckler ES, Mitchell SE: A Robust, Simple Genotyping-by-Sequencing (GBS) approach for high diversity species. PLoS One 2011, 6:e19379.,PubMed CentralView ArticlePubMedGoogle Scholar
- Aulchenko YS, Ripke S, Isaacs A, Van Duijn CM: GenABEL: an R library for genome-wide association analysis. Bioinformatics. 2007, 23: 1294-1296. 10.1093/bioinformatics/btm108.View ArticlePubMedGoogle Scholar
- Caniato FF, Hamblin MT, Guimaraes CT, Zhang Z, Schaffert RE, Kochian LV, Magalhaes JV: Association mapping provides insights into the origin and the fine structure of the sorghum aluminum tolerance locus, AltSB. PLoS ONE 2014, 9:e87438.,PubMed CentralView ArticlePubMedGoogle Scholar
- Valluru R, Vadez V, Hash CT, Karanam P: A minute P application contributes to a better establishment of pearl millet (Pennisetum glaucum (L.) R. Br.) seedling in P deficient soils. Soil Use Manag. 2010, 26: 36-43. 10.1111/j.1475-2743.2009.00245.x.View ArticleGoogle Scholar
- Leiser WL, Rattunde HFW, Weltzien E, Haussmann BIG: Phosphorus uptake and use efficiency of diverse West and Central African sorghum genotypes under field conditions in Mali. Plant Soil. 2014, 377: 383-394. 10.1007/s11104-013-1978-4.View ArticleGoogle Scholar
- Atlin GN, Frey KJ: Predicting the relative effectiveness of direct versus indirect selection for oat yield in three types of stress environments. Euphytica. 1989, 44: 137-142. 10.1007/BF00022608.View ArticleGoogle Scholar
- Beebe S, Lynch J, Galwey N, Tohme J, Ochoa I: A geographical approach to identify phosphorus-efficient genotypes among landraces and wild ancestors of common bean. Euphytica. 1997, 95: 325-338. 10.1023/A:1003008617829.View ArticleGoogle Scholar
- Hammond JP, Broadley MR, White PJ, King GJ, Bowen HC, Hayden R, Meacham MC, Mead A, Overs T, Spracklen WP, Greenwood DJ: Shoot yield drives phosphorus use efficiency in Brassica oleracea and correlates with root architecture traits. J Exp Bot. 2009, 60: 1953-1968. 10.1093/jxb/erp083.View ArticlePubMedGoogle Scholar
- Parentoni S, de Souza C, de Carvalho Alves V, Gama E, Coelho A, de Oliveira A, Guimarães P, Guimarães C, Vasconcelos M, Pacheco CAP, Meirelles WF, de Magalhães JV, Moreira Guimarães LJ, da Silva AR, Ferreira Mendes F, Schaffert RE: Inheritance and breeding strategies for phosphorus efficiency in tropical maize (Zea Mays L.). Maydica. 2010, 55: 1-15.Google Scholar
- Ding G, Zhao Z, Liao Y, Hu Y, Shi L, Long Y, Xu F: Quantitative trait loci for seed yield and yield-related traits, and their responses to reduced phosphorus supply in Brassica napus. Ann Bot. 2012, 109: 747-759. 10.1093/aob/mcr323.PubMed CentralView ArticlePubMedGoogle Scholar
- Deu M, Sagnard F, Chantereau J, Calatayud C, Hérault D, Mariac C, Pham J-L, Vigouroux Y, Kapran I, Traore PS, Mamadou A, Gerard B, Ndjeunga J, Bezançon G: Niger-wide assessment of in situ sorghum genetic diversity with microsatellite markers. Theor Appl Genet. 2008, 116: 903-913. 10.1007/s00122-008-0721-7.View ArticlePubMedGoogle Scholar
