Genetic Architecture of Chile Pepper (Capsicum Spp.) QTLome Revealed Using Meta-QTL Analysis

Chile peppers (Capsicum spp.) are among the most important vegetable crops in the world due to their health-related, economic, and industrial uses. In recent years, quantitative trait loci (QTL) mapping approaches have been widely implemented to identify genomic regions affecting variation for different traits for marker-assisted selection (MAS) in peppers. Meta-QTL analysis for different traits in Capsicum remains lacking, and therefore it would be necessary to re-evaluate identied QTL for a more precise MAS for genetic improvement. We report the rst known meta-QTL analysis for diverse traits in the chile pepper QTLome. A literature survey using 29 published linkage mapping studies identied 766 individual QTL from ve different trait classes. A total of 311 QTL were projected into a consensus map. Meta-analysis identied 30 meta-QTL regions distributed across the 12 chromosomes of Capsicum. MQTL5.1 and MQTL5.2 related to Phytophthora capsici fruit and root rot resistance were delimited to < 1.0 cM condence intervals in chromosome P5. Candidate gene analysis for the anking sequences for the P5 meta-QTL revealed biological functions related to DNA repair and transcription regulation. Moreover, epigenetic mechanisms such as histone and RNA methylation and demethylation were predicted, indicating the potential role of epigenetics for P. capsici resistance. Allele specic SNP markers for the meta-QTL will be developed and validated using different breeding populations of Capsicum for MAS of P. capsici resistant lines. Altogether, results from metaQTL analysis for chile pepper QTLome rendered further insights into the genetic architecture of different traits for this valuable horticultural crop. on primary axis, and anthracnose resistance, among others. This demonstrated potential pleiotropy and/or effects of close linkage between the underlying QTL 44 . Such colocation of QTL related with diverse sets of traits for the identied meta-QTL across different chromosomes of chile pepper indicates the possibility of multi-trait improvement using genomic information from multiple linkage mapping studies. In the current study, meta-QTL analysis was used to dissect the genetic architecture of diverse traits in Capsicum. traits. Each QTL was characterized according to the number of lines used for QTL type of mapping population (e.g. F 2 , RIL, BC, DH), QTL name, trait, chromosome and linkage group designations, LOD, phenotypic variation explained (R 2 ), and chromosome positions (in cM). QTL with LOD scores of < 2.0 and with R 2 values not reported in the original study were excluded in further analysis. Genomewide association studies were not included in the literature review.


Introduction
The wealth of genomic information available for different crop species and diverse traits has expanded signi cantly in recent years due to the decreasing costs of high-throughput genotyping and the development of novel and powerful tools to dissect quantitative trait loci (QTL), which are genomic regions affecting variation for quantitative traits 1 . A 'QTLome' refers to the collection of QTL and their allelic variation shown to affect any quantitative trait for a given trait and species 2,3 . The rst step to interpret information from QTLome is meta-QTL analysis 2,4 . QTL metaanalysis integrates results from multiple QTL studies and could provide more insights into the genetic architecture of traits associated with different genomic regions 5 . The QTL identi ed from a group of different QTL using metaanalysis at a 95% con dence interval are called meta-QTL 6 . In meta-QTL analysis, various models can be implemented for identifying the consensus QTL from different studies thereby validating results and re ning positions of QTL in the consensus map 4,7 . Using information from meta-analysis could therefore allow for a more precise marker-assisted breeding for the genetic improvement of traits across different crops.
