Transcriptome-wide association analyses reveal the impact of regulatory variants on rice panicle architecture and causal gene regulatory networks.

Panicle architecture is a key determinant of rice grain yield and is mainly determined at the 1-2 mm young panicle stage. Here, we investigated the transcriptome of the 1-2 mm young panicles from 275 rice varieties and identified thousands of genes whose expression levels were associated with panicle traits. Multimodel association studies suggested that many small-effect genetic loci determine spikelet per panicle (SPP) by regulating the expression of genes associated with panicle traits. We found that alleles at cis-expression quantitative trait loci of SPP-associated genes underwent positive selection, with a strong preference for alleles increasing SPP. We further developed a method that integrates the associations of cis- and trans-expression components of genes with traits to identify causal genes at even small-effect loci and construct regulatory networks. We identified 36 putative causal genes of SPP, including SDT (MIR156j) and OsMADS17, and inferred that OsMADS17 regulates SDT expression, which was experimentally validated. Our study reveals the impact of regulatory variants on rice panicle architecture and provides new insights into the gene regulatory networks of panicle traits.

Bract suppression regulated by the miR156/529-SPLs-NL1-PLA1 module is required for the transition from vegetative to reproductive branching in rice.

Reproductive transition of grasses is characterized by switching the pattern of lateral branches, featuring the suppression of outgrowth of the subtending leaves (bracts) and rapid formation of higher-order branches in the inflorescence (panicle). However, the molecular mechanisms underlying such changes remain largely unknown. Here, we show that bract suppression is required for the reproductive branching in rice. We identified a pathway involving the intrinsic time ruler microRNA156/529, their targets SQUAMOSA PROMOTER BINDING PROTEIN LIKE (SPL) genes, NECK LEAF1 (NL1), and PLASTOCHRON1 (PLA1), which regulates the bract outgrowth and thus affects the pattern switch between vegetative and reproductive branching. Suppression of the bract results in global reprogramming of transcriptome and chromatin accessibility following the reproductive transition, while these processes are largely dysregulated in the mutants of these genes. These discoveries contribute to our understanding of the dynamic plant architecture and provide novel insights for improving crop yields.

ATAC-seq with unique molecular identifiers improves quantification and footprinting.

ATAC-seq (Assay for Transposase-Accessible Chromatin with high-throughput sequencing) provides an efficient way to analyze nucleosome-free regions and has been applied widely to identify transcription factor footprints. Both applications rely on the accurate quantification of insertion events of the hyperactive transposase Tn5. However, due to the presence of the PCR amplification, it is impossible to accurately distinguish independently generated identical Tn5 insertion events from PCR duplicates using the standard ATAC-seq technique. Removing PCR duplicates based on mapping coordinates introduces increasing bias towards highly accessible chromatin regions. To overcome this limitation, we establish a UMI-ATAC-seq technique by incorporating unique molecular identifiers (UMIs) into standard ATAC-seq procedures. UMI-ATAC-seq can rescue about 20% of reads that are mistaken as PCR duplicates in standard ATAC-seq in our study. We demonstrate that UMI-ATAC-seq could more accurately quantify chromatin accessibility and significantly improve the sensitivity of identifying transcription factor footprints. An analytic pipeline is developed to facilitate the application of UMI-ATAC-seq, and it is available at https://github.com/tzhu-bio/UMI-ATAC-seq .