Telomere-to-telomere gap-free genome assembly and chromosome segment substitution lines facilitated gene identification in wild rice

doi:10.21203/rs.3.rs-2587158/v1

Download PDF

Article

Telomere-to-telomere gap-free genome assembly and chromosome segment substitution lines facilitated gene identification in wild rice

https://doi.org/10.21203/rs.3.rs-2587158/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 29 May, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

The abounding variations hidden in wild rice (Oryza rufipogon) provided potential reservoirs of useful genes for rice breeding. However, the genome of wild rice has not yet been comprehensively assessed. In this study, we constructed the telomere-to-telomere (T2T) genome assembly and annotation of wild rice Y476 with a total length of 421.1 Mb and 38,055 predicted protein-coding genes, which features evident improvements in the continuity and quality of sequence compared with the existing wild rice genomes. Of Y476 genome assembly, 128 Mb and 5,984 novel genes were not found in ‘Nipponbare’ reference, leading to 690 gene families expanded and 247 gene families contracted, including disease resistance and grain regulation gene family. Two sets of chromosome segments substitution lines (CSSLs) were constructed, using Y476 as the donor parent and cultivated rice as the recurrent parents, covering the whole genome of Y476, providing materials for introducing novel genes and identification of their function. Combining the T2T reference genome and CSSL population, 254 QTLs associated with agronomic traits, biotic and abiotic stresses were identified. A receptor-like kinase gene associated with rice blast resistance was identified, a 7.8-kb insertion was found in the intron of this gene which increased its expression level. Using CRISPR/Cas9 technology and transcriptomic data, we confirmed that this wild rice allele improves rice blast resistance via regulating OsMADS26 expression. Collectively, our findings provided not only a T2T gap-free reference genome but also a highly efficient platform for novel gene identification of O. rufipogon, and therefore, have important implications for wild rice utilization.

Biological sciences/Plant sciences/Plant molecular biology

Biological sciences/Genetics/Agricultural genetics

Oryza rufipogon

T2T genome assembly

chromosome segment substitution line

gene identification.

The domestication of cultivated rice (Oryza sativa L.) from its ancestral progenitor wild diploid rice, Oryza rufipogon, is considered one of the most important developments in human history, and rice is now the staple food feeding more than half of the world’s population ^1,2. O. rufipogon has adapted to different ecological environments, spanning a broad geographical range of global pantropical regions, extensively in diverse natural habitats in South China^3,4. During the course of domestication and breeding of cultivated rice, many desirable traits such as abiotic tolerance and biotic resistance were lost, profoundly decreasing genetic diversity⁵. The genomes of wild rice constitute an important reservoir of agronomic, biotic resistance and abiotic tolerance traits for rice genetic improvement, and also provide fundamental data for studying plant genome evolution within a short timeframe. Strategies to harness such traits for rice improvement showed clear promise, as exemplified by the introgression of “wild-abortive sterile” gene from O. rufipogon that lead to a great success of hybrid rice⁶.

The importance of O. rufipogon for rice improvement has been well documented, and substantial progress has been achieved in O. rufipogon utilization based on the advances in genomics and genetics^5,7,8. Harnessing the genetic diversity in O. rufipogon for rice improvement requires: 1) a high-quality reference genome of wild rice, and 2) a suitable and permanent genetic population in clearly genetic backgrounds. Assembly of a high-quality reference genome of O. rufipogon is not easy because of its relatively high heterozygosity. To date, many O. rufipogon accessions have been sequenced and two accessions had chromosome-scale assemblies^3,9. The large or super pan-genomic studies of Oryza family were performed in recent years ^8,10, many O. rufipogon accessions were de novo sequenced in these studies. However, a telomere-to-telomere gap-free genome is needed to better understand the genomic/genetic diversity of O. rufipogon. More representative genomes of high quality from different geographically distributed wild rice populations will facilitate the study and utilization of the valuable resources of the wild rice.

Chromosome segment substitution lines (CSSLs) is a powerful tool for identifying naturally occurring, favorable alleles in crop wild relative germplasm¹¹. In the last two decades, the utility of CSSLs in identification of novel genomic regions and quantitative trait loci (QTL) hot spots influencing a wide range of traits has been well demonstrated in wild rice¹². However, their incompleteness and fragmentation limit their use for wild rice functional studies. Further efforts are needed to fill the gaps between genomic studies and gene identification, so that practical breeding can be performed in a rapid, high-throughput and precise manner, taking full advantage of whole-genome information of O. rufipogon.

Herein we reported the first telomere-to-telomere (T2T) genome assembly and annotation for one typical Chinese O. rufipogon accession by integrating HiC, BioNano, Nanopore and HiFi techniques. Compared with previously assembled genomes, this genome assembly showed considerable improvements in contiguity, completeness and correctness of O. rufipogon assembly. With the near-complete genome assemblies, comparative genomics analyses demonstrated a large number of structure variations and gene function changes in wild rice. Some novel genes and specific gene clusters suggested the cause of diseases resistance in wild rice. We constructed two CSSL population to introduce favorable chromosome segments from O. rufipogon to cultivated rice, allowing us to identify a large amount of QTLs associated with agronomic, biotic and abiotic stresses. These resources will accelerate functional genomic studies of wild rice and inform future breeding programs for rice improvement.

Morphology and genetic structure analysis of an O. rufipogon accession

Firstly, we investigated more than 1000 accessions from their original habitats all over China and have high geographical and agricultural phenotypic diversity. The biotic stress resistance and abiotic stress tolerance of this panel were evaluated. One accession Y476, which was collected from Sanya city, Hainan province had excellent overall resistance to various biotic and abiotic stresses. It has the typical characteristics of wild rice, including creeping and vigorous growth capacity, long awns, purple and completely exserted stigma, black-hulls, and reddish-brown pericarps (Fig. 1A). It was immune to the rice blast pathogen, and survived after treated with 200 mM NaCl for 5 days (Fig. 1B, 1C).

To characterize the genome composition of Y476, we generated 26.2 Gb paired-end (PE) Illumina short reads (Supplementary Table 1) and performed population structure analysis combined with whole-genome re-sequencing data from 444 O. rufipogon accessions, 268 O. sativa aus, 1882 O. sativa indica, and 1194 O. sativa japonica accessions^2,13. We performed a phylogenetic and principal component analysis (PCA) of the rice pool. Japonica, indica and Aus rice formed completely isolated clusters, whereas the genomes of O. rufipogon were classified into three types, Or-I, Or-II and Or-III. The phylogenetic analysis and PCA showed that Y476 was closed to Or-II type clusters (Fig. 2A, Supplementary Fig. 1A), which mainly distribute in South and South-East Asia. We further analyzed the admixture pattern of Y476 to confirm its genomic composition using the program ADMIXTURE. With a K-value of 6, six genome types (Or-I to Or-III, aus, japonica-type, and indica-type) were identified. Y476 comprises 69.5% of the wild rice components, and 8.4% of indica, 9% of japonica, and 4% of aus component. (Supplementary Fig. 1B).

