Exploring DNA variant segregation types in pooled genome sequencing enables effective mapping of weeping trait in Malus

Abstract

To unlock the power of next generation sequencing-based bulked segregant analysis in allele discovery in out-crossing woody species, and to understand the genetic control of the weeping trait, an F1 population from the cross 'Cheal's Weeping' × 'Evereste' was used to create two genomic DNA pools 'weeping' (17 progeny) and 'standard' (16 progeny). Illumina pair-end (2 × 151 bp) sequencing of the pools to a 27.1× (weeping) and a 30.4× (standard) genome (742.3 Mb) coverage allowed detection of 84562 DNA variants specific to 'weeping', 92148 specific to 'standard', and 173169 common to both pools. A detailed analysis of the DNA variant genotypes in the pools predicted three informative segregation types of variants: <lm×mm> (type I) in weeping pool-specific variants, and <lm×ll> (type II) and <hk×hk> (type III) in variants common to both pools, where the first allele is assumed to be weeping linked and the allele shown in bold is a variant in relation to the reference genome. Conducting variant allele frequency and density-based mappings revealed four genomic regions with a significant association with weeping: a major locus, Weeping (W), on chromosome 13 and others on chromosomes 10 (W2), 16 (W3), and 5 (W4). The results from type I variants were noisier and less certain than those from type II and type III variants, demonstrating that although type I variants are often the first choice, type II and type III variants represent an important source of DNA variants that can be exploited for genetic mapping in out-crossing woody species. Confirmation of the mapping of W and W2, investigation into their genetic interactions, and identification of expressed genes in the W and W2 regions provided insight into the genetic control of weeping and its expressivity in Malus.

Fig. 1.

A flowchart illustrating the major steps in MAFD and AFDDD mappings of the weeping phenotype. (This figure is available in color at JXB online.)

Fig. 3.

Schematic representations of the segregation of DNA variants linked to allele W in either phase under varying segregation types inferred for four of the five variant genotype groups: G1 (A), G2 (B), G3 (C), and G4 (D). Each segregation type is illustrated in a colored rectangle that includes the two parents at the top, four representative weeping progeny in the middle, and four standard progeny at the bottom. The long vertical lines in blue represent the chromosomal segment harboring W. The red and orange short vertical lines represent allele W and DNA variants in relation to the reference genome, respectively. The tree-like drawings with upward and downward ‘branches’ indicate standard and weeping tree phenotypes, respectively. The expected allele frequency (%) of DNA variants in the weeping and standard pools is given accordingly. In each segregation type denotation, the allele at the first position is designated as being linked to weeping phenotype in the seed parent ‘Cheal’s Weeping’ (e.g. letter ‘l’ in <lm×mm>), and those in bold are DNA variants in relation to the apple reference genome (e.g. letters ‘l’ in <lm×ll>). Segregation types informative for mapping allele W are shown in green rectangles (See Supplementary Table S5 for more details). ‘Common’, variants common to both pools; SP, standard pool; WP, weeping pool.

Fig. 4.

Distribution of allele frequency and density of variants specific to the weeping pool. (A) Distribution of allele frequency of 84562 variants. (B) Distribution of density of 18604 variants of allele frequency ranging from 40% to 60%. The colored bar at the bottom represents the assembled reference genome of 17 chromosomes as numbered. Based on z-score test, significant variant density peaks were detected on seven chromosomes, including 5 (z=4.0, P=6.4 × 10⁻⁵), 8 (z=3.4, P=6.7 × 10⁻⁴), 10 (z=4.2, P=2.6 × 10⁻⁵), 12 (z=3.4, P=6.7 × 10⁻⁴), 13 (z=5.1, P=3.4 × 10⁻⁶), 14 (z=7.9, P=0), and 16 (z=3.8, P=1.4 × 10⁻⁴).

Fig. 5.