Chile peppers, belonging to the genus Capsicum and family Solanaceae, are one of the most important horticultural crops in the world due to their culinary uses, health bene ts, and economic impact. Nutrients such as vitamins A, C, and folate are present in varying degrees in peppers 8 . Chile peppers have been used to help combat chronic pain, and it was suggested that capsaicin, the compound found predominantly in pungent lines, can induce depletion in nerve sensory terminals, and is commonly found in pain relieving creams 9 . In the state of New Mexico in the United States, chile peppers are important cash crops for farmers 10 . A present disadvantage of the cultivation of Capsicum species, however, is an inadequate supply due to low yield from farmers. Low yield can be attributed to many factors, including pests and diseases, and undesired agronomic traits such as subpar fruit size or uneven biomass distribution [11][12][13] . In the past, marker-assisted selection through genetic mapping has been implemented in chile peppers to facilitate genetic improvement for different traits including yield and resistance to major diseases. While many QTL for diverse traits have been identi ed for chile peppers, there is no known report of meta-study across diverse traits in Capsicum. Most of the meta-QTL studies by far focused on diverse sets of traits across the Solanaceous relatives of chile peppers such as potatoes (Solanum tuberosum) 14 and tomatoes (S. lycopersicum) 15 ; and in major eld crops such as rice (Oryza sativa) 6,16 , wheat (Triticum aestivum) 17,18 , barley (Hordeum vulgare) 19 , maize (Zea mays) 20,21 , cotton (Gossypium spp.) 22 , and soybeans (Glycine max) 23 . In chile peppers, the only known report for meta-QTL to date was a study by Mallard et al. 24 , who identi ed meta-genomic regions for P. capsici resistance, designated as MetaPc5.1, MetaPc5.2, and MetaPc5.3, in chromosome P5. Meta-analysis indicated that MetaPc1 confers resistance against eight isolates, whereas MetaPc5.2 and MetaPc5.3 exhibit resistance for three isolates of P. capsici.
Given that meta-QTL studies remain lacking in chile peppers, it would be necessary to implement meta-analysis to determine signi cant genomic regions involved in variation for important traits. The current study aims to perform analysis of meta-QTL to identify regions associated with diverse traits for targeted genetic improvement in chiles.
Speci cally, our objectives were to (1) develop a consensus genetic map in chile peppers using SNP markers; (2) identify meta-QTL related with different traits such as heat levels (pungency), yield, adaptation, and resistance to diseases in peppers; (3) determine SNP markers linked to meta-QTL; and (4) identify candidate genes present in the meta-QTL regions. We report the rst known meta-QTL study of the chile pepper QTLome that could provide insights into the genetic architecture of different traits for this important horticultural crop.

Results
Classi cation of QTL for diverse traits in Capsicum spp. Overall, 766 individual QTL from 130 unique traits and ve different trait classes across the 12 chromosomes of chile pepper were identi ed. These results were based from 29 linkage mapping studies published within a 10-year period (2010-2020) ( Table 1; Fig. 1). Among these studies, majority (90%) of the reported QTL were identi ed using either a recombinant inbred line (RIL) or an F 2 biparental mapping population. The number of individuals used for QTL mapping ranged between 63 and 440. Marker types used in identifying QTL included ampli ed fragment length polymorphism (AFLP), simple sequence repeats (SSR), and single nucleotide polymorphism (SNP) markers, among others. Across the various linkage-mapping studies surveyed, among the most commonly used parents in creating a biparental mapping population included 'CM-334', a P. capsici resistant line for identi cation of QTL linked to chile pepper blight resistance; 'Bhut Jolokia', a 'superhot' chile pepper for the discovery of QTL associated with capsaicinoid content; and 'Yolo Wonder', a bell-pepper type, for linkage mapping of stem and fruit-related traits and disease resistance QTL.    Table S1). These ranged between ve and 318 candidate genes for Scaffold 3155.426970 (MQTL5.1) and Marker635294 (MQTL5.2). A wide range of biological functions for the candidate genes, including those related with DNA replication, repair, transcription regulation, phosphorylation, and glycosylation, among others, for MQTL5.1 and MQTL5.2 has been identi ed. Moreover, several functions related with epigenetic mechanisms such as RNA, DNA, tRNA, and histone methylation have been predicted. Among the candidate genes identi ed for MQTL5.2, PHT67052 and PHT65976, have functions related to transcription regulation, histone H3-K36 methylation, and histone lysine methylation; and regulation of gene expression, respectively, in C. annuum. Solyc11g033270.2.1 (MQTL5.2) has functions associated with regulation of defense response to bacterial and fungal infections, as well as activation of protein kinase activity, and stress-activated protein kinase signaling cascade in S. lycopersicum. Gene Solyc11g013370.2.1 has roles in defense response to fungus, leaf morphogenesis, seed dormancy, and positive regulation of cell division in tomatoes. Genes OIS99998 and OIS96773 have roles in methylation in wild tobacco (Nicotiana attenuata), whereas PGSC0003DMT400072321 has functions related to double-strand break repair via homologous recombination; DNA replication, recombination, and repair, and PGSC0003DMT400003724 has functions in tRNA N2-guanine methylation, tRNA processing, and methylation in S. tuberosum.