Sequencing, assembly and annotation of a gap-free wild rice genome

Different sequencing platforms were applied to develop a high-quality genome assembly for Y476. Approximately, 26.2Gb 150 bp Illumina PE reads (~ 62.2X) was used for K-mer analysis. The genome size of Y476 was estimated to be ~ 429Mb with a heterozygosity of 1.06% (Supplementary Fig. 2). The quality of long reads sequencing determines the quality of genome assembly, and the sequencing data generated in this study is particularly remarkable. 29.7Gb (~ 70.5X) HiFi reads were generated by the PacBio sequel II platform (Supplementary Table 1), with 99.9% accuracy of long reads and the N50 length of greater than 15 kb (Supplementary Fig. 3A). For the PacBio HiFi read assemblies, the preliminary assembly applied Hifiasm on HiFi reads with -hic mode, and generated contigs with the N50 length of 33.2Mb, which is 3 ~ 30 times larger than that of the reported wild rice genomes^3,9 (Table 1). Nanopore ONT sequencing applies the latest technology Ultra-long sequencing, and the N50 is about 100 kb (Supplementary Fig. 3B), which is especially helpful for the assembly of highly repetitive regions, such as centromeric and telomeric regions, and finally generated telomere-to-telomere genome. About 40X Nanopore Ultra-long reads (~ 17Gb) were generated to assemble Y476. The NextDenovo was used to assemble the ONT data, forming 50 contigs with the N50 length of 9.4Mb. For highly heterozygous species with more heterozygous fragments in the genome sequence, the assembled contig sequences were further filtered using purge_dups. To correctly anchor the contigs of the Y476 primary assembly, 68.9Gb chromosome conformation capture sequencing (Hi-C) data of Y476 were generated. HiC-Pro was applied to generate chromosomal interaction maps and confirmed the correct ordering and orientation of all the pseudomolecules. With interactive signals, de novo assembled contig was anchored to each of 12 pseudomolecules. The chromosome ID and orientation were tuned in accordance with R498¹⁴, the gapless Y476 genome was generated, which contained 12 chromosomes and 3 gaps in chromosome (Chr.) 5, 8, and Chr. 12. The ONT-assembled genome was used to patch the HiFi genome with the three gaps by RagTag, which successfully closed 3 gaps (Supplementary Fig. 4). The final assembly was comprised of 12 gap-free chromosomes and 3 gaps achieved extremely high mapping rate (> 99.90%) and coverage of Illumina sequencing reads, further supporting the high quality of gap-closing. The de novo Bionano optical maps were aligned to the genome to verify the correct sequence and direction of the genome assemblies (Supplementary Fig. 5). Finally, the gap-free Y476 genome with a total length of 421.1Mb was assembled. HiCPlotter was applied to generate chromosomal interaction heatmaps of 12 chromosomes (Fig. 2B).

Table 1

Comparison of the Y476 genome with previously published assemblies of the wild rice genomes. * indicated data lack.
Accession number	Y476	IRGC106162⁹	Yuanjiang wild rice of China³
Assembly
Genome size (Mb)	421.1	377.1	376.5
Number of Chromosome(gaps)	12(0)	*	12 (1219)
Contig N50 (Mb)	33.2	13.2	1.1
Genome BUSCO (%)	98.80	96.20	97.36
QV	64.61	*	*
LAI	24.76	10.19	16.55
Telomeres number	18	0	0
Centromere number	12	0	0
Annotation
Number of predicted genes	38,055	33,903	34,830
Average gene length(bp)	2726	2397	2921
Average CDS length(bp)	1168	*	1,125
Protein BUSCO (%)	98.00	93.40	94.20
Repeat sequences (%)	56.80	49.37	44.14

For the genome annotation, our pipeline is shown in Supplementary Fig. 6. First, we screened the genome repetitive sequences to annotate transposable elements (TEs). The Extensive de novo TE Annotator (EDTA) was used to generate a high-quality repetitive sequence library and then the RepeatMasker was used to annotate TEs. A total of 687,167 TEs accounted for 59.21% of Y476 genome (Supplementary Table 2). After that, we used the sequences after repetitive sequence mask to predict the gene structure. Several de novo prediction software, homologous prediction, and transcript prediction with different principles produced an accurate gene model. The EVM integrated the results of different software outputs and filtered out some low-quality gene models (Supplementary Table 3). Finally, 38,055 protein-coding genes were predicted in our assembled genome, with an average coding sequence size of 1,168 bp and an average of 4.42 exons per gene (Supplementary Table 4). Homology-based annotation of the non-coding RNA sequences predicted 772 tRNA and 465 ribosomal RNA genes for the Y476 genome (Supplementary Table 5). Using MUMmer to further compare Y476 and MSU7.0 genomes (Nipponbare), we identified 2,726,973 SNPs and 404,723 InDels between the two genomes. Using 500kb intervals across the 12 chromosomes, we plotted the densities of GC content, genes, TEs, SNPs, and InDels in the genome (Fig. 2C).

Quality assessment and validation of Y476 assembly

The quality and completeness of the Y476 assembly were evaluated in multiple ways. First, Illumina short reads and long reads were mapped to the assembled genome, and mapping rates and coverage were used to evaluate the accuracy of the genome (Table 1). For the Illumina reads mapping rate was 99.15% and the coverage was 99.97%. Secondly, we used BUSCO to evaluate genomic completeness. About 98.8% of the core conserved plant genes (1,595 out of 1,614 BUSCOs) were found complete in the Y476 genome assembly (Supplementary Table 6), and 98.0% (1,581 out of 1,614) were found complete in the Y476 predicted genes (Supplementary Table 7). Of the protein-coding genes, 94.8% were predicted based on the information from NR (non-redundant, NCBI), Swissprot, InterPro, GO, and the KEGG database (Supplementary Table 8). Collectively, these results illustrated the high quality, reliability and accuracy of the Y476 assembly. We also assessed the k-mer-based quality estimations (k = 17) for the Y476 genome using Merqury with both Illumina PCR-free and HiFi reads^15,16. The Merqury-estimated quality value (QV) of the Y476 gap-free genome was 64.61 (Table 1). Finally, the LTR assembly index (LAI) value for the Y476 genome was 24.76 (Table 1), which is higher than all the published wild rice genomes^8,17. In addition, we identified the centromeric region on each chromosome (Supplementary Table 9). Using seven base telomeric repeats (CCCATTT at 5 'end and TTTAGGG at 3' end) as sequence query, we identified 18 telomeres in the Y476 genome (Supplementary Table 10). These results showed that our Y476 genome assembly has high reliability and quality. We also evaluated the two published wild rice genomes^3,9 using the same analysis, and the quality of Y476 was significantly better in comparison (Table 1).