- Sagnard F, Deu M, Dembélé D, Leblois R, Touré L, Diakité M, Calatayud C, Vaksmann M, Bouchet S, Mallé Y, Togola S, Traoré PCS: Genetic diversity, structure, gene flow and evolutionary relationships within the Sorghum bicolor wild–weedy–crop complex in a western African region. Theor Appl Genet. 2011, 123: 1231-1246. 10.1007/s00122-011-1662-0.View ArticlePubMedGoogle Scholar
- Billot C, Ramu P, Bouchet S, Chantereau J, Deu M, Gardes L, Noyer J-L, Rami J-F, Rivallan R, Li Y, Lu P, Wang T, Folkertsma RT, Arnaud E, Upadhyaya HD, Glaszmann J-C, Hash CT: Massive sorghum collection genotyped with SSR markers to enhance use of global genetic resources. PLoS One 2013, 8:e59714.,PubMed CentralView ArticlePubMedGoogle Scholar
- Bouchet S, Pot D, Deu M, Rami J-F, Billot C, Perrier X, Rivallan R, Gardes L, Xia L, Wenzl P, Kilian A, Glaszmann J-C: Genetic structure, linkage disequilibrium and signature of selection in sorghum: lessons from physically anchored DArT markers. PLoS One 2012, 7:e33470.,PubMed CentralView ArticlePubMedGoogle Scholar
- Morris GP, Ramu P, Deshpande SP, Hash CT, Shah T, Upadhyaya HD, Riera-Lizarazu O, Brown PJ, Acharya CB, Mitchell SE, Harriman J, Glaubitz JC, Buckler ES, Kresovich S: Population genomic and genome-wide association studies of agroclimatic traits in sorghum. Proc Natl Acad Sci. 2013, 110: 453-458. 10.1073/pnas.1215985110.PubMed CentralView ArticlePubMedGoogle Scholar
- Barro-Kondombo C, Sagnard F, Chantereau J, Deu M, Vom Brocke K, Durand P, Gozé E, Zongo JD: Genetic structure among sorghum landraces as revealed by morphological variation and microsatellite markers in three agroclimatic regions of Burkina Faso. Theor Appl Genet. 2010, 120: 1511-1523. 10.1007/s00122-010-1272-2.View ArticlePubMedGoogle Scholar
- Mace ES, Tai S, Gilding EK, Li Y, Prentis PJ, Bian L, Campbell BC, Hu W, Innes DJ, Han X, Cruickshank A, Dai C, Frère C, Zhang H, Hunt CH, Wang X, Shatte T, Wang M, Su Z, Li J, Lin X, Godwin ID, Jordan DR, Wang J: Whole-genome sequencing reveals untapped genetic potential in Africa’s indigenous cereal crop sorghum. Nat Commun. 2013, 4: 1-9.Google Scholar
- Rattunde HFW, Weltzien E, Diallo B, Diallo AG, Sidibe M, Touré AO, Rathore A, Das RR, Leiser WL, Touré A: Yield of photoperiod-sensitive sorghum hybrids based on Guinea-race germplasm under farmers’ field conditions in Mali. Crop Sci 2013, 53:2454.,View ArticleGoogle Scholar
- Voortman RL, Brouwer J, Albersen PJ: Characterization of spatial soil variability and its effect on millet yield on Sudano-Sahelian coversands in SW Niger. Geoderma. 2004, 121: 65-82. 10.1016/j.geoderma.2003.10.006.View ArticleGoogle Scholar
- Manu A, Wilding LP, Hossner LR, Pfordresher AA, Geiger SC: Soil parameters related to crop growth variability in Western Niger, West Africa. Soil Sci Soc Am J. 1996, 60: 283-288. 10.2136/sssaj1996.03615995006000010043x.View ArticleGoogle Scholar
- Melo JO, Lana UGP, Piñeros MA, Alves VMC, Guimarães CT, Liu J, Zheng Y, Zhong S, Fei Z, Maron LG, Schaffert RE, Kochian LV, Magalhaes JV: Incomplete transfer of accessory loci influencing SbMATE expression underlies genetic background effects for aluminum tolerance in sorghum. Plant J. 2013, 73: 276-288. 10.1111/tpj.12029.View ArticlePubMedGoogle Scholar
