Besides its importance as a crop, rye is one of the parents of a man-made species Triticale and the short arm of rye chromosome 1 (1RS) has been introgressed into several hundreds of wheat cultivars [1,2]. In fact, some of the most successful wheat varieties carry the 1BL.1RS translocation as the presence of 1RS in the wheat genome increases both yield and protein content in grains [3]. Moreover, 1RS carries a cluster of genes encoding resistance to stem, leaf and yellow rust – Sr31, Lr26 and Yr9, respectively [4] and a self-incompatibility locus [5]. On the down side, 1RS carries the Sec-1 locus coding for ε-secalin, which negatively affects bread making quality [6]. Thus, it would be of great advantage to isolate those genes individually through map-based cloning and develop markers for marker assisted selection in rye and wheat.

Despite the economic importance of rye, little is known about its genetic make up at the DNA sequence level. To our knowledge, there is no ongoing sequencing project in rye, and there are no plans to target gene-rich fractions of its genome. Rye is underrepresented in the sequence databases compared to wheat and barley for which 1,104,002 and 529,839 sequences respectively, are deposited in GenBank. There are only 9,807 rye sequences (about 5 Mbp) available, of which about 90% are expressed sequence tags (ESTs). Updated list of rye genes, markers and linkage data was created by Schlegel and Korzun [7]. The lack of sequence information is a major limitation for marker development and gene cloning in this species.

The monoploid genome size of rye (1Cx = 7,917 Mbp) is the largest among temperate cereals, almost 40% larger than that of bread wheat (Table ​(Table1).1). This is due to the presence of a large amount of highly repetitive sequences. Flavell et al. [8] estimated the repetitive DNA content of rye to be 92%. Despite the progress in sequencing technology and bioinformatics, sequencing the whole rye genome remains a very difficult and expensive task. In particular, genome shotgun sequencing of such a large and repetitive genome seems currently impossible. On the other hand, the short arm of rye chromosome 1 represents only 5.6% of the rye genome and with the molecular size of 441 Mbp, 1RS is comparable in size to the whole rice genome, which was recently sequenced [9-11]. Recently, a method has been developed to dissect large plant genomes into individual chromosomes using flow cytometric sorting (reviewed in [12,13]). A protocol for sorting individual rye chromosomes has been set up by Kubaláková et al. [14], and Šimková et al. [15] created two BAC libraries from flow sorted 1RS arms. The library represents a valuable tool for map-based cloning, targeted sequencing and marker development.

End sequencing of BAC clones enables generating random sequence information distributed across the whole genome. Kelley et al. [16] developed a protocol for high throughput BAC end sequence (BES) generation using automated sequencers. This protocol is now a routine in large sequencing centers, reducing cost and enabling the creation of large data sets. Nevertheless, the number of BESs from the Triticeae tribe is currently limited with only 37,609 hexaploid wheat BESs and 32 Triticum monococcum BESs in GenBank, representing the whole tribe. Beyond the sequence information itself, BESs are a valuable source of molecular markers. Shultz et al. [17] used BESs derived from BACs, representing minimum tiling path of soybean, to develop new microsatellite markers. Among the first 135 primer pairs tested, more than 60% were polymorphic. Paux et al. [18] took advantage of a BAC library specific for wheat chromosome 3B [19], and sequenced BAC ends to isolate chromosome-specific molecular markers based on inserted transposable elements (ISBP – Insertion Site Based Polymorphism). Paux et al. [18] succeeded in developing thirty-nine 3B-specific markers to anchor BAC contigs to the genetic/deletion map and have since then developed several hundreds of ISBP markers from 3B (unpublished). According to the authors' estimate, about 5% of BESs are suitable for ISBP development.