Distribution of allele frequency directional difference (AFDD) and density of variants common to both pools on the apple reference genome. (A) Distribution of AFDD of the 6377 variants of AFDD ≥30 percentage points between the weeping and standard pools. (B) Distribution of density of the 6377 variants. The colored bar at the bottom represents the assembled reference genome of 17 chromosomes as numbered. Significant variant density peaks were identified on chromosomes 13 (z=8.7, P=0), 10 (z=5.7, P=1.2 × 10⁻⁸), 16 (z=4.4, P=1.1 × 10⁻⁵), and 5 (z=2.9, P=3.7 × 10⁻³).

Fig. 6.

Distribution of allele frequency directional difference (AFDD) and density of variants with AFDD ≥30 percentage points on chromosomes 13 (A, B), 10 (C, D), 16 (E, F), and 5 (G, H).

Fig. 7.

Assessing AFDDD mapping targeted variant genotype groups using the 6377 variants of AFDD ≥30 percentage points. (A–C) Number and frequency (%) of such variants observed in the five genotype groups at the genome scale (A), on chromosome 13 (B), and in the W region (C). (D) Number and frequency (%) of all the 173169 variants common to both pools in the five genotype groups. G1, heterozygous in weeping (He-W)/heterozygous in standard (He-S); G2, homozygous in weeping (Ho-W)/He-S; G3, He-W/homozygous in standard (Ho-S); G4, Ho-W/Ho-S; and G5, ‘Complex’. (This figure is available in color at JXB online.)

Fig. 8.

Chromatogram of DNA sequences of parents ‘Cheal’s Weeping’ and ‘Evereste’ covering three single nucleotide variants (SNVs; indicated by the red box) of segregation types <hk×hk> (A) and <lm×ll> (B, C) in the W region on chromosome 13. The SNV genotypes in the two parents, the reference genome, and the weeping and standard pools are listed accordingly. (A) SNV at position 7923460 in gene LOC103452418. (B) SNV at position 8209678 in gene LOC103452141. (C) SNV at position 8210175 in gene LOC103452141. Letters in bold represent DNA polymorphism (variant) in relation to the reference.

Fig. 2.

Phenotypic evaluation of growth habit in populations ‘Cheal’s Weeping’ × ‘Evereste’ (A), NY-051 × ‘Louisa’ (B), and NY-011 × NY-100 (C) segregating for weeping phenotype. Chi-square tests (excluding the intermediates) showed that the segregation of weeping (-like) and standard (-like) growth habits fit the 1:1 ratio in all the three populations: ‘Cheal’s Weeping’ × ‘Evereste’, χ²=0.1111, P=0.74; NY-051 × ‘Louisa’, χ²=0.2687, P=0.60; and NY-011 × NY-100, χ²=0.4100, P=0.52. (This figure is available in color at JXB online.)

Fig. 9.

Confirmation of mapping of loci W and W2. (A) Analysis of four SSR markers Ch13-7641, Ch13-8181, Ch13-8547, and Ch13-9530 from the W region on chromosome 13 in population NY-051 × ‘Louisa’. The image shows the markers’ polyacrylamide gel electrophoresis profile in 38 of the 140 individuals. The SSR bands Ch13-7641-119bp, Ch13-8181-125bp, Ch13-8547-168bp (the vertical line between lanes 10 and 11 indicates that this marker was run in two gels), and Ch13-9530-175bp of ‘Louisa’ origin and linked to the weeping phenotype are indicated with an arrow. W, weeping; WL, weeping-like; S, standard; SL, standard-like; –, seedling tree was dead before phenotyping. (B–E) Weeping trait association of SSR markers Ch13-8547 (in the W region) and Ch10-20017 (in the W2 region) in populations ‘Cheal’s Weeping’ × ‘Evereste’ (B, C) and NY-011 × NY-100 (D, E). Marker alleles linked to weeping and standard are suffixed with ‘-W’ and ‘-S’, respectively. (E, F) Effect of genetic interactions between the alleles of W and those of W2 (deduced from marker alleles Ch13-8547-W and Ch10-20017-W, respectively) on the expressivity of the weeping phenotype in populations ‘Cheal’s Weeping’ × ‘Evereste’ (E) and NY-011 × NY-100 (F).