Global comparison, identification of unique gene family and novel genes of Y476

To dissect the genome variation between Y476 and cultivated rice, we compared the Y476 genome and indica (Xian) variety ‘9311’ and japonica (Geng) variety ‘Nipponbare’ (Nip). The Y476 genome contains a higher proportion of repeat sequences, which resulted in larger genome size. Synteny comparison of the genomic structure revealed high collinearity between Y476 and Nip and 9311, and the fragment inversions, duplications, and translocations were showed in Fig. 3A. Three long fragment inversions, 4 Mb on Chr. 9 of 9311, 4.5 Mb on Chr. 6 and 3.3 Mb on Chr. 9 of Nip, were confirmed by PCR and sequencing (Supplementary Fig. 7). SNPs, small InDels, and the presence-absence variations (PAVs) were listed in Table 2. The genome similarity between Y476 and Nip is higher than between Y476 and 9311. PAV analysis identified 47,284 PAVs between Y476 and Nip, including 80 Mb deletion and 128 Mb insertion, and some randomly selected large PAVs were also verified by PCR amplification (Fig. 3B; Supplementary Fig. 8; Supplementary Table 11). Moreover, 18,656 genes were affected by these PAVs. 50,889 PAVs were identified between Y476 and 9311, (Fig. 3B; Supplementary Table 12), 11,205 genes were affected by these PAVs.

Table 2

Global comparison of Y476 and Nip and 9311
Statistics of variation	Nip vs Y476	9311 vs Y476
No. SNPs	2,726,973	3,262,275
No. small indels	404,723	522,076
Deletion number	21,340	24,030
Deletion total length (bp)	80,447,894	91,882,432
Deletion length range (bp)	50 − 1,076,270	50–838,432
Insertion number	25,944	26,859
Insertion total length (bp)	128,011,421	127,762,275
Insertion length range (bp)	50 − 2,624,538	50 − 1,024,568
Number of genes affected by PAV	18, 656	11,205
Duplications	527	467
Inversions	226	218
Translocations	320	357

We performed orthologous clustering on Y476, Nip and 9311. In the Y476 genome, 27,454 gene families were identified. Among these, 20,675 of the gene families were common to all three genomes, whereas 303 gene families were specific to the Y476 genome (Supplementary Fig. 9). 690 gene families were expanded and 247 gene families were contracted in Y476 genome, including NBS-LRR and RNA_pol_Rpc4 gene family, which play important roles in disease resistance and grain regulation, respectively. An analysis of Gene Ontology (GO) terms for these specific families revealed that several biological processes, such as DNA integration, viral genome integration into host DNA, and ion transport are enriched in the Y476 genome (Supplementary Table 13). A total of 5,984 novel protein-coding genes absent in the MSU7(Nip)/9311 were predicted in the Y476 genome. Of these novel genes, 18.1% (1,085) were annotated by Pfam, 6.9% (415) genes were annotated by GO terms. Prominently, 155 of 1,085 genes are potential disease resistance genes (R genes), including 106 NB-ARC genes and 49 genes with LRR domain. Some of these domains may lead to the excellent disease resistance of Y476 (Supplementary Table 14).

We identified two large structural variations on Chr. 4 and 11, which contained tandem repeats of gene clusters. We found a gene cluster on Chr. 4 (32 genes) and LOC_Os04g32350 belongs to RNA_pol_Rpc4 gene family and can regulate the expression of genes involving in grain development. The gene cluster on Chr.11 (39 genes) and three NBS-LRR genes in Nip belong to a common gene family, and were discovered by tandem repeat duplication of these three genes (Supplementary Table 15). The number of NBS-LRR resistance genes in Y476 was significantly increased. By RNA-seq with the conditions of rice blast challenge, we identified that the gene expression of these two gene clusters was significantly changed compared with that of Nip gene (Fig. 3C-F).

Development of chromosome segments substitution lines in different cultivated rice genetic backgrounds

Based on the observation that there is wide genetic variation between wild rice accession Y476 and cultivated rice, it is anticipated that some of these variations could be responsible for phenotypic differences that include resistance to biotic and abiotic stresses. In order to dissect the wild rice genomic information and manipulate these variations for directional breeding, we developed two sets of chromosome segment substitution lines (CSSLs) using Y476 as the donor parent, 9311 and Nip as the receipt/recurrent parent, respectively. The procedure for the development of two sets of CSSLs was summarized in Supplementary Fig. 10. The CSSL/9311 population is an advanced backcrossed generation population, contains 198 lines and covers 85% of the wild rice genome, each line has an average of 96.5% recurrent parent genome, and harbor 7.3 substitution segments (Fig. 4A-C, Supplementary Table 16,17). The CSSL/Nip population, a less backcrossed population, contains 225 lines and covers the whole genome of wild rice. Each line has an average of 80.6% recurrent parent Nip genome, and harbor 21 substitution segments (Fig. 4D). The genetic constitution of CSSL/Nip population was analyzed. In the CSSL/Nip population, the sizes of the substituted segments ranged from 0.37 to 106 cM, with an average of 14.4 cM. Twenty-six percent of substituted segments were smaller than 5 cM, 22% of substituted segments were from 5 to 10 cM, and 11% were over 30 cM. 40% of CSSLs contained no more than 20 substituted segments (Fig. 4D-F, Supplementary Table 18,19).

The QTLs mapping was performed based on the phenotype and genotype association using these two CSSL populations. Some domestication-related trait associated genes, such as sd1 for plant height^18,19, sh4 for seeds shattering²⁰, and C1 for red or purple traits^21,22, were observed in these two CSSL populations and easily fine mapped on their exact positions (Supplementary Fig. 11). Variates in these three loci from CSSLs have the wild-type haplotypes which were previously reported. This demonstrated that CSSLs were useful resources for identification of wild rice genes.

QTL mapping of favorable agronomic traits using CSSLs

Using the CSSL/Nip population, nine agronomical traits, including grain length, grain width, 1000-grain weight, grain length to grain width ratio, plant height, panicle length, tiller number, length of flag leaf and width of flag leaf were investigated in three environments and showed a large range of variation (Supplementary Fig. 12, Supplementary Table 20). We identified 244 QTLs associated with these nine agronomic traits in three environments (Fig. 5A, Supplementary Fig. 13, Supplementary Table 21). Among these QTLs, 214 were previously uncharacterized. Approximately 130 novel genes which were not found in the cultivated rice were identified in these QTLs, the new loci from wild rice provided resources for further studies. PAV identification was conducted on the QTL interval, and 1,323 PAVs were identified, which distributed on 199 QTLs on 12 chromosomes (Fig. 5B, Supplementary Table 22).

