A novel approach to develop wheat chromosome-specific KASP markers for detecting Amblyopyrum muticum segments in doubled haploid introgression lines

Many wild relative species are being used in pre-breeding programmes to increase the genetic diversity of wheat. Genotyping tools such as single nucleotide polymorphism (SNP)-based arrays and molecular markers have been widely used to characterise wheat-wild relative introgression lines. However, due to the polyploid nature of the recipient wheat genome, it is difficult to develop SNP-based KASP markers that are codominant to track the introgressions from the wild species. Previous attempts to develop KASP markers have involved both exome- and PCR-amplicon-based sequencing of the wild species. But chromosome-specific KASPs assays have been hindered by homoeologous SNPs within the wheat genome. This study involved whole genome sequencing of the diploid wheat wild relative Amblyopyrum muticum and development of a SNP discovery pipeline that generated ∼38,000 SNPs in single-copy wheat genome sequences. New assays were designed to increase the density of Am. muticum polymorphic KASP markers. With a goal of one marker per 60 Mbp, 335 new KASP assays were validated as functional. Together with assays validated in previous studies, 498 well distributed chromosome-specific markers were used to recharacterize previously genotyped wheat-Am. muticum doubled haploid (DH) introgression lines. The chromosome specific nature of the KASP markers allowed clarification of which wheat chromosomes were involved with recombination events or substituted with Am. muticum chromosomes and the higher density of markers allowed detection of new small introgressions in these DH lines. Key Message A novel methodology to generate chromosome-specific SNPs between wheat and its wild relative Amblyopyrum muticum and their use in the development of KASP markers to genotype wheat-Am. muticum introgression lines.


Declarations 23
Funding 24 Institute while this work was carried out. 30 The remaining authors declare that the research was conducted in the absence of any 31 commercial or financial relationships that could be construed as a potential conflict of 32 interest. 33

Availability of data and material 34
Raw reads data for Am. muticum has been made available through the Grassroots data 35 repository hosted by the Earlham Institute and funded by DFW programme 36 (https://opendata.earlham.ac.uk/wheat/under_license/toronto/Grewal_et_al_2021-09-37 13_Amybylopyrum_muticum/). All DH lines used in this study are available through the 38

INTRODUCTION 85
Bread wheat (Triticum aestivum L., 2n = 6x = 42, AABBDD) is one of the most widely 86 grown and consumed crops worldwide. After two spontaneous interspecific hybridisation 87 events (Dvořák et al. 1993 Genotyping of introgression lines is more efficient when using wild-relative genome-112 specific SNPs. The Axiom ® Wheat-Relative Genotyping SNP Array was developed ( The genotyping procedure was as described by Grewal  performed using a QuantStudio 5 (Applied Biosystems) and the data analysed using the 216 QuantStudio TM Design and Analysis Software V1.5.0 (Applied Biosystems). 217

Multi-colour Genomic in situ Hybridisation (mc-GISH) 218
Preparation of the root-tip metaphase chromosome spreads, the protocol for mcGISH and 219 the image capture was as described in Grewal et al. (2020b). Briefly, genomic DNA from 220 T. urartu (to detect the A-genome), Aegilops speltoides (to detect the B-genome), and 221 Aegilops tauschii (to detect the D-genome) and Am. muticum were isolated as described 222 above. The genomic DNA of (1) T. urartu was labelled by nick translation with 223 ChromaTide TM Alexa Fluor TM 488-5-dUTP (Invitrogen; C11397; coloured green), (2) Ae.  for counterstaining all slides. Metaphases were detected using a high-throughput, fully 232 automated Zeiss Axio ImagerZ2 upright epifluorescence microscope (Carl Zeiss Ltd., 233 Oberkochen, Germany). Image capture was performed using a MetaSystems Coolcube 1m 234 CCD camera and image analysis was carried out using Metafer4 (automated metaphase  235 image capture) and ISIS (image processing) software (Metasystems GmbH, Altlussheim, 236 Germany). 237  robust and their positions on the wheat chromosomes are indicated in Fig. 2d. 286

Results
In total, 498 well-distributed, chromosome-specific KASP markers (Online Resource 3), 287 polymorphic between wheat and Am. muticum, were used for downstream genotyping of 288 introgression lines. Fig. 2e shows a line plot of the physical distance between these 289 markers in wheat where each gridline of the y axis represents 10 Mb physical distance on 290 a chromosome. The distance between the markers ranged from just 3 bases to ~82.5 Mb 291 with an average distance of 26 Mb. The average distance between the tip of the short arm 292 and the first marker on the arm was 2.9 Mb while that from the last marker to the end of 293 the long arm was 2.3 Mb. There were only seven instances where the gap between two 294 KASP markers exceeded the desired 60 Mb and these are shown with a red stroke in the 295 line in Fig. 2e. All these gaps were due to poor availability of SNPs within the desired bin 296 as shown by the corresponding SNP density plot (Fig. 2b). 297 specific cases of aneuploidy in some DH lines to support the GISH observations but also 320 suggested disparities with previously reported results. 321 Table 1 Details of the type of introgression, its code (as indicated in Fig. 2g) still present in these lines (Fig. 3a). However, GISH indicated that the 2T was potentially 356

