Phenotypic analysis
Wide phenotypic variation in response to peanut web blotch was observed among RILs under all the five environments tested in this study (Fig. 1). The distribution of the disease scale recorded during the years 2007 and 2008 was shown in Fig. 1A, whereas the distribution of the disease index recorded during the years 2012 and 2018 (both in the field and indoor) was shown in Fig. 1B-D.
Sequencing, SNP and bin markers discovery
A whole-genome resequencing strategy was applied to construct paired-end libraries for the parental lines and their 212 RIL progenies. The length of DNA fragments in the libraries was about 350bp. Approximately, 490 Gb of clean data (Q20>96%) were produced, resulting from 6,285 million reads, each with a length of 150 bp. In total, 600 million reads were generated for each of the two parents, whereas reads generated for each of the RILs varied from 22.37 to 24.83 million (Supplementary Table S1). Coverage rate, mapped reads rate, sequence depth and other results from alignment to the reference genome are shown in Supplementary Table S1. In particular, the coverage rate associated with the two parents Zheng8903 and Yuhua4 were 98.51% and 99.05%, respectively, whereas it ranged from 53% to 63.63% in the RIL population (Table S1). The sequencing depth was 35.23× for both parents, whereas it ranged from 1.31× to 1.46× for the RIL population. Originally, 636,831 SNPs were called from the 214 samples using the GATK protocol. Then, 556,615 SNPs were retained after filtering for low quality loci in the two parents, due to missing values, heterozygosis, depth < 10 and GQ < 20. Finally, 138,039 SNPs which were homozygous and polymorphic between two parents were used for further analyses.
Construction of physical recombination maps and high density genetic linkage map
To avoid errors caused by low coverage associated with RIL sequencing, a sliding window with 15 consecutive SNPs was used to find the more accurate recombined breakpoints. The physical recombination map of 212 RILs was constructed based on the recombination map of each progeny (Supplementary Figure S1). After that, all chromosomes of the 212 RILs were aligned and compared for the minimal of 100-kb intervals. As a result, a total of 3,634 bin markers for the 212 lines were obtained in this way, and the genotypes and physical locations of the bins are given in Supplemental Table S2.
The obtained 3,634 bin markers were used to construct a genetic linkage map by the software JoinMap®v5.0 [22]. Twenty linkage groups were generated and assigned to the 20 chromosomes of the cultivated peanut according to the physical positions. The total genome length was 1,817.91 cM and the marker density across the 20 linkage groups ranged from 0.39 to 0.66 cM with an average of 0.50 cM (Table 1). The LG16 had the lowest marker number (129) and the lowest genetic length (54.58 cM), while LG3 had the highest marker number (277) and the highest genetic length (135.61 cM) (Table 1 and Fig. 2). More than 97.5% of the inter-markers distance was lower than 3 cM. The highest inter-marker distance (16.06 cM) was associated with LG16 (Table 1).
QTL mapping and candidate genes prediction for peanut web blotch resistance
QTL mapping of peanut web blotch resistance was performed with MapQTL® v6.0 [23], using phenotypic data collected across five environments. Eight QTLs associated with peanut web blotch resistance, located in eight different LGs, were confirmed in at least two environments, explaining from 2.8 to 15.1% of phenotypic variation and displaying LOD values ranging from 1.32 to 7.45 (Table 2). Two QTLs (qWBRA04 and qWBRA14) located on LG04 and LG14 were significantly associated with resistance in all the five testing environments in this study (Table 2, Fig. 3 and Fig. 4) and explained more than 10% of phenotypic variation, indicating they are probably the major QTLs with stable expression. Except for qWBRA13 and qWBRA05, which were detected in four and two environments respectively, the other four QTLs qWBRA03, qWBRA16, qWBRA17, qWBRA19 were detected in three environments (Table 2). Absolute values displayed by the additive effect parameter ranged from 0.11 to 8.29, and were negative for the QTLs on LG4, LG5, LG13, LG14, LG19 (indicating that the favorable allele originates from the resistant parent Zheng8903), and positive for the QTLs on LG3, LG16 and LG17 (indicating that the favorable allele originates from the susceptible parent Yuhua4).
To identify candidate genes for peanut web blotch resistance, coding sequences in the genomic region associated with the QTLs qWBRA04 and qWBRA14 were examined for predicted function, according to the Arachis hypogaea cv. Tifrunner reference genome annotation database [18]. A total of 41 candidate genes were identified with a putative role in disease resistance (Table 3). In detail, the region of qWBRA04, spanning a linkage interval of 1.10 cM, corresponds to a physical interval of ~86kb and contains four nucleotide binding site-leucine rich repeat (NBS-LRR) genes. The genes Arahy.Q7VTCQ and Arahy.9YX67Z contain coiled-coil (CC) domains and the genes Arahy.SK6LYR and Arahy.1RZ0PJ contain Toll-interleukin receptor (TIR) domain (Table 3). The region of qWBRA14, spanning 0.48 cM, physically corresponds to ~2.8 Mb and contains 37 genes encoding disease resistance protein. Among them, 19 genes contain TIR domains and one gene contains a CC domain, whereas the remaining 17 genes encode other proteins with a putative role in disease resistance, such as a bZIP transcription factor and a WRKY transcription factor-like protein.