Several documented genes in QTLs were confirmed for their PAVs between Y476 and Nip. For example, a known OsSWEET14 gene (LOC_Os11g31190), which encodes a sugar transporter and negatively regulated grain weight²³, harbors a 663bp PAV upstream. CSSL with this allele from Y476 showed a reduced expression level and increased grain weight compared with Nip (Supplementary Fig. 14). Therefore, we believe that the existence of these PAVs may be the cause for significant differences in the expression of these genes and thus affect the phenotype.

Identification of chromosome loci controlling salt tolerance using CSSLs

To identify beneficial alleles for abiotic stresses, the salt tolerance was investigated under 0.5% NaCl stress during the whole growth period using the CSSL/Nip population. Phenotypic transgressive variation was observed in this population, nearly half of CSSLs were more salt tolerant than their recurrent parent Nip and half were the same (Supplementary Fig. 15A). Three QTLs were identified and one QTL near loci S2_4579633 associated with a large LOD value of 12.3 were selected (Fig. 6A, Supplementary Table 23), four genes were found in this interval (Supplementary Table 24), and the variation between wild rice and Nip were investigated. One CSSL line, N133, which harbors this locus and is highly salt tolerant (Level 3), was selected for further study (Supplementary Fig. 15B, C). The 20-day seedling of N133 also showed higher salt tolerance than Nip, the survival rate, the fresh and dry weight of above-ground seedlings of N133 were much higher than those of Nip under salt stress (Fig. 6B-F).

Based on the transcriptomic data, we detected the relative expression levels of four genes in this QTL near loci S2_4579633, only one gene, LOC_Os02g08540 has a significantly different transcriptional level between N133 and Nip under salt treatments (Supplementary Fig. 15D-G). LOC_Os02g08540 encodes an unknown expressed protein and was strongly induced by salt stress in N133 plants, which was confirmed with Real-time PCR (Fig. 6G). According to Y476 and Nip genomic sequence, two PAVs, 87-bp and 240-bp deletion, were found in the promoter and downstream regions of this gene respectively (Fig. 6H). Therefore, integrating the transcriptome and genome variant data identified LOC_Os02g08540 as a candidate gene associated with salt tolerance.

As an innovative rice germplasm harbor wild rice gene, line N133 showed a comprehensive excellent salt tolerance, both at the seedling and adult plant stages in the field, whereas N133 survived under 0.5% salt stress during the whole growth period (Supplementary Fig. 15B, C). Additionally, the yield traits of line N133 were also better than those of Nip (Supplementary Fig. 16). Our results provided promising gene and germplasm resources for future rice salt tolerance breeding.

Identification of rice blast resistant gene from wild rice using CSSLs

Seven QTLs for rice blast resistance were identified, and contained 70 opening read frames. Among them, the QTL located near S7_21365207 has largest LOD value of 10.9 and phenotypic variance explained (PVE) of 12.1% (Fig. 7A, Supplementary Table 25). One CSSL/Nip line, N154, which harbors this QTL and was highly resistance to rice blast, was selected for further study. The lesion length on line N154 was almost zero (Fig. 7B). Three genes located near the S7_21365207 locus, i.e. LOC_Os07g35660, LOC_Os07g35670, LOC_Os07g35680 (Supplementary Table 26). Only LOC_Os07g35680 had significant expression level difference between N154 and Nip, and was confirmed by Real-time PCR (Fig. 7C; Supplementary Fig. 17). A 7.8-kb PAV was found in the LOC_Os07g35680 intron region between N154 and Nip (Fig. 7D). qPCR analysis revealed that the expression level of LOC_Os07g35680 was significantly higher in the leaves of N154 than that in Nip, and was significantly increased after rice blast treatment (Fig. 7C, E). Combining with genome variant data, LOC_Os07g35680 was identified as a candidate gene from wild rice for rice blast resistance. LOC_Os07g35680 encoding a receptor-like kinase (RLK). Phylogenetic analysis revealed that LOC_Os07g35680 belonged to RLK family in rice germplasm and clustered with several known RLK genes (Supplementary Fig. 18).

We next compared the transcriptomes of N154 with Nip after rice blast infection using RNA-seq analysis. Global gene expression differences were found between Nip and N154. The analysis revealed 3,788 differentially expressed genes (DEGs) in N154 without blast infection when compared to Nip. Among this set of genes, with blast infection, 841 DEGs were found in N154 when compared to Nip, including 87 DEGs enriched in the GO term of “Cell death”, ‘‘response to stimulus”, ‘‘defense response”, and ‘‘immune response” (Fig. 7F; Supplementary Fig. 19A, B). There are four cloned resistance genes related to rice blast among the 841 DEGs, Pi9²⁴ were significantly up-regulated while OsWAK112d ²⁵, OsMADS26 ^26,27 and Pish were significantly down-regulated in N154 (Supplementary Fig. 19C). The expression pattern of these four rice blast resistance genes were confirmed by Real-time PCR (Fig. 7G).

Using CRISPR/Cas9 technology, we generated knock-out mutants of LOC_Os07g35680 in the N154 background (N154-KO) (Supplementary Fig. 20). As expected, the rice blast resistance of the knock-out plants decreased significantly. The N154-KO lines displayed enhanced susceptibility to Magnaporthe oryzae similar to Nip, and the lesion length on the N154-KO lines was significantly increased compared with N154 (Fig. 7H, I). Subsequently, we detected the expression levels of these four rice blast resistance genes in N154-KO lines, and found that only OsMADS26 expression in N154-KO lines was significantly changed, and the expression alteration was consistent with Nip, which is significantly different from N154 (Fig. 7J). These results demonstrated that the natural variation in the LOC_Os07g35680 locus is critical for the rice resistance to M. oryzae and, led to the transcriptional changes of rice blast resistance gene OsMADS26, which might contribute to an adaptive molecular response to the rice blast resistance in line N154 and wild rice.

Asian cultivated rice O. sativa was domesticated from wild rice O. rufipogon and probably in the wild rice diversity center, which is located in south China. Abounding beneficial alleles were maintained in the wild rice genome due to less modern breeding selection. The primary objectives of this study were to construct a platform and genomic and germplasm resources for high-throughput gene identification from O. rufipogon. The strategy was from “a high-quality reference genome to the genetic population to QTLs to genes”. A high-quality reference genome is critical for understanding genome structure and genetic variation. However, assembling a typical O. rufipogon genome has been challenging because of its high degree of heterozygosity. In the past five years, well documented genome assemblies of O. rufipogon accessions include W1943 with a contig N50 of 34 kb^5,7, an accession from Yunnan province of China, with a contig N50 of 1.1Mb³, IRGC106162 from Laos with a contig N50 of 13.2Mb⁹. All of these accessions are not typical Chinese wild rice and the data are far from reaching a gap-free assembly.