Discussion
The identi cation and analysis of notable QTL regions associated with different traits has been one of the cornerstones of modern molecular breeding for plant genetic improvement. In Capsicum spp., previous studies focused on identifying QTL linked to diverse traits including but not limited to resistance to major diseases such as chile pepper blight caused by the oomycete P. capsici, yield and yield components, capsaicin (heat) content, and agronomic traits, among others (Table 1). Given this wealth of information from previous QTL studies, it would be necessary to re-evaluate results from linkage mapping using a meta-analysis approach to re ne genomic regions associated with important traits resulting in a more e cient implementation of MAS in chile pepper breeding programs. QTL meta-analysis for different traits in Capsicum remains lacking, where a major focus in the past has been the identi cation of meta-QTL for resistance to P. capsici 24 . This status quo of meta-studies for chile peppers has thus driven us to explore meta-QTL for the Capsicum QTLome for diverse traits, i.e. not only for those QTL related with disease resistance, but also for those loci linked with other important yield and agronomic characters in chiles. Here, we report the rst known meta-analysis of the chile pepper QTLome rendering deeper insights into the genetic architecture of diverse sets of traits for this valuable crop.
We employed a relatively stringent method in declaring a QTL cluster as a meta-QTL: (1) each meta-QTL should be composed of at least two different QTL; and (2) these QTL should come from at least two independent studies. Accordingly, from an initial set of 39 meta-QTL, only 30 were regarded to be 'true' meta-QTL across the 12 chromosomes of chile peppers, with con dence intervals between 0.55 cM (MQTL5.2) and 14.90 cM (MQTL12.3). These criteria were therefore relevant for a more accurate representation of the meta-QTL identi ed for the chile pepper QTLome. In other crop species, varying numbers of meta-QTL have been identi ed. Only 11 meta-QTL were detected for seedling stage salinity tolerance in rice 6 , whereas 60 meta-QTL were identi ed for Fusarium head blight resistance in wheat 25 . In another meta-study of QTL in pea plants, 27 meta-QTL were resolved for seed protein content and yield-related traits 26 . Such differences could be a consequence of the genome size, reliability of the consensus map used for meta-analysis, number of QTL regions identi ed, as well as the intrinsic properties of the reported QTL, such as phenotypic variation explained and LOD scores. As precision in QTL positions are dependent on population size and trait variation explained 7 , re-calculating positions based on the type of mapping populations used for analyses could facilitate a better representation of the genetic positions for each of the QTL evaluated.
One of the objectives of a meta-QTL study is to delimit the region of a QTL using information from multiple linkagemapping studies. Chromosome P5 represents a major chromosome for P. capsici resistance in chile peppers, with large-effect QTL reported in previous studies 27− 29 . In the current study, we reported two meta-QTL regions in chromosome P5, namely MQTL5.1, and MQTL5.2 delimited to < 1.0 cM con dence interval, i.e. 0.79 cM and 0.55 cM, respectively, comprised of QTL mapped for P. capsici fruit and root rot resistance. These corresponded to the genomic regions having the most re ned genetic distance among all the meta-QTL identi ed in the present work.
