Characterization of Capsicum annuum genetic diversity and population structure based on parallel polymorphism discovery with a 30K unigene Pepper GeneChip

Abstract

The widely cultivated pepper, Capsicum spp., important as a vegetable and spice crop world-wide, is one of the most diverse crops. To enhance breeding programs, a detailed characterization of Capsicum diversity including morphological, geographical and molecular data is required. Currently, molecular data characterizing Capsicum genetic diversity is limited. The development and application of high-throughput genome-wide markers in Capsicum will facilitate more detailed molecular characterization of germplasm collections, genetic relationships, and the generation of ultra-high density maps. We have developed the Pepper GeneChip® array from Affymetrix for polymorphism detection and expression analysis in Capsicum. Probes on the array were designed from 30,815 unigenes assembled from expressed sequence tags (ESTs). Our array design provides a maximum redundancy of 13 probes per base pair position allowing integration of multiple hybridization values per position to detect single position polymorphism (SPP). Hybridization of genomic DNA from 40 diverse C. annuum lines, used in breeding and research programs, and a representative from three additional cultivated species (C. frutescens, C. chinense and C. pubescens) detected 33,401 SPP markers within 13,323 unigenes. Among the C. annuum lines, 6,426 SPPs covering 3,818 unigenes were identified. An estimated three-fold reduction in diversity was detected in non-pungent compared with pungent lines, however, we were able to detect 251 highly informative markers across these C. annuum lines. In addition, an 8.7 cM region without polymorphism was detected around Pun1 in non-pungent C. annuum. An analysis of genetic relatedness and diversity using the software Structure revealed clustering of the germplasm which was confirmed with statistical support by principle components analysis (PCA) and phylogenetic analysis. This research demonstrates the effectiveness of parallel high-throughput discovery and application of genome-wide transcript-based markers to assess genetic and genomic features among Capsicum annuum.

Figure 1. SPPs identified in the Pun1 locus (Pepper GeneChip CAPS_Contig.2339).

At the top, the diagram represents an alignment of CAPS_Contig.2339 with GenBank sequences for Pun1(FJ755173.1, GU300812.1, AY819028.1, AY819029.1, AY819032.1, AB206919.1), pun1¹ which has a large 5' deletion (gb AY819031.1), AT3-2 (FJ687524.1) and Pepper Chip assembly unigenes with significant similarity (>80% identity, >50 nucleotides aligned). The number of SNPs per 50 bp window between CAPS_Contig.2339 and aligned sequences are indicated by color boxes with key shown above. Regions not aligning are indicated by black lines. Below are the (40, option1 or 43 option 2) C. annuum lines with SPP calls (red  =  A, blue  =  B) along the length of the CAPS_CONTIG.2339 from 5' to 3'shown left to right with positions shown above the allele calls. Black lines link the positions of the SPP calls to the alignment cartoon. SPPs at positions with no additional sequence information are indicated (ns).SPP positions, indicated in the top row of SPP calls, highlighted in blue were identified as duplicated sequences among the pepper assemblies and thus were removed prior to subsequent analyses.

Figure 2. Polymorphism among pungent and non-pungent lines on chromosome P2.

Minor allele frequency by cM position across (A) non-pungent and (B) pungent lines. An 8.75 cM region of P2 containing pun1¹ (↓) is monomorphic among the non-pungent lines.

Figure 3. Polymorphisms among 19 non-pungent, 21 pungent C. annuum and 3 non-annuum lines.

(A) Venn diagram depicting the number of SPPs and unigenes (shown in parentheses) polymorphic within each group out of 9,272 SPPs within 5,712 unigenes having >2 minor alleles across the 43 lines. (B) Allele frequency matrix for 1729 informative SPPs among pungent and non-pungent C. annuum lines. Numbers indicate SPPs found at each minor allele frequency pair. At the corners are shown the total number of SPPs for minor allele frequencies pairs divided by 0<0.25 and 0.25<0.5. At left and bottom are shown total SPPs at each minor allele frequency for pungent and non-pungent lines respectively.

Figure 4. Consensus tree and population substructure estimated from SPP markers.

At the top, a Fitch & Margoliash tree of 40 C. annuum lines, rooted with C. frutescens and C. Chinense. The majority-rule consensus cladogram (overall equal branch lengths) was generated from 13,621 SPP markers. Numbers associated with branches indicate percent support based on 7,500 bootstrap replicates. Branches with less than 50% support have been collapsed. At the bottom is shown C. annuum population substructure determined using Structure with 2,712 mapped SPP markers for K = 2 to K = 6. Each genotype is represented by a vertical column and genotypes are ordered according to the cladogram. Each color bar represents a different subpopulation and the proportion of a given variety's color bar represents the proportion that variety belongs to the corresponding subpopulation. The branches of the cladogram are colored according to the highest proportion subpopulation assignment when K = 6 with grey branches indicating highly admixed individuals, having no more than a 0.60 fraction assigned to any subpopulation. Grouping by common structure subpopulation constitution is indicated by colored border with assigned names shown below. Long Yellow Marconi and Lange Westlandse Rode are abbreviated L. Yellow Marconi and L. W. Rode, respectively.

Figure 5. Principle component analysis & phylogram derived from SPP markers.

(A) A representative Fitch and Margoliash phylogram with branch lengths reflecting phylogenetic distances based on Nei and Li genetic distances. Boxed clades and labels correspond to clusters based on Structure simulation at K = 6. Long Yellow Marconi and Lange Westlandse Rode are abbreviated L. Yellow Marconi and L. W. Rode, respectively. (B) Graph showing the coordinates of the first 3 principle components for each variety. Clusters are circled and labeled. Non-clustered lines PI 201234 and CM334 are labeled separately. In this view the Anaheim/Ancho, Mixed and Bell clusters, boxed in black, are difficult discriminate, therefore, this area was expanded in (C) which shows the arrangement of these clusters. Symbols correspond to those representing horticultural classifications in Figure 4.