In this study, we carefully selected one wild rice accession, Y476, which represented a typical Chinese wild rice and has high resistance and tolerance to biotic and abiotic stresses (Fig. 1). The Y476 plant used for genome assembly in this study contains a relatively high degree of heterozygosity (1.06%). We de novo assembled its T2T genome (contigs with the N50 length of 33.2Mb) by combining the latest HiFi reads, the ONT Ultra-long sequencing, and genome assembly methods. The genome of Y476 featured evident improvements in the continuity and quality compared with the existing wild rice genomes (Fig. 2, Table 1) and showed a high degree of synteny with those of japonica and indica. One important application of high-quality de novo assembly of the reference genome is to detect and characterize genetic variations in the whole genome, especially structure variations. Recent studies show that PAVs had a greater impact on genomic polymorphism and functional gene variation than SNPs²⁸. PAVs have been found to affect many rice agronomic traits such as hybrid sterility, flowering time, grain size, and disease resistance^29,30. From the global comparison of genomes of Y476 with japonica and indica, abounding PAVs, unique gene family, and novel genes which were absent in the cultivated rice genomes were identified (Fig. 3). This suggested that during rice domestication, many wild rice genes have been lost or changed in the cultivated rice, and many of these genes may be beneficial for the improvement of modern rice varieties.

Development of CSSL is laborious. The two CSSL populations we constructed in the last ten years provided a platform and valuable resources to dissect the wild rice genome and identify elite alleles. The less backcrossed generation of CSSLs which covered the whole wild rice genome were used for QTLs/genes identification, and the advanced backcrossed CSSLs provided single segment substitution lines or near isogenic lines of wild rice genes (Fig. 4). Many key genes involved in domestication were easily identified using these two CSSL populations, demonstrating that CSSLs are effective for gene mapping. QTL identification was performed using the CSSL population, abounding PAVs were found in QTLs according to the genomic data. Combining the genomic data and CSSLs, we have identified novel QTLs/genes associated with agronomic and yield traits and salt stress, and that PAVs played an important role in gene expression regulation (Fig. 5, 6). Moreover, we have identified a novel rice blast resistant gene, wild rice LOC_Os07g35680 allele has a 7.8-kb insertion in its intron region, which led to higher expression level. Further study showed that LOC_Os07g35680 improved rice blast resistance via inhibited the expression of OsMADS26, a negative regulatory factor of rice blast disease (Fig. 7). These QTLs and genes provided a framework for detailed functional analysis of genomic segments in O. rufipogon and should be exploited further for rice breeding in the future.

In summary, we assembled a T2T reference genome of a typical Chinese common wild rice, and were able to identify extensive variations between wild and cultivated rice. These variations had been introgressed into cultivated rice genetic background by the construction of CSSL populations. Our results highlighted the roles of PAVs. We explored beneficial genomic sequences conferring resistance to biotic and abiotic stresses, which could be used for transfer desirable traits in rice breeding. The reference genome and CSSL populations published herein will accelerate wild rice functional genomics studies and genome-enabled stress resistant improvement.

Plant materials, DNA extraction, and library construction

The O. rufipogon accession Y476 was originally collected in Sanya, Hainan Province, China (N18.15°, E109.31°), and conserved in Chinese National Wild rice Germplasm Garden. For genome sequencing, high-quality genomic DNA was extracted from leaves using a modified CTAB method³¹. The quality of DNA was checked by agarose gel electrophoresis, DNA were sequenced with both Illumina HiSeq X Ten (Illumina Inc., San Diego, CA) and PacBio Sequel (Pacific Biosciences of California, Menlo Park, CA) platforms. Portions of the DNA were sent to Frasergen to construct circular consensus sequencing (CCS) libraries and sequence them using a PacBio Sequel platform and Illumina Hiseq platforms, and other portion of DNA was sent to Benagen to construct libraries and sequence using ONT Ultra-long reads. Short reads generated from the Illumina platform were used for the estimation of the genome size, the level of heterozygosity, the mapping rate, and the coverages of the genome, and long reads from the PacBio platform were used for genome assembly. Trizol (Invitrogen) was used to extract RNA from different development stages of rice, including seedling, tillering, and heading stages. Tissues collected for RNA sequencing included panicles, leaves, shoots, and roots. The RNA-seq library was prepared with an insertion size of 300 bp and sequenced on the Illumina platform.

Genome estimation using k-mer analysis

The Illumina short reads were used for survey analysis. The k-mer distributions were estimated using Jellyfish³² with parameters -m 21 -t 1 -s 5G -C. Then the genome sizes and heterozygosity were calculated by the GenomeScope with default parameters³³.

Genome de novo assembly and quality assessment

For the PacBio assemblies, SMRTbell libraries were sequenced on a PacBio Sequel II system and consensus reads (HiFi reads) were generated using CCS software (https://github.com/pacificbiosciences/unanimity) with default parameters. These long (~ 15 kb) and highly accurate (> 99%) HiFi reads were assembled using Hifiasm to generate a draft genome³⁴. Some short contigs were removed and the genomes were manually corrected. For the Nanopore assemblies, we first checked the N50 length of the Ultra-long reads, and then used NextDenove (https://github.com/Nextomics/NextDenovo) for the preliminary assembly of the genome with parameters of genome_size = 430m, read_type = ont, rerun = 3. Some short contigs were removed and the genome was manually correct. Hi-C data were classified as valid or invalid interaction pairs using HiC-Pro v2.11.1³⁵, and only valid interaction pairs were retained for subsequent assembly. 3D-DNA³⁶ was used to cluster, order, and orientate contigs. Then, we compared the genome assembled using HiFi sequences with the genome assembled by NextDenove and filled the gaps by RagTag. After the assembly phase, the gap-free genome was polished using short reads and HiFi reads by Racon and Merfin with six iterative rounds¹⁴. Finally, the heatmap of genomic interactions was plotted by HiCPlotter software³⁷.

The genome completeness was evaluated by BUSCO using the embryophyta_odb10 database³⁸. The genome continuity was evaluated by calculating the contig N50 length. The accuracy of the genome is evaluated by mapping the WGS sequencing data to the genome and calculating the mapping rate and coverages by qualimap2^39,40. Finally, the new method LAI value which is using the repeat sequences was used to evaluate the assembly level of the genome⁴¹. We also assessed the genome assembly using Merqury QV based on 19-mer “hybrid” Merqury k-mer database combining Illumina PCR-free and HiFi reads^15,16.

