High quality and complete genotype data. The genotypes of all accessions were imputed and evaluated, resulting in an overall missing data rate of <3% and an estimated accuracy greater than 99%. The SNP/INDEL genotypes of all accessions are available for online queries and download. To facilitate population genetic analysis, RiceVarMap also offers ancestral allele information and allele distribution data of subpopulations.

Comprehensive annotations of genomic variations. RiceVarMap now provides more precise variations and annotations. Haplotype-based approaches and conservatism of a segment in species were used to evaluate the impact of missense variations. To understanding the impact of variation located in the noncoding regions, we quantified genome-wide chromatin accessibility, which indicate the binding sites and the binding intensites of transcription factors. Based on these high-quality data, we then used Basenji (Kelley et al., 2018), a deep learning method to train the effect of nucleotide sequence on changes in chromatin openness. Thirdly, we used variant sites identified in rice populations and trained model to predict the variant effect in different tissues.

Phenotype data and GWAS results. The database provides geographical details and phenotype images, agronomic and metabolic traits for some rice accessions. Plant scientists and breeders can also search for significant SNPs associated with various traits to develop useful molecular markers or pick up candidate genes.

New query methods and more tools. Besides all functions offered by RiceVarMap v1.0, we have designed more powerful visualization tools based on bokeh and JBrowse. Specifically, for non-coding variation, we provide an online prediction tool, Regulatory Variant Prioritization, users can query the effect of variants online by submitting variant IDs in the database or a VCF file.

Genotype data

Currently, we collected sequencing data from three sets of rice germplasms consisting of totally 4,726 accessions of cultivated rice (Oryza sativa L.):

The first set of germplasm consisted of 533 accessions selected to represent both the usefulness in rice improvement and the genetic diversity in the cultivated species. We sequenced the 533 accessions using the Illumina HiSeq 2000 in the form of 90-bp paired-end reads to generate high-quality sequences of more than one gigabase per accession (>2.5x per genome, total 6.7 billion reads). These raw data is available in NCBI with BioProject accession number PRJNA171289. We provide phenotype images, agronomic and metabolic traits for these accessions.

The second set of germplasm was 950 rice accessions sequenced by Huang et al. (2012, Nat Genet, 44:32-39) that were downloaded from the EBI European Nucleotide Archive (accession number ERP000106 and ERP000729), which consists of 4.6 billion 73-bp paired-end reads (~1x per genome).

The third set of germplasm was 3024 rice accessions from The 3,000 Rice Genomes Project (2014, GigaScience, 3:7) that were downloaded from the EBI European Nucleotide Archive (accession number PRJEB6180), and unpublished resequencing data from 219 rice accessions, which has an average sequencing depth of 14x per genome. 

Recognizing the limitations of using the Japonica reference genome (e.g., Nipponbare) for comprehensively analyzing variation unique to the Indica subgroup, RiceVarMap incorporates a specialized set of 4K genotype data derived from a cooperative project. This data was specifically aligned to the Indica reference genome (MH63RS3) to better capture Indica-specific genomic features. For this MH63-aligned dataset, we currently provide limited but high-value services crucial for comparative studies: subpopulation genotype and frequency queries, and non-coding region Basenji variant annotations.

Phenotype data and GWAS results

RiceVarMap currently provides phenotypic data and GWAS results for 13 agronomic traits (including heading date,  plant height, and grain weight et al.) and 840 metabolite traits (Xie et al., 2015, Proc Natl Acad Sci USA, 112: E5411-E5419; Chen et al., 2014, Nat Genet, 46:714-721).  Phenotype information is available on this page.

Chromatin accessibility data

To generate a comprehensive landscape of accessible chromatin in rice (Oryza sativa), we took advantage of an improved ATAC-seq protocol (UMI-ATAC-seq, which incorporates unique molecular identifiers into the regular ATAC-seq technique for accurate quantification and footprinting) to perform chromatin accessibility profiling in 23 tissues/organs spanning the entire life cycle of rice. The representative tissues include callus, radicle, plumule, leaf, leaf sheath, root, apical meristem (AM1/AM2), dormant buds (DBuds), shoot apical meristem (SAM1/SAM2/SAM3), panicle neck node (PNN), stem, young panicle (Panicle1/Panicle2/Panicle3/Panicle4), lemma, palea, pistil, stamen, and seed coat (Seed1/Seed2/Seed3). The experiments were conducted in three representative rice varieties, namely Nipponbare (NIP; japonica subspecies), Minghui 63 (MH63; indica subspecies type II), and Zhenshan 97 (ZS97; indica subspecies type I), with each experiment consisting of at least two biological replicates. In total, 145 genome-wide chromatin accessibility datasets with high sequencing depth (~30.7 M reads on average) were generated.