Similarly, in peanut (Arachis hypogea), a recent meta-analysis of QTL for late leaf spot resistance delimited a region to 0.38 cM and 0.70 cM 30 , whereas in wheat, genomic regions associated with Fusarium head blight resistance and root-related traits were narrowed to 0.82 cM 25 and 0.50 cM intervals 31 , respectively. Capsicum spp. MQTL5.1 and MQTL5.2 consisted largely of major effect QTL, with percent variation explained ranging between 10 and 52.7% (MQTL5.1) and 8.9 and 67.7% (MQTL5.2) identi ed from ve independent QTL mapping studies, with N ranging between 63 and 297 individuals. Notably, these constituent QTL also represent those with the highest phenotypic trait variation explained in the Capsicum QTLome evaluated; this could be a reason for a more re ned meta-QTL region for disease resistance. Furthermore, MQTL5.2 consisted of 24 QTL, which was next to MQTL1.3 identi ed to having the highest number of individual QTL. Among the criteria for choosing a meta-QTL for selection are (1) a small con dence interval, (2) a high number individual QTL comprising the meta-QTL, and (3) a high trait variation explained of initial QTL 7 . Considering these factors, MQTL5.1 and MQTL5.2 could serve as potential targets for marker-assisted breeding and selection for improved P. capsici resistance in chile peppers. The identi cation of meta-QTL linked DNA-based markers will help prioritize different QTL for introgression through MAS in plant breeding programs 3,31 . In this regard, information from the anking sequences for MQTL5.1 and MQTL5.2 identi ed in chromosome P5 will be utilized for the development of Kompetitive allele speci c (KASP®) 32 SNP assays for marker-assisted breeding. These KASP assays will be further validated using a recombinant inbred line population previously developed at New Mexico State University 33 , and on a diverse population of New Mexican chile peppers to screen for resistance to different races of P. capsici.
The power of meta-QTL analysis lies in determining genomic regions that are most frequently involved in phenotypic variation and in delimiting the QTL intervals, therefore enabling candidate gene identi cation for positional cloning 31 . Also, meta-QTL are potentially genomic regions that are highly rich in genes 25 thereby facilitating pyramiding or stacking of important loci. Putative blight resistant protein homologues and leucine rich repeat (LRR) receptor-like serine/threonine protein kinases have been previously identi ed as candidate genes for P. capsici resistance in chile peppers 34,35 . In the current study, candidate gene analysis using sequences for markers anking MQTL5.1 and MQTL 5.2 in chromosome P5 identi ed genes with diverse biological functions related to disease resistance, including DNA repair, DNA strand renaturation, ion transport, and several epigenetic mechanisms such as DNA, RNA, and histone methylation/demethylation, indicating the possible function of epigenetics in controlling gene expression for disease resistance in chile peppers. Epigenetics and its relationship with conferring disease resistance has been well recognized in other crops such as Arabidopsis 36 , rice 37 , and maize 38 . The denser cytosine methylation pro le of the Capsicum genome relative to that of the tomato and potato genomes 39 could indicate the relevance of epigenetics for the expression of different genes in peppers. Accordingly, identifying epialleles near the meta-QTL regions in chromosome P5 could be important in breeding towards improving resistance to P. capsici in chile peppers. Nevertheless, while the candidate genes identi ed here represent promising targets for future breeding, it is not known whether they are the true functional regulators of the detected meta-QTL, as many other genes could be present within the meta-QTL regions 5 . It would therefore be relevant to perform functional validation of the effects of these candidate genes using different chile pepper germplasm. Overall, meta-QTL analysis con rmed the relevance of chromosome P5 as a major genomic region harboring QTL and different candidate genes for P. capsici resistance in Capsicum.