Identification of centromere and telomere sequence

We collected the 155-165bp CentO satellite DNA sequences in rice. Then HMMER was used to search the location of centromeres in our genome⁴². Centromeric regions were defined as the region containing all the units with high hit scores. The telomeric sequence 5’-CCCTAAA-3’ and the reverse complement of the seven bases were searched directly.

Genome annotation

RepeatMasker was used to mask the genome and annotate the TE elements with a rice high-quality non-redundant TE library⁴³. This TE library was generated by EDTA, and this software had been tested and worked well on the MSU rice genome and was robust for both plant and animal species⁴⁴.

We predicted the structure of the protein-coding genes in the rice genome using three gene prediction methods, including ab initio, homology-based, and RNA-seq data. Before gene prediction, the assembled genome was hard and soft masked using RepeatMasker. We adopted Augustus (Unsupervised training), GlimmerHMM (Models available), and Braker2 (Prediction based on transcriptome data) to perform ab initio gene prediction^45–47. Exonerate (v2.2.0) was used to conduct homology-based gene prediction. The protein sequences from MSUv7, MH63RS2, ZS97RS2, R498 rice genome, and TAIR10 Arabidopsis thaliana were aligned to our genome assembly and coding genes were predicted using Exonerate with the default parameters⁴⁸. RNA sequencing data were ~ 6Gb each. Evidence of transcriptome included de novo assembly of transcripts (based on Trinity to PASA) and genome-guided (Hisat2 to StringTie to Transdecoder^49–52. EVidenceModeler was used to integrate all prediction results of the three methods to predict genes models⁵³. Finally, gene models were filtered by removing those gene coding sequences overlapping with TE sequence by more than 20% or the coding region length was less than 150bp.

There are three ways to predict the functions of protein-coding genes. The first is to use BlastP to search against protein sequence in the NCBI non-redundant protein database (NR) and Swiss-Prot (http://web.expasy.org/docs/swiss-prot/guideline.html) database⁵⁴. Second, protein domain and gene ontology term annotations were performed using InterProScan⁵⁵. The third is KEGG annotation with KEGG Automatic Annotation Server⁵⁶. tRNAscan-SE was used to identify tRNA genes with default parameters⁵⁷. RNAmmer was used to predict rRNA sequences⁵⁸.

Synteny analysis

Nucmer was used to perform comparisons between genomes with parameters --mum --mincluster 200 --minmatch 100⁵⁹. Delta-filter was used to filter the results of the alignment file produced by nucmer with parameters -i 95 -l 100 -1. The dotplot was generated by mummer’s function mummerplot.

Genome-wide comparisons, and SNP, InDel, and structure variations identification

MUMmer4 was used to compare genomes of Y476 to Nip with parameters: -maxmatch -c 100 -l 50. Then, we filtered delta files using the delta filter − 1 parameter, which allowed us to only keep the best alignments. Finally, we identified SNPs and InDels using show-snps (with -ClrT parameter) for filtered delta files. Structure variations were identified using MUMandCo with default parameters based on the whole genome alignment information provided by MUMmer (v4) to detect variants⁵⁹.

Distribution of structure variations relative to gene position and annotation

We classified PAVs into five categories according to the overlapping genomic regions with PAVs. PAVs share overlaps with the coding region, intron, ± 2 kb of genes, and intergenic regions. If an PAV shared overlaps with two or more different genomic regions, the PAV was classified into different genomic regions at the same time. Then we calculated the percentage of each PAV category in all PAVs. If a PAV overlaps with the gene region or the upstream and downstream 2 kb region, we consider that this PAV will affect this gene, and annotate this PAV as the gene it affects.

Construction and genotyping of CSSLs

The CSSL populations used in this study were constructed with wild rice Y476 as the donor parent, and ‘9311’ and ‘Nip’ as the recipient parents respectively. Schematic of the development of CSSLs was shown in Supplementary Fig. 10. The whole genome survey during the construction of CSSL/9311 population was performed suing SSR and InDel molecular markers. And the whole genome survey of CSSL/Nip population was performed by genome sequencing. We eliminated some lines with a lot of genetic background noises and selected lines with relatively few segments, which were self-pollinated to produce homozygous lines as much as possible.

Genotype identification of CSSL/Nip population was completed by genotyping by target sequencing (GBTS) with 10 k SNPs. To identify introgression segments from Y476 to Nip, a sliding-window approach was applied⁶⁰. Based on the SNP sequencing data completed by GBTS technology, consecutive SNPs were examined in a sliding window of 15 SNPs. The ratio between the numbers of SNPs from the Y476 and Nip was calculated in each window. Based on the allele ratio (Nip/Y476), each window genotype was then defined as having a homozygous Nip genotype (larger than 11:4) or a homozygous Y476 genotype (smaller than 2:13). In this way, we identified the recombination breakpoints in all the lines, then we aligned all the chromosomes of all the lines and compared them in a minimal of 100-kb intervals. Adjacent 100-kb intervals with the same genotype across the entire population were recognized as a single recombination bin. Finally, 1,542 recombination bins were generated for CSSL/Nip population genotyping.

Phenotype identification and QTLs mapping

Phenotypes of nine agronomic traits were investigated in three environments (Supplementary Table 27). For identification of salt tolerance, all CSSLs were planted in a salt stress pool, with 0.5% salt stress treatment during the whole growth period. The salt tolerance level of each line was evaluated after heading. All experiments were repeated three times independently, and investigated data were performed according to the evaluation criteria of rice germplasm⁶¹. The identification of QTLs for each trait was performed with QTL IciMapping 4.1⁶². The RSETP-LRT-ADD mapping method was applied with a logarithm of odds (LOD) threshold of 2.5.

RNA isolation and qRT-PCR

Total RNA was extracted using RNA Easy Fast Plant Tissue Kit (TIANGEN, Beijing, China) according to the manufacturer’s protocols. Reverse transcription (RT) reactions were performed with PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa, Dalian, China) according to the manufacturer’s instructions. Real-time qPCR experiments were performed using a SuperReal PreMix Plus (SYBR Green) (TIANGEN, Beijing, China) on a CFX96 Real-Time PCR System (Bio-Rad, Beijing, China). The Actin gene (LOC_Os03g50885) was used for the normalization of all qRT-PCR data. All the qRT-PCR analyses were performed for at least three independent biological replicates. The 2^−△△CT method was used to calculate the relative expression levels with three technical replicates⁶³. Primers are listed in Supplementary Table 28.

RNA-seq analysis

Total RNA was isolated from the leaves of Nipponbare and N154 at 0, 24 h and 48 h after inoculating rice blast fungus with three biological replicates. Leaves of Nipponbare and N133 were collected at 0, 12 h and 24 h after treatment with 200 mmol/L NaCl. RNA-seq data were analyzed following the standard protocol as described previously⁶⁴. The reads were mapped to the rice reference genome (MSU v7) using Hisat2^51,65. Calculation of read counts and the identification of DEGs were performed using DEseq2 ^64,66. Genes with an expression |log2Fold Change| >= 1 and FDR < 0.05 were defined as a statistically significantly differentially expressed. GO enrichment analysis was performed using PlantGSEA and AgriGO^67,68.