Validation of KASP markers through genotyping of introgression lines 298
introgressed as a whole chromosome due to the presence of Am. muticum telomeric repeat 357 signals on both ends of this introgression (Fig. 3d). If the 2AS segment had recombined 358 with 2T or translocated onto another wheat chromosome, it would not be visible via GISH 359 due to its small size. 360 The markers also showed that DH lines 16-21 had a small 6T segment (up to 10 Mb) on 361 the distal end of 6DL (6T.D2; Fig. 3a) which was not previously detected by the Axiom 362 array and is not visible by GISH. Genoytping analysis of four other DH lines 62, 71, 74 and 363 348, showed that in addition to the 4T.D2 segment, a very small segment (up to 20 Mb) 364 from 7T was present at the distal end of 7AS (7T.A1; Fig. 3b) which had not been detected 365 before in these lines. This very small segment on the distal end of chromosome 7AS was 366 also detected by GISH in this study (Fig. 3e). The KASP markers were also able to detect 367 another small segment (between 20-30 Mbp) from chromosome 5T in DH lines 121 and 368 122 (5T.D1; Fig. 3c). Due to its slightly bigger size, this Am. muticum segment can be 369 viewed by GISH on the distal end of chromosome 5DL in DH-122 as shown in Fig. 3f

Discussion 390
Previous studies have reported chromosome-specific KASP markers between wheat and 391 Am. muticum (Grewal et al. 2020a) and other wild relative species (Grewal et al. 2020b;  392 Grewal et al. 2021), which have been used for genotyping wheat-wild relative introgression 393 lines. The objective of this work was to fill in the gaps with more KASP markers to increase 394 the efficiency of genotyping by using an approach that involved faster SNP discovery and 395 a more robust, chromosome-specific assay design than the ones reported in previous 396 studies. In this work, we produced ~38K SNPs between wheat and its wild relative Am. 397 muticum in single-copy regions of the wheat genome and then converted some of these 398 into wheat chromosome-specific KASP markers. In combination with previously designed 399 chromosome KASP markers, a new set of well-distributed markers was obtained and used 400 to re-genotype wheat-Am. muticum DH introgression lines ) to validate 401 the functionality of these markers as efficient genotyping tools and detect as many Am. 402 muticum introgressions as possible. 403 A recently developed set of KASP markers (Set 2) was tested on Am. muticum accessions 404 in this study but only 5.8% of the 224 assays were found to be polymorphic with wheat. 405 This was as expected since this set of markers was originally developed to detect T. urartu 406 introgressions in a wheat background (Grewal et al. 2021). When the 13 KASP markers 407 were added to the 150 Am. muticum KASP markers developed during the original study 408 (Grewal et al. 2020a), numerous gaps between markers were still present (Fig. 2c)  409 preventing a uniform spread of markers able to detect Am. muticum introgressions across 410 the whole of the wheat genome. 411

SNP Discovery 412
A major bottleneck at this stage was the lack of SNPs between wheat and Am. muticum 413 that could be converted to KASP markers in regions that lacked an existing assay. With 414 the advent of cheaper sequencing costs, it was possible to sequence the wild relative 415 species to gain abundant SNPs for KASP assay design, some of which would be polymorphic 416 between the species. However, in polyploid crops like bread wheat, it is challenging to 417 generate chromosome-specific KASP assays able to distinguish heterozygous from 418 homozygous individuals (co-dominant SNPs) and requires extensive validation (Allen et  419 al here was to find SNPs in single-copy regions of the wheat genome using bioinformatic 423 tools, thereby, resulting in ~38K SNPs, each specific to a wheat chromosome (Fig. 1). 424 When the wheat genome assembly RefSeq1.0 was published ( Fig. 3a-c). The chromosome-specificity of the KASP markers 511 allowed detection of the wheat chromosome that was involved in the recombinant 512 chromosome or that had been substituted. Thus, it was observed that in DH lines 15 and 513 16, 2T.A1 was a whole chromosome that had replaced both 2A chromosomes rather than 514 recombined with B genome chromosomes as previously reported. Where possible due to 515 the size of the introgression, some of these results were validated by mcGISH in this work 516 (Fig. 3d-f). 517 The chromosome-specificity of these KASP markers also allowed the detection of a number 518 of wheat chromosome deletions in the DH lines as shown in Table 1. However, these were 519 limited to the detection of homozygous deletions. These homozygous deletions included 520 both whole wheat chromosomes or segments (both large and small) from the wheat 521 chromosomes. In this context, one of the main observations involved the 15 sister DH 522 lines (codes between DH-124 to 147 and DH-355 to 357) that showed that the pair of 1B 523 chromosomes had been deleted in these lines expect for a small segment at the distal end 524 of 1BL. We proposed that it was this 1BL segment that had translocated/recombined with 525 a pair of A genome chromosomes, most likely chromosomes 1A. These lines also have 16 526 A genome chromosomes (King et al. 2019) and so it is possible that the pair of 1A-1BL 527 recombinant chromosomes are present in addition to the pair of 1A chromosomes since 528 the KASP markers at the distal end of 1A do not indicate the absence of any of the 1A 529 wheat alleles. 530

Conclusion 531
Unlike previous work that relied on PCR-based amplicon sequencing (Grewal et al. 2020a), 532 this method of generating SNPs between wheat and Am. muticum in single-copy regions 533 of the wheat genome, made possible due to whole genome sequencing of the wild species, 534 is rapid and allows for the development of chromosome-specific KASP assays. A variety of 535 wild relative species are being used to increase the genetic diversity in hexaploid wheat. 536 This approach can therefore be applied to other wheat wild relative species for SNP 537 discovery, highlighting the need for greater investment in whole genome sequencing of 538 these wild species. These KASP markers have greatly increased our capability to 539 characterise, screen and identify both introgressions and wheat chromosomal aberrations 540 in wheat-wild relative introgression lines. However, it is important to note that their 541 efficiency is dependent on their density across the wheat genome and small introgressions 542 existing between two KASP markers could have gone undetected. With the reducing cost 543 of DNA sequencing, we envisage that the next improvement in characterisation of such 544 introgressions, with the potential to give higher resolution, would be low-coverage whole 545 genome resequencing of wheat-wild relative introgression lines. 546 547