Chile peppers are unique among the members of family Solanaceae due to their ability to produce capsaicinoids which render distinct avors and heat pro les. Previously, gene mapping, allele sequence data, and expression pro le analyses collectively identi ed the pungency gene Pun1 in chromosome P2 responsible for the biosynthesis of capsaicinoids in chile peppers 40  In the current study, meta-QTL analysis was used to dissect the genetic architecture of diverse traits in Capsicum. Genomic regions for disease resistance to P. capsici in chile peppers were re ned, and the role of chromosome P5 as a major genomic region harboring disease resistance QTL has been con rmed. Two meta-QTL, MQTL5.1 and MQTL5.2, in chromosome P5 have been delimited to < 1.0 cM intervals. Analysis of candidate genes for these meta-QTL revealed biological functions related to DNA repair, response to bacterial and fungal infection, and DNA, RNA, and histone methylation, which demonstrate the potential role of epigenetics on resistance to P. capsici. The colocalization of several unrelated QTL on similar chromosomal regions demonstrates potential pleiotropic effects and the effect of linkage due to location. SNP assays will be developed for these meta-QTL and will be used for MAS for resistance to pepper blight. This study by far is the largest reported meta-analysis of different traits and the rst known study of the Capsicum QTLome. The information presented here could serve as a valuable resource for the genomic improvement of diverse sets of traits in chile peppers.

Materials And Methods
Collection and characterization of QTL for different traits in chile peppers. A comprehensive bibliographic review of 29 QTL mapping studies published between 2010 and 2020 (Table 1) was rst conducted to generate a QTL database for Capsicum. The evaluated traits in chile peppers were divided into ve major categories, namely (1) adaptation; (2) agronomic, quality, and yield; (3) disease resistance; (4) heat pro le (pungency); and (5) biochemical and physiological traits. Each QTL was characterized according to the number of lines used for QTL mapping, type of mapping population (e.g. F 2 , RIL, BC, DH), QTL name, trait, chromosome and linkage group designations, LOD, phenotypic variation explained (R 2 ), and chromosome positions (in cM). QTL with LOD scores of < 2.0 and with R 2 values not reported in the original study were excluded in further analysis. Genomewide association studies were not included in the literature review.
Projection to a consensus map and analysis of meta-QTL. Three different genetic maps from four cultivated species of Capsicum were used to develop a consensus map for the analysis of meta-QTL in chile peppers. These included an interspeci c SNP genetic map derived from the hybridization between C. annuum and C. frutescens 45 ; a SLAFbased SNP array resulting from genotyping an F 2 population of a cross between C. chinense and C. annuum 46 ; and an intraspeci c SNP linkage map derived from the hybridization between two C. baccatum varieties 47 . Consensus maps for each of the 12 chromosomes of chile pepper were created using the 'LPmerge' package 48 in R 49 . This function implements a linear programming (LP) algorithm to effectively reduce the mean absolute error in combining different genetic or linkage maps.
For QTL projection to the consensus map, the con dence interval (CI) for each QTL were calculated according to Darvasi  Meta-analysis was performed using Biomercator v.4.2.3 which implements a maximum likelihood algorithm developed by Go net and Gerber 4 . In this method, an N number of QTL linked to the same trait or set of related traits detected in independent experiments and located in the same genomic regions is determined to be consistent with ve different QTL models, namely, 1, 2, 3, 4, and 5-N QTL models. An Akaike information criterion was used to determine the best model in identifying a meta-QTL, or "real" QTL which best represent the original QTL 25 . Additionally, only those genomic regions where QTL from at least two different genetic mapping studies co-localized to form a meta-QTL were regarded as a meta-QTL 5 . Therefore, no meta-QTL identi ed in this study consist of only a single QTL (i.e. a singleton). Designations for each identi ed meta-QTL were based on the corresponding chromosome number and their position relative to the short arm of the chromosome (e.g. "MQTL1.1", "MQTL2.3"). Pepper chromosomes with the identi ed meta-QTL and their anking markers were redrawn using the 'LinkageMapView' package 53 in R.
Candidate gene identi cation for the meta-QTL. Identi cation of candidate genes was conducted using the sequences of the anking markers for the identi ed meta-QTL that has a < 1.0 cM con dence interval. Flanking sequences were BLASTn searched in EnsemblPlants (https://plants.ensembl.org/index.html) 54 against the genome of C. annuum and annotated genes and their biological functions were listed. Additionally, orthologous genes from Solanaceous plant species, including tomato (S. lycopersicum), potato (S. tuberosum), and wild tobacco (N. attenuata) were identi ed.