Fungus and inoculation

Magnaporthe oryzae isolates were grown on oatmeal agar for about two weeks at 28℃ before producing spores. Conidia were induced under light for 2–3 days and then the spores were collected in sterile water with Tween 20. The blast resistance of rice plants in blast nursery was artificially identified by transplanting the diseased plants as spreader lines into the field. The identification of resistance to rice blast in the laboratory is spray inoculation and punch inoculation. For punch inoculation, 4 µL spore suspension was pipetted at three spots on each leaf, and leaves were kept in a culture dish containing 0.1% 6-benzoaminopurine sterile water to keep moist. Lesion length was measured five days post inoculation⁶⁹. All inoculation experiments were repeated three times independently.

Acknowledgments

This work was supported by the National Key R&D Program of China (2021YFD1200100, and 2021YFD1200501), Hainan Yazhou Bay Seed Laboratory (project of B21HJ0215), Project of Sanya Yazhou Bay Science and Technology City (SKJC-2020-02-001), and the Agricultural Science and Technology Innovation Program of Chinese Academy of Agricultural Sciences.

Author contributions

W.H.Q. and H.H. conceived and designed the experiments. J.F.H and Y.L.Z. performed most of the experiments. Y.P.L. performed the rice blast resistance identification and transgenic experiments. Q.W.Y. and Q.Q. supervised the project. J.F.H., Y.L.Z., W.H.Q. and H.H. analyzed data and wrote the manuscript. The other authors performed some of the experiments. All authors read and approved the manuscript.

Competing interests

The authors declare no competing interests.

