Identification and reproducibility of diagnostic DNA markers for tuber starch and yield optimization in a novel association mapping population of potato (Solanum tuberosum L.)

Abstract

SNPs in candidate genes Pain - 1, InvCD141 (invertases), SSIV (starch synthase), StCDF1 (transcription factor), LapN (leucine aminopeptidase), and cytoplasm type are associated with potato tuber yield, starch content and/or starch yield. Tuber yield (TY), starch content (TSC), and starch yield (TSY) are complex characters of high importance for the potato crop in general and for industrial starch production in particular. DNA markers associated with superior alleles of genes that control the natural variation of TY, TSC, and TSY could increase precision and speed of breeding new cultivars optimized for potato starch production. Diagnostic DNA markers are identified by association mapping in populations of tetraploid potato varieties and advanced breeding clones. A novel association mapping population of 282 genotypes including varieties, breeding clones and Andean landraces was assembled and field evaluated in Northern Spain for TY, TSC, TSY, tuber number (TN) and tuber weight (TW). The landraces had lower mean values of TY, TW, TN, and TSY. The population was genotyped for 183 microsatellite alleles, 221 single nucleotide polymorphisms (SNPs) in fourteen candidate genes and eight known diagnostic markers for TSC and TSY. Association test statistics including kinship and population structure reproduced five known marker-trait associations of candidate genes and discovered new ones, particularly for tuber yield and starch yield. The inclusion of landraces increased the number of detected marker-trait associations. Integration of the present association mapping results with previous QTL linkage mapping studies for TY, TSC, TSY, TW, TN, and tuberization revealed some hot spots of QTL for these traits in the potato genome. The genomic positions of markers linked or associated with QTL for complex tuber traits suggest high multiplicity and genome wide distribution of the underlying genes.

Fig. 1

Boxplots of the adjusted entry means for tuber starch content (TSC), tuber yield (TY), tuber starch yield (TSY), tuber number (TN), and tuber weight (TW) evaluated in two-year field trials in 191 tetraploid cultivars (CUL), 73 tetraploid breeding clones (BRE) and 16 Andean landraces (LAN) of the QUEST population

Fig. 2

Principal coordinate plots of the QUEST population including (a) and excluding (b) the landraces, based on 183 alleles at 29 microsatellite loci. Genotypes were separated by the first two principal coordinates (PC) which were calculated on the basis of Jaccard’s distances. Numbers in parentheses are the percentage of explained variance by the PC

Fig. 3

Phenotypic effects of genotype classes of SNPs StCDF1_snp1812 (a), SSIV_snp2679 (b), and cytoplasm type (c) in the QUEST(−LAN) population. Means (grey bars) and standard deviations of TSC, TY, and/or TSY are shown. Significant differences between genotype classes are indicated by a and b (ANOVA post hoc test LSD, p < 0.05). The number of individuals in each genotype class is indicated at the bottom of the bar. The three genotypes with the highest dosage (ATTT) of the allele StCDF1_T ₁₈₁₂ associated with increased TY and TSY were the variety Isla and the breeding clones 2001Q29-10 and 2003P54-4. The four genotypes with the highest dosage (AATT) of the allele SSIV_T ₂₆₇₉ associated with increased yield were the varieties Murato, Melody, Riviera, and Opal (Online Resource 1)

Fig. 4

Physical maps of the 12 potato chromosomes, showing to the right of each chromosome the genomic positions of the markers genotyped in the QUEST population, and to the left markers linked to previously mapped QTL for tuber starch content (TSC, specific gravity (SG) in Bonierbale et al. 1993), yield (TY), starch yield (TSY), tuber weight (TW), tuber number (TN), and tuberization (TZ, in vitro tuberization (ivt) and greenhouse tuberization (gt) in Šimko et al. 1999). Numbers in parenthesis after the markers are numerical codes for the corresponding reference: 1 (Bonierbale et al. 1993), 2 (Schäfer-Pregl et al. 1998), 3 (van den Berg et al. 1996), 4 (Šimko et al. 1999), 5 (Zhou et al. 2014), and 6 (Navarro et al. 2011). RFLP marker sequences were retrieved from the databases Sol Genomics Network (TG, CT, and CD markers, http://www.sgn.cornell.edu/) and GABI Primary Database (GP and CP markers, GluA, SK2, SbeI, AGPaseB, pat, prp1, http://www.gabipd.org/). Microsatellite primer sequences were retrieved from the literature cited in Materials and Methods. Gene sequences were obtained from GenBank entries (http://www.ncbi.nlm.nih.gov/). Sequences were BLASTed against the potato genome sequence (pseudomolecules v4.03, http://potato.plantbiology.msu.edu/blast.shtml). QTL are shown red next to the linked or associated marker locus or next to the brackets indicating genome segments harbouring the QTL. Loci associated with TSC, TY, TSY, TW, and/or TN in the QUEST(+LAN) population but not in the QUEST(−LAN) population (MM-PK model, p < 0.01, MAF >1 %, see Online Resource 5) are shown in blue letters. Loci associated with TSC, TY, and/or TSY in both the QUEST(−LAN) and QUEST(+LAN) population are shown in green letters (see Table 5). Genes with known function are in italics