- Iuchi S, Koyama H, Iuchi A, Kobayashi Y, Kitabayashi S, Kobayashi Y, Ikka T, Hirayama T, Shinozaki K, Kobayashi M: Zinc finger protein STOP1 is critical for proton tolerance in Arabidopsis and coregulates a key gene in aluminum tolerance. Proc Natl Acad Sci. 2007, 104: 9900-9905. 10.1073/pnas.0700117104.PubMed CentralView ArticlePubMedGoogle Scholar
- Kobayashi Y, Ohyama Y, Kobayashi Y, Ito H, Iuchi S, Fujita M, Zhao C-R, Tanveer T, Ganesan M, Kobayashi M, Koyama H: STOP2 activates transcription of several genes for Al- and low pH-tolerance that are regulated by STOP1 in Arabidopsis. Mol Plant. 2014, 7: 311-322. 10.1093/mp/sst116.View ArticlePubMedGoogle Scholar
- Ohyama Y, Ito H, Kobayashi Y, Ikka T, Morita A, Kobayashi M, Imaizumi R, Aoki T, Komatsu K, Sakata Y, Iuchi S, Koyama H: Characterization of AtSTOP1 orthologous genes in tobacco and other plant species. Plant Physiol. 2013, 162: 1937-1946. 10.1104/pp.113.218958.PubMed CentralView ArticlePubMedGoogle Scholar
- Thibaud M-C, Arrighi J-F, Bayle V, Chiarenza S, Creff A, Bustos R, Paz-Ares J, Poirier Y, Nussaume L: Dissection of local and systemic transcriptional responses to phosphate starvation in Arabidopsis. Plant J. 2010, 64: 775-789. 10.1111/j.1365-313X.2010.04375.x.View ArticlePubMedGoogle Scholar
- Valdés-López O, Hernández G: Transcriptional regulation and signaling in phosphorus starvation: what about legumes?. J Integr Plant Biol. 2008, 50: 1213-1222. 10.1111/j.1744-7909.2008.00758.x.View ArticlePubMedGoogle Scholar
- Jung YJ, Lee IH, Nou IS, Lee KD, Rashotte AM, Kang KK: BrRZFP1 a Brassica rapa C3HC4-type RING zinc finger protein involved in cold, salt and dehydration stress. Plant Biol. 2013, 15: 274-283. 10.1111/j.1438-8677.2012.00631.x.View ArticlePubMedGoogle Scholar
- Caniato FF, Guimarães CT, Hamblin M, Billot C, Rami JF, Hufnagel B, Kochian LV, Liu J, Garcia AAF, Hash CT: The relationship between population structure and aluminum tolerance in cultivated sorghum. PLoS One 2011, 6:e20830.,PubMed CentralView ArticlePubMedGoogle Scholar
- Walton JD: HC-toxin. Phytochemistry. 2006, 67: 1406-1413. 10.1016/j.phytochem.2006.05.033.View ArticlePubMedGoogle Scholar
- Sugawara F, Strobel G, Strange RN, Siedow JN, Duyne GDV, Clardy J: Phytotoxins from the pathogenic fungi Drechslera maydis and Drechslera sorghicola. Proc Natl Acad Sci. 1987, 84: 3081-3085. 10.1073/pnas.84.10.3081.PubMed CentralView ArticlePubMedGoogle Scholar
- Sindhu A, Chintamanani S, Brandt AS, Zanis M, Scofield SR, Johal GS: A guardian of grasses: specific origin and conservation of a unique disease-resistance gene in the grass lineage. Proc Natl Acad Sci. 2008, 105: 1762-1767. 10.1073/pnas.0711406105.PubMed CentralView ArticlePubMedGoogle Scholar
- Mohnen D: Pectin structure and biosynthesis. Curr Opin Plant Biol. 2008, 11: 266-277. 10.1016/j.pbi.2008.03.006.View ArticlePubMedGoogle Scholar
- Lam E: Controlled cell death, plant survival and development. Nat Rev Mol Cell Biol. 2004, 5: 305-315. 10.1038/nrm1358.View ArticlePubMedGoogle Scholar