Khush, G. S. What will it take to feed 5.0 million rice consumers in 2030. Plant Mol. Biol. 59, 1-6 (2005).
Huang, X. et al. A map of rice genome variation reveals the origin of cultivated rice. Nature 490, 497-501 (2012).
Li, W. et al. SMRT sequencing of the Oryza rufipogon genome reveals the genomic basis of rice adaptation. Commun Biol. 3, 167 (2020).
Gao, L., Zhang, S., Zhou, Y., Ge, S. & Hong, D. A survey of the current status of wild rice in China. Biodiv. Sci. 4, 160-166 (1996).
Zhao, Q. et al. Pan-genome analysis highlights the extent of genomic variation in cultivated and wild rice. Nat. Genet. 50, 278-284 (2018).
Lin, S. & Yuan, L. Innovative approaches to rice breeding. International rice research conference 35-51 (1980).
Stein, J. C. et al. Genomes of 13 domesticated and wild rice relatives highlight genetic conservation, turnover and innovation across the genus Oryza. Nat. Genet. 50, 285-296 (2018).
Zhang. F. et al. Long-read sequencing of 111 rice genomes reveals significantly larger pan-genomes. Genome Res. 32, 853-863 (2022).
Xie, X. et al. A chromosome-level genome assembly of the wild rice Oryza rufipogon facilitates tracing the origins of Asian cultivated rice. Sci. China Life Sci. 64, 282-293 (2021).
Shang, L. et al. A super pan-genomic landscape of rice. Cell Res. 32, 878-896 (2022).
Ali, M. L., Sanchez, P. L., Yu, S., Lorieux, M. & Eizenga, G. C. Chromosome Segment Substitution Lines: A Powerful Tool for the Introgression of Valuable Genes from Oryza Wild Species into Cultivated Rice (O. sativa). Rice 3, 218-234 (2010).
Balakrishnan, D., Surapaneni, M., Mesapogu, S. & Neelamraju, S. Development and use of chromosome segment substitution lines as a genetic resource for crop improvement. Theor. Appl. Genet. 132, 1-25 (2019).
Li, J., W, J. & Zeigler, R. S. The 3,000 rice genomes project: new opportunities and challenges for future rice research. GigaSci. 3, 7 (2014).
Du, H. et al. Sequencing and de novo assembly of a near complete indica rice genome. Nat. Commun. 8, 15324 (2017).
Mc Cartney, A. M. et al. Chasing perfection: validation and polishing strategies for telomere-to-telomere genome assemblies. Nat. Methods 19, 687-695 (2022).
Rhie, A., Walenz, B. P., Koren, S. & Phillippy, A. M. Merqury: reference-free quality, completeness, and phasing assessment for genome assemblies. Genome Biol. 21, 245 (2020).
Yu, H. et al. A route to de novo domestication of wild allotetraploid rice. Cell 184, 1156-1170 (2021).
Spielmeyer, W., Ellis, M. H. & Chandler, P. M. Semidwarf (sd-1), "green revolution" rice, contains a defective gibberellin 20-oxidase gene. Proc. Natl. Acad. Sci. USA. 99, 9043-9048 (2002).
Zhang, L. et al. Identification and genetic analysis of qCL1.2, a novel allele of the “green revolution” gene SD1 from wild rice (Oryza rufipogon) that enhances plant height. BMC Genetics 21, 62 (2020).
Li, C., Zhou, A. & Sang, T. Rice Domestication by Reducing Shattering dressed. Science 311, 1936-1939 (2006).
Saitoh, K., Onishi, K., Mikami, I., Thidar, K. & Sano, Y. Allelic diversifcation at the C (OsC1) locus of wild and cultivated rice: nucleotide changes associated with phenotypes. Genetics 168, 997-1007 (2004).
Qiao, W. et al. A functional chromogen gene C from wild rice is involved in a different anthocyanin biosynthesis pathway in indica and japonica. Theor. Appl. Genet. 134, 1531-1543 (2021).
Kim, P. et al. Tissue-specific activation of DOF11 promotes rice resistance to sheath blight disease and increases grain weight via activation of SWEET14. Plant Biotech. J. 19, 409-411 (2021).
Liu, G., Lu, G., Zeng, L. & Wang, G. Two broad-spectrum blast resistance genes, Pi9(t) and Pi2(t), are physically linked on rice chromosome 6. Mol. Genet. Genomics 267, 472-480 (2002).
Delteil, A. et al. Several wall-associated kinases participate positively and negatively in basal defense against rice blast fungus. BMC Plant Biol. 16, 17 (2016).
Lee, S. et al. Further characterization of a rice AGL12 group MADS-box gene, OsMADS26. Plant Physiol. 147, 156-168 (2008).
Khong, G. N. et al. OsMADS26 Negatively regulates resistance to pathogens and drought tolerance in rice. Plant Physiol. 169, 2935-2949 (2015).
Wellenreuthe, M., Mérot, C., Berdan, E. & Bernatchez, L. Going beyond SNPs: The role of structural genomic variants in adaptive evolution and species diversification. Mol Ecol. 28, 1203–1209 (2019).
Shen, R. et al. Genomic structural variation-mediated allelic suppression causes hybrid male sterility in rice. Nat. Commun. 8, 1310 (2017).
Lye, Z. N. & Purugganan, M. D. Copy number variation in domestication. Trends Plant Sci. 24, 352-365 (2019).
Porebski, S., Bailey, L. G. & Baum, B. R. Modification of a CTAB DNA extraction protocol for plants containing high polysaccharide and polyphenol components. Plant Mol. Biol. Rep. 15, 8-15 (1997).
Marçais, G. & Kingsford, C. A fast, lock-free approach for efficient parallel counting of occurrences of k-mers. Bioinformatics 27, 764-770 (2011).
Vurture, G. W. et al. GenomeScope: fast reference-free genome profiling from short reads. Bioinformatics 33, 2202-2204 (2017).
Cheng, H., Concepcion, G. T., Feng, X., Zhang, H. & Li, H. Haplotype-resolved de novo assembly using phased assembly graphs with HiFIasm. Nat. Methods 18, 170-175 (2021).
Servant, N. et al. HiC-Pro: an optimized and flexible pipeline for Hi-C data processing. Genome Biol. 16, 259 (2015).
Burton, J. N. et al. Chromosome-scale scaffolding of de novo genome assemblies based on chromatin interactions. Nat. Biotechnol. 31, 1119-1125 (2013).
Akdemir, K. C. & Chin, L. HiCPlotter integrates genomic data with interaction matrices. Genome Biol. 16, 198 (2015).
Manni, M., Berkeley, M. R., Seppey, M., Simão, F. A. & Zdobnov, E. M. BUSCO update: novel and streamlined workflows along with broader and deeper phylogenetic coverage for Scoring of eukaryotic, prokaryotic, and viral Genomes. Mol. Biol. Evol. 38,4647-4654 (2021).
Okonechnikov, K., Conesa, A. & García-Alcalde, F. Qualimap2: advanced multi-sample quality control for high throughput sequencing data. Bioinformatics 32,292-294 (2016).
Li, H. & Durbin, R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics 26, 589-595 (2010).
Ou, S., Chen, J. & Jiang, N. Assessing genome assembly quality using the LTR Assembly Index (LAI). Nucleic Acids Res. 46, e126 (2018).
Finn, R. D., Clements, J. & Eddy, S. R. HMMER web server: interactive sequence similarity searching. Nucleic Acids Res. 39, W29-W37 (2011).
Shi, J. & Liang, C. Generic Repeat Finder: A High-Sensitivity Tool for Genome-Wide De Novo Repeat Detection. Plant Physiol. 180, 1803-1815 (2019).
Ou, S. et al. Benchmarking transposable element annotation methods for creation of a streamlined, comprehensive pipeline. Genome Biol. 20, 275 (2019).
Urnov, F. D. et al. Highly efficient endogenous human gene correction using designed zinc-finger nucleases. Nature 435, 646-651(2005).
Majoros, W. H., Pertea, M. & Salzberg, S. L. TigrScan and GlimmerHMM: two open source ab initio eukaryotic gene-finders. Bioinformatics 20, 2878-2879 (2004).
Br°una, T., Hoff, K.J., Lomsadze, A., Stanke, M. & Borodovsky, M. BRAKER2: automatic eukaryotic genome annotation with GeneMark-EP+ and AUGUSTUS supported by a protein database. NAR Genom. Bioinform. 3, 1-21 (2021).
Kikuchi, K., Terauchi, K., Wada, M. & Hirano, H. Y. The plant MITE mPing is mobilized in anther culture. Nature 421, 167 (2003).
Grabherr, M. G. et al. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat. Biotechnol. 29,644-652 (2011).
Haas, B. J. et al.De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nat. Protoc. 8,1494-1512 (2013).
Pertea, M., Kim, D., Pertea, G. M., Leek, J. T. & Salzberg, S. L. Transcript-level expression analysis of RNA-seq experiments with HISAT, StringTie and Ballgown. Nat. Protoc. 11, 1650-1667 (2016).
Kovaka, S. et al. Transcriptome assembly from long-read RNA-seq alignments with StringTie2. Genome Biol. 20, 278 (2019).
Haas, B. J. et al. Automated eukaryotic gene structure annotation using EVidenceModeler and the Program to Assemble Spliced Alignments. Genome Biol. 9, R7 (2008).
Camacho, C. et al. BLAST+: architecture and applications. BMC Bioinformatics 10,421 (2009).
Quevillon, E. et al. InterProScan: protein domains identifier. Nucleic Acids Res. 33, W116-W120 (2005).
Kanehisa, M. et al. Data, information, knowledge and principle: back to metabolism in KEGG. Nucleic Acids Res. 42, D199-D205 (2014).
Lowe, T. M. & Eddy, S. R. tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25, 955-964 (1997).
Nawrocki, E. P. & Eddy, S. R. Infernal 1.1: 100-fold faster RNA homology searches. Bioinformatics 29, 2933-2935 (2013).
Marçais, G. et al. MUMmer4: a fast and versatile genome alignment system. PLoS Comput. Biol. 14, e1005944 (2018).
Huang, X. et al. High-throughput genotyping by whole-genome resequencing. Genome Res. 19, 1068-1076 (2018).
Han, L., Wei, X... Descriptors and data standard for rice (Oryza sativa L.) (China Agriculture. Press, Beijing, 2006).
Meng, L., Li, H., Zhang, L. & Wang, J. QTL IciMapping: integrated software for genetic linkage map construction and quantitative trait locus mapping in biparental populations. Crop J. 3,269–283 (2015).
Livak, K. J. & Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2^-[Δ][Δ]Ct method. Methods 25,402-408 (2001).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Kawahara, Y. et al. Improvement of the Oryza sativa Nipponbare reference genome using next generation sequence and optical map data. Rice 6, 4 (2013).
Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30,923-930 (2014).
Yi, X., Du, Z. & Su, Z. PlantGSEA: a gene set enrichment analysis toolkit for plant community. Nucleic Acids Res. 41, W98-W103 (2013).
Tian, T. et al. agriGO v2.0: a GO analysis toolkit for the agricultural community, 2017 update. Nucleic Acids Res. 45,W122-W129 (2017).
Li, W. et al. A natural allele of a transcription factor in rice confers broad-spectrum blast resistance. Cell 170, 114-126 (2017).

There is NO Competing Interest.

SupplementaryFiguresandTables.zip
Dataset 1

Download PDF

Journal Publication

published 29 May, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Telomere-to-telomere gap-free genome assembly and chromosome segment substitution lines facilitated gene identification in wild rice

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1