The observed content of repetitive sequences (84.2%) is lower than expected and is similar to that found in the wheat B genome (85.9%) by Paux et al. [18]. As indicated above, rye has a significantly higher 1Cx value than bread wheat. Thus, assuming an equal number of genes, the rye genome should contain more repeats than the wheat genome. In fact, using Cot analysis, Flavell et al. [8] estimated the content of repetitive elements in rye genome to be of about 92%. Our low estimate could be due to insufficient representation of rye repeats in databases that were used in our analyses. For example, the TREPtotal database, which showed most of the significant matches, comprises only 39 entries of rye repeats. This limited information, compared to wheat and barley with 663 and 554 elements, respectively (in the same database), could result is short sequence alignments and hence underestimation of the amount of repeated DNA in rye BES. As expected, Class I elements (retrotransposons) were the most abundant repetitive fraction in the rye genome similar to what was found for the wheat and maize genomes [18,37].

A comparison of the frequency of various types of repeats in genomes of rye and wheat B and D genomes suggested a greater similarity between the rye and wheat B genome than between the rye and wheat D genome. This close relationship of rye and wheat B genome was also supported at the sequence level. It is interesting to note that rye and the putative B genome progenitor Aegilops speltoides have the same mating system, both being outcrossers. Moreover, the B genome is largest of the three homeoelogous wheat genomes and similar in size to the rye genome. On the other hand, the similarity in repeat composition between both genomes may simply reflect similar trends in the mode of their expansion via the LTR-retrotransposon activation.

Until now, the lack of sequence data did not permit estimation of the number of genes in rye. By analyzing 2 Mbp sequence from chromosome arm 1RS, we estimate 2,000 genes being present on 1RS and thus 36,000 genes in the rye genome. This first estimate for rye is consistent with the predicted gene numbers in other plants. Most recent estimate for gene number in A. thaliana genome is 33,000 (TAIR8 release) [38], the current TIGR rice genome annotation (Release 5) [39] estimates 41,000 genes in rice genome. Both numbers are close to our prediction for rye.

In addition to the analysis of the rye genome composition, we used BAC end sequences for marker development. There is still a low density of markers available for rye genome and additional markers are urgently needed. Development of a genetic linkage map of rye with 183 markers was reported by Korzun et al. [40]. Bednarek et al. [41] presented a genetic map of rye containing 480 markers including 200 RFLPs, 179 AFLPs, 88 RAPDs and 13 proteins. Khlestkina et al. [42] mapped 99 SSRs derived from EST sequences (SSR-ESTs), nine of which mapped to chromosome 1R. Several attempts were made to transfer SSR and/or EST-SSR markers from wheat and barley into rye [43-45]. Recently Varshney et al. [46] succeeded in transferring and mapping 12 barley SNP markers in rye. Thus, to date the density of markers is quite low and does not allow efficient map based cloning or MAS in rye.

This work presents a method for targeted development of molecular markers from specific parts of the rye genome using the 1RS chromosome arm as a case study. Until now, there are only a few markers available for this critical part of the rye genome. This hampers marker assisted breeding not only in rye, but also in wheat where the markers from 1RS would permit selection of lines with introgressions of desired parts of 1RS without the harmful genes. We have developed twelve new ISBPs markers for 1RS and designed almost 200 additional primer pairs for potential ISBP markers that remain to be tested. If needed, generation of additional BAC end sequences from 1RS is possible. In addition to the development of new ISBP markers, BES containing microsatellites were used to develop new 1RS-specific SSR markers. Out of the 63 tested microsatellites that were tested, 21 were specific for 1RS [47]. Thus, in total 33 1RS-specific markers were obtained using our strategy. On the other hand, Badnarek et al. [41] isolated 198 AFLP and RAPD markers from genomic DNA and only 29 markers were specific for 1RS. This clearly demonstrates the efficiency of the targeted approach.

One of the potential problems in targeted marker development from the 1RS-specific BAC library is the contamination with clones from wheat DNA. This could explain the specificity of amplification for several ISBP markers with wheat DNA. As the 1RS arm was flow-sorted from a wheat-rye ditelosomic addition line, contamination by fragments of wheat chromosomes cannot be avoided. However, the contamination of the 1RS BAC library was estimated to be only 14% [15] and thus should not seriously compromise the efficiency of marker development. Further improvement could be achieved by selecting BAC clones from contigs after fingerprinting the library, as the contigs originate from the chromosome of interest while the infrequent contaminating BACs remain as singletons.