- Roberts IN, Caputo C, Criado MV, Funk C: Senescence-associated proteases in plants. Physiol Plant. 2012, 145: 130-139. 10.1111/j.1399-3054.2012.01574.x.View ArticlePubMedGoogle Scholar
- Bray R, Kurtz LT: Determination of total, organic, and available forms of phosphorus in soils. Soil Sci. 1945, 59: 39-46. 10.1097/00010694-194501000-00006.View ArticleGoogle Scholar
- Leiser WL, Rattunde HF, Piepho H-P, Parzies HK: Getting the most out of sorghum low-input field trials in West Africa using spatial adjustment. J Agron Crop Sci. 2012, 198: 349-359. 10.1111/j.1439-037X.2012.00529.x.View ArticleGoogle Scholar
- Cullis BR, Smith AB, Coombes NE: On the design of early generation variety trials with correlated data. J Agric Biol Environ Stat. 2006, 11: 381-393. 10.1198/108571106X154443.View ArticleGoogle Scholar
- Yan W, Kang MS: GGE Biplot Analysis: A Graphical Tool for Breeders, Geneticists, and Agronomists. CRC Press, Boca Raton 2002.View ArticleGoogle Scholar
- Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES: TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics. 2007, 23: 2633-2635. 10.1093/bioinformatics/btm308.View ArticlePubMedGoogle Scholar
- Paterson AH, Bowers JE, Bruggmann R, Dubchak I, Grimwood J, Gundlach H, Haberer G, Hellsten U, Mitros T, Poliakov A, Schmutz J, Spannagl M, Tang H, Wang X, Wicker T, Bharti AK, Chapman J, Feltus FA, Gowik U, Grigoriev IV, Lyons E, Maher CA, Martis M, Narechania A, Otillar RP, Penning BW, Salamov AA, Wang Y, Zhang L, Carpita NC, et al: The Sorghum bicolor genome and the diversification of grasses. Nature. 2009, 457: 551-556. 10.1038/nature07723.View ArticlePubMedGoogle Scholar
- Roberts A, McMillan L, Wang W, Parker J, Rusyn I, Threadgill D: Inferring missing genotypes in large SNP panels using fast nearest-neighbor searches over sliding windows. Bioinformatics. 2007, 23: i401-i407. 10.1093/bioinformatics/btm220.View ArticlePubMedGoogle Scholar
- Shin J-H, Blay S, McNeney B, Graham J: LDheatmap: an R function for graphical display of pairwise linkage disequilibria between single nucleotide polymorphisms. J Stat Softw. 2006, 16: 1-9.View ArticleGoogle Scholar
- Zheng X, Levine D, Shen J, Gogarten SM, Laurie C, Weir BS: A High-performance Computing Toolset for Relatedness and Principal Component Analysis of SNP Data. Bioinformatics. 2012, 28: 3326-3328. 10.1093/bioinformatics/bts606.PubMed CentralView ArticlePubMedGoogle Scholar
- Alexander DH, Novembre J, Lange K: Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 2009, 19: 1655-1664. 10.1101/gr.094052.109.PubMed CentralView ArticlePubMedGoogle Scholar
- Lipka AE, Tian F, Wang Q, Peiffer J, Li M, Bradbury PJ, Gore MA, Buckler ES, Zhang Z: GAPIT: genome association and prediction integrated tool. Bioinformatics. 2012, 28: 2397-2399. 10.1093/bioinformatics/bts444.View ArticlePubMedGoogle Scholar
- Zhang Z, Ersoz E, Lai C-Q, Todhunter RJ, Tiwari HK, Gore MA, Bradbury PJ, Yu J, Arnett DK, Ordovas JM, Buckler ES: Mixed linear model approach adapted for genome-wide association studies. Nat Genet. 2010, 42: 355-360. 10.1